Skip to contentSkip to conversation

5.1.6-stable review

PATCH In review Greg Kroah-Hartman 6 participants 414 messages May 30, 2019 – May 31, 2019 Page 2 of 3

Subsystems (12)
lore
Share & tools
lore
author maintainer
#141 lore raw

PATCH 141/405 mm/uaccess: Use unsigned long to placate UBSAN warnings on older GCC versions

In reply to Greg Kroah-Hartman

[ Upstream commit 29da93fea3ea39ab9b12270cc6be1b70ef201c9e ]

Randy reported objtool triggered on his (GCC-7.4) build:

lib/strncpy_from_user.o: warning: objtool: strncpy_from_user()+0x315: call to __ubsan_handle_add_overflow() with UACCESS enabled lib/strnlen_user.o: warning: objtool: strnlen_user()+0x337: call to __ubsan_handle_sub_overflow() with UACCESS enabled

This is due to UBSAN generating signed-overflow-UB warnings where it should not. Prior to GCC-8 UBSAN ignored -fwrapv (which the kernel uses through -fno-strict-overflow).

Make the functions use 'unsigned long' throughout.

---
 lib/strncpy_from_user.c | 5 +++--
 lib/strnlen_user.c      | 4 ++--
 2 files changed, 5 insertions(+), 4 deletions(-)
lib/strncpy_from_user.c +3 -2
index 58eacd41526c5..023ba9f3b99f0 100644
@@ -23,10 +23,11 @@
  * hit it), 'max' is the address space maximum (and we return
  * -EFAULT if we hit it).
  */
-static inline long do_strncpy_from_user(char *dst, const char __user *src, long count, unsigned long max)
+static inline long do_strncpy_from_user(char *dst, const char __user *src,
+					unsigned long count, unsigned long max)
 {
 	const struct word_at_a_time constants = WORD_AT_A_TIME_CONSTANTS;
-	long res = 0;
+	unsigned long res = 0;
 
 	/*
 	 * Truncate 'max' to the user-specified limit, so that
lib/strnlen_user.c +2 -2
index 1c1a1b0e38a5f..7f2db3fe311fd 100644
@@ -28,7 +28,7 @@
 static inline long do_strnlen_user(const char __user *src, unsigned long count, unsigned long max)
 {
 	const struct word_at_a_time constants = WORD_AT_A_TIME_CONSTANTS;
-	long align, res = 0;
+	unsigned long align, res = 0;
 	unsigned long c;
 
 	/*
@@ -42,7 +42,7 @@ static inline long do_strnlen_user(const char __user *src, unsigned long count,
 	 * Do everything aligned. But that means that we
 	 * need to also expand the maximum..
 	 */
-	align = (sizeof(long) - 1) & (unsigned long)src;
+	align = (sizeof(unsigned long) - 1) & (unsigned long)src;
 	src -= align;
 	max += align;
 
Signature · 1 line
2.20.1
author maintainer
#142 lore raw

PATCH 142/405 Bluetooth: hci_qca: Fix crash with non-serdev devices

In reply to Greg Kroah-Hartman

[ Upstream commit ecf2b768bd11e2ff09ecbe621b387d0d58e970cf ]

qca_set_baudrate() calls serdev_device_wait_until_sent() assuming that the HCI is always associated with a serdev device. This isn't true for ROME controllers instantiated through ldisc, where the call causes a crash due to a NULL pointer dereferentiation. Only call the function when we have a serdev device. The timeout for ROME devices at the end of qca_set_baudrate() is long enough to be reasonably sure that the command was sent.

Fixes fa9ad876b8e0 ("Bluetooth: hci_qca: Add support for Qualcomm Bluetooth chip wcn3990")
---
 drivers/bluetooth/hci_qca.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/bluetooth/hci_qca.c +2 -1
index 237aea34b69f1..340c3c750b180 100644
@@ -992,7 +992,8 @@ static int qca_set_baudrate(struct hci_dev *hdev, uint8_t baudrate)
 	while (!skb_queue_empty(&qca->txq))
 		usleep_range(100, 200);
 
-	serdev_device_wait_until_sent(hu->serdev,
+	if (hu->serdev)
+		serdev_device_wait_until_sent(hu->serdev,
 		      msecs_to_jiffies(CMD_TRANS_TIMEOUT_MS));
 
 	/* Give the controller time to process the request */
Signature · 1 line
2.20.1
author maintainer
#143 lore raw

PATCH 143/405 Bluetooth: hci_qca: Give enough time to ROME controller to bootup.

In reply to Greg Kroah-Hartman

[ Upstream commit 7f09d5a6c33be66a5ca19bf9dd1c2d90c5dfcf0d ]

This patch enables enough time to ROME controller to bootup after we bring the enable pin out of reset.

---
 drivers/bluetooth/hci_qca.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/bluetooth/hci_qca.c +2 -0
index 340c3c750b180..d3b467792eb3d 100644
@@ -508,6 +508,8 @@ static int qca_open(struct hci_uart *hu)
 		qcadev = serdev_device_get_drvdata(hu->serdev);
 		if (qcadev->btsoc_type != QCA_WCN3990) {
 			gpiod_set_value_cansleep(qcadev->bt_en, 1);
+			/* Controller needs time to bootup. */
+			msleep(150);
 		} else {
 			hu->init_speed = qcadev->init_speed;
 			hu->oper_speed = qcadev->oper_speed;
Signature · 1 line
2.20.1
author maintainer
#144 lore raw

PATCH 144/405 Bluetooth: btbcm: Add default address for BCM43341B

In reply to Greg Kroah-Hartman

[ Upstream commit 5035726128cd2e3813ee44deedb9898509edb232 ]

The BCM43341B has the default MAC address 43:34:1B:00:1F:AC if none is given. This address was found when enabling Bluetooth on multiple Intel Edison modules. It also contains the sequence 43341B, the name the chip identifies itself as. Using the same BD_ADDR is problematic when having multiple Intel Edison modules in each others range. The default address also has the LAA (locally administered address) bit set which prevents a BNEP device from being created, needed for BT tethering.

Add this to the list of black listed default MAC addresses and let the user configure a valid one using f.i.
btmgmt -i hci0 public-addr xx:xx:xx:xx:xx:xx

---
 drivers/bluetooth/btbcm.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/bluetooth/btbcm.c +3 -1
index d5d6e6e5da3bf..62d3aa2b26f60 100644
@@ -37,6 +37,7 @@
 #define BDADDR_BCM43430A0 (&(bdaddr_t) {{0xac, 0x1f, 0x12, 0xa0, 0x43, 0x43}})
 #define BDADDR_BCM4324B3 (&(bdaddr_t) {{0x00, 0x00, 0x00, 0xb3, 0x24, 0x43}})
 #define BDADDR_BCM4330B1 (&(bdaddr_t) {{0x00, 0x00, 0x00, 0xb1, 0x30, 0x43}})
+#define BDADDR_BCM43341B (&(bdaddr_t) {{0xac, 0x1f, 0x00, 0x1b, 0x34, 0x43}})
 
 int btbcm_check_bdaddr(struct hci_dev *hdev)
 {
@@ -82,7 +83,8 @@ int btbcm_check_bdaddr(struct hci_dev *hdev)
 	    !bacmp(&bda->bdaddr, BDADDR_BCM20702A1) ||
 	    !bacmp(&bda->bdaddr, BDADDR_BCM4324B3) ||
 	    !bacmp(&bda->bdaddr, BDADDR_BCM4330B1) ||
-	    !bacmp(&bda->bdaddr, BDADDR_BCM43430A0)) {
+	    !bacmp(&bda->bdaddr, BDADDR_BCM43430A0) ||
+	    !bacmp(&bda->bdaddr, BDADDR_BCM43341B)) {
 		bt_dev_info(hdev, "BCM: Using default device address (%pMR)",
 			    &bda->bdaddr);
 		set_bit(HCI_QUIRK_INVALID_BDADDR, &hdev->quirks);
Signature · 1 line
2.20.1
author maintainer
#145 lore raw

PATCH 145/405 Bluetooth: mediatek: Fixed incorrect type in assignment

In reply to Greg Kroah-Hartman

[ Upstream commit cac63f9b163700fb70a609ad220697c61b797d6b ]

Fixed warning: incorrect type in assignment reported by kbuild test robot. The detailed warning is shown as below.

make ARCH=x86_64 allmodconfig
make C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__'

All warnings (new ones prefixed by >>):

btmtkuart.c:671:18: sparse: warning: incorrect type in assignment
(different base types)
btmtkuart.c:671:18: sparse: expected unsigned int [usertype] baudrate
btmtkuart.c:671:18: sparse: got restricted __le32 [usertype]

sparse warnings: (new ones prefixed by >>) btmtkuart.c:671:18: sparse: warning: incorrect type in assignment
(different base types)
btmtkuart.c:671:18: sparse: expected unsigned int [usertype] baudrate
btmtkuart.c:671:18: sparse: got restricted __le32 [usertype]

vim +671 drivers/bluetooth/btmtkuart.c

659
660 static int btmtkuart_change_baudrate(struct hci_dev *hdev)
661 {
662 struct btmtkuart_dev *bdev = hci_get_drvdata(hdev);
663 struct btmtk_hci_wmt_params wmt_params;
664 u32 baudrate;
665 u8 param;
666 int err;
667
668 /* Indicate the device to enter the probe state the host is
669 * ready to change a new baudrate.
670 */

Quoted 1 line

671 baudrate = cpu_to_le32(bdev->desired_speed);

672 wmt_params.op = MTK_WMT_HIF;

Fixes 22eaf6c9946a ("Bluetooth: mediatek: add support for MediaTek MT7663U and MT7668U UART devices")
---
 drivers/bluetooth/btmtkuart.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/bluetooth/btmtkuart.c +1 -1
index b0b680dd69f49..f5dbeec8e2748 100644
@@ -661,7 +661,7 @@ static int btmtkuart_change_baudrate(struct hci_dev *hdev)
 {
 	struct btmtkuart_dev *bdev = hci_get_drvdata(hdev);
 	struct btmtk_hci_wmt_params wmt_params;
-	u32 baudrate;
+	__le32 baudrate;
 	u8 param;
 	int err;
 
Signature · 1 line
2.20.1
author maintainer
#146 lore raw

PATCH 146/405 HID: logitech-hidpp: use RAP instead of FAP to get the protocol version

In reply to Greg Kroah-Hartman

[ Upstream commit 096377525cdb8251e4656085efc988bdf733fb4c ]

According to the logitech_hidpp_2.0_specification_draft_2012-06-04.pdf doc:
https://lekensteyn.nl/files/logitech/logitech_hidpp_2.0_specification_draft_2012-06-04.pdf

We should use a register-access-protocol request using the short input / output report ids. This is necessary because 27MHz HID++ receivers have a max-packetsize on their HIP++ endpoint of 8, so they cannot support long reports. Using a feature-access-protocol request (which is always long or very-long) with these will cause a timeout error, followed by the hidpp driver treating the device as not being HID++ capable.

This commit fixes this by switching to using a rap request to get the protocol version.

Besides being tested with a (046d:c517) 27MHz receiver with various 27MHz keyboards and mice, this has also been tested to not cause regressions on a non-unifying dual-HID++ nano receiver (046d:c534) with k270 and m185 HID++-2.0 devices connected and on a unifying/dj receiver (046d:c52b) with a HID++-2.0 Logitech Rechargeable Touchpad T650.

---
 drivers/hid/hid-logitech-hidpp.c | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)
drivers/hid/hid-logitech-hidpp.c +13 -4
index 199cc256e9d9d..ffd30c7492df8 100644
@@ -836,13 +836,16 @@ static int hidpp_root_get_feature(struct hidpp_device *hidpp, u16 feature,
 
 static int hidpp_root_get_protocol_version(struct hidpp_device *hidpp)
 {
+	const u8 ping_byte = 0x5a;
+	u8 ping_data[3] = { 0, 0, ping_byte };
 	struct hidpp_report response;
 	int ret;
 
-	ret = hidpp_send_fap_command_sync(hidpp,
+	ret = hidpp_send_rap_command_sync(hidpp,
+			REPORT_ID_HIDPP_SHORT,
 			HIDPP_PAGE_ROOT_IDX,
 			CMD_ROOT_GET_PROTOCOL_VERSION,
-			NULL, 0, &response);
+			ping_data, sizeof(ping_data), &response);
 
 	if (ret == HIDPP_ERROR_INVALID_SUBID) {
 		hidpp->protocol_major = 1;
@@ -862,8 +865,14 @@ static int hidpp_root_get_protocol_version(struct hidpp_device *hidpp)
 	if (ret)
 		return ret;
 
-	hidpp->protocol_major = response.fap.params[0];
-	hidpp->protocol_minor = response.fap.params[1];
+	if (response.rap.params[2] != ping_byte) {
+		hid_err(hidpp->hid_dev, "%s: ping mismatch 0x%02x != 0x%02x\n",
+			__func__, response.rap.params[2], ping_byte);
+		return -EPROTO;
+	}
+
+	hidpp->protocol_major = response.rap.params[0];
+	hidpp->protocol_minor = response.rap.params[1];
 
 	return ret;
 }
Signature · 1 line
2.20.1
author maintainer
#147 lore raw

PATCH 147/405 pinctrl: pistachio: fix leaked of_node references

In reply to Greg Kroah-Hartman

[ Upstream commit 44a4455ac2c6b0981eace683a2b6eccf47689022 ]

The call to of_get_child_by_name returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./drivers/pinctrl/pinctrl-pistachio.c:1422:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 1360, but without a corresponding object release within this function.

---
 drivers/pinctrl/pinctrl-pistachio.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/pinctrl/pinctrl-pistachio.c +2 -0
index aa5f949ef219c..5b0678f310e52 100644
@@ -1367,6 +1367,7 @@ static int pistachio_gpio_register(struct pistachio_pinctrl *pctl)
 		if (!of_find_property(child, "gpio-controller", NULL)) {
 			dev_err(pctl->dev,
 				"No gpio-controller property for bank %u\n", i);
+			of_node_put(child);
 			ret = -ENODEV;
 			goto err;
 		}
@@ -1374,6 +1375,7 @@ static int pistachio_gpio_register(struct pistachio_pinctrl *pctl)
 		irq = irq_of_parse_and_map(child, 0);
 		if (irq < 0) {
 			dev_err(pctl->dev, "No IRQ for bank %u: %d\n", i, irq);
+			of_node_put(child);
 			ret = irq;
 			goto err;
 		}
Signature · 1 line
2.20.1
author maintainer
#148 lore raw

PATCH 148/405 pinctrl: st: fix leaked of_node references

In reply to Greg Kroah-Hartman

[ Upstream commit 483d70d73beaecab55882fcd2a357af72674e24c ]

The call to of_get_child_by_name returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./drivers/pinctrl/pinctrl-st.c:1188:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 1175, but without a corresponding object release within this function. ./drivers/pinctrl/pinctrl-st.c:1188:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 1175, but without a corresponding object release within this function. ./drivers/pinctrl/pinctrl-st.c:1199:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 1175, but without a corresponding object release within this function. ./drivers/pinctrl/pinctrl-st.c:1199:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 1175, but without a corresponding object release within this function.

---
 drivers/pinctrl/pinctrl-st.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)
drivers/pinctrl/pinctrl-st.c +10 -5
index e66af93f2cbf8..195b442a23434 100644
@@ -1170,7 +1170,7 @@ static int st_pctl_dt_parse_groups(struct device_node *np,
 	struct property *pp;
 	struct st_pinconf *conf;
 	struct device_node *pins;
-	int i = 0, npins = 0, nr_props;
+	int i = 0, npins = 0, nr_props, ret = 0;
 
 	pins = of_get_child_by_name(np, "st,pins");
 	if (!pins)
@@ -1185,7 +1185,8 @@ static int st_pctl_dt_parse_groups(struct device_node *np,
 			npins++;
 		} else {
 			pr_warn("Invalid st,pins in %pOFn node\n", np);
-			return -EINVAL;
+			ret = -EINVAL;
+			goto out_put_node;
 		}
 	}
 
@@ -1195,8 +1196,10 @@ static int st_pctl_dt_parse_groups(struct device_node *np,
 	grp->pin_conf = devm_kcalloc(info->dev,
 					npins, sizeof(*conf), GFP_KERNEL);
 
-	if (!grp->pins || !grp->pin_conf)
-		return -ENOMEM;
+	if (!grp->pins || !grp->pin_conf) {
+		ret = -ENOMEM;
+		goto out_put_node;
+	}
 
 	/* <bank offset mux direction rt_type rt_delay rt_clk> */
 	for_each_property_of_node(pins, pp) {
@@ -1229,9 +1232,11 @@ static int st_pctl_dt_parse_groups(struct device_node *np,
 		}
 		i++;
 	}
+
+out_put_node:
 	of_node_put(pins);
 
-	return 0;
+	return ret;
 }
 
 static int st_pctl_parse_functions(struct device_node *np,
Signature · 1 line
2.20.1
author maintainer
#149 lore raw

PATCH 149/405 pinctrl: samsung: fix leaked of_node references

In reply to Greg Kroah-Hartman

[ Upstream commit 44b9f86cd41db6c522effa5aec251d664a52fbc0 ]

The call to of_find_compatible_node returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./drivers/pinctrl/samsung/pinctrl-exynos-arm.c:76:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 66, but without a corresponding object release within this function. ./drivers/pinctrl/samsung/pinctrl-exynos-arm.c:82:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 66, but without a corresponding object release within this function.

---
 drivers/pinctrl/samsung/pinctrl-exynos-arm.c | 1 +
 1 file changed, 1 insertion(+)
drivers/pinctrl/samsung/pinctrl-exynos-arm.c +1 -0
index 44c6b753f692a..85ddf49a51885 100644
@@ -71,6 +71,7 @@ s5pv210_retention_init(struct samsung_pinctrl_drv_data *drvdata,
 	}
 
 	clk_base = of_iomap(np, 0);
+	of_node_put(np);
 	if (!clk_base) {
 		pr_err("%s: failed to map clock registers\n", __func__);
 		return ERR_PTR(-EINVAL);
Signature · 1 line
2.20.1
author maintainer
#150 lore raw

PATCH 150/405 clk: rockchip: undo several noc and special clocks as critical on rk3288

In reply to Greg Kroah-Hartman

[ Upstream commit f4033db5b84ebe4b32c25ba2ed65ab20b628996a ]

This is mostly a revert of commit 55bb6a633c33 ("clk: rockchip: mark noc and some special clk as critical on rk3288") except that we're keeping "pmu_hclk_otg0" as critical still.

NOTE: turning these clocks off doesn't seem to do a whole lot in terms of power savings (checking the power on the logic rail). It appears to save maybe 1-2mW. ...but still it seems like we should turn the clocks off if they aren't needed.

About "pmu_hclk_otg0" (the one clock from the original commit we're still keeping critical) from an email thread:

Quoted 12 lines

pmu ahb clock

Function: Clock to pmu module when hibernation and/or ADP is enabled. Must be greater than or equal to 30 MHz.

If the SOC design does not support hibernation/ADP function, only have hclk_otg, this clk can be switched according to the usage of otg. If the SOC design support hibernation/ADP, has two clocks, hclk_otg and pmu_hclk_otg0.
Hclk_otg belongs to the closed part of otg logic, which can be switched according to the use of otg.

Quoted 4 lines

pmu_hclk_otg0 belongs to the always on part.

As for whether pmu_hclk_otg0 can be turned off when otg is not in use, we have not tested. IC suggest make pmu_hclk_otg0 always on.

For the rest of the clocks:

atclk: No documentation about this clock other than that it goes to the CPU. CPU functions fine without it on. Maybe needed for JTAG?

jtag: Presumably this clock is only needed if you're debugging with JTAG. It doesn't seem like it makes sense to waste power for every rk3288 user. In any case to do JTAG you'd need private patches to adjust the pinctrl the mux the JTAG out anyway.

pclk_dbg, pclk_core_niu: On veyron Chromebooks we turn these two clocks on only during kernel panics in order to access some coresight registers. Since nothing in the upstream kernel does this we should be able to leave them off safely. Maybe also needed for JTAG?

hsicphy12m_xin12m: There is no indication of why this clock would need to be turned on for boards that don't use HSIC.

pclk_ddrupctl[0-1], pclk_publ0[0-1]: On veyron Chromebooks we turn these 4 clocks on only when doing DDR transitions and they are off otherwise. I see no reason why they'd need to be on in the upstream kernel which doesn't support DDRFreq.

---
 drivers/clk/rockchip/clk-rk3288.c | 13 ++++---------
 1 file changed, 4 insertions(+), 9 deletions(-)
drivers/clk/rockchip/clk-rk3288.c +4 -9
index 5a67b7869960e..f3bbcdfa88ead 100644
@@ -313,13 +313,13 @@ static struct rockchip_clk_branch rk3288_clk_branches[] __initdata = {
 	COMPOSITE_NOMUX(0, "aclk_core_mp", "armclk", CLK_IGNORE_UNUSED,
 			RK3288_CLKSEL_CON(0), 4, 4, DFLAGS | CLK_DIVIDER_READ_ONLY,
 			RK3288_CLKGATE_CON(12), 6, GFLAGS),
-	COMPOSITE_NOMUX(0, "atclk", "armclk", CLK_IGNORE_UNUSED,
+	COMPOSITE_NOMUX(0, "atclk", "armclk", 0,
 			RK3288_CLKSEL_CON(37), 4, 5, DFLAGS | CLK_DIVIDER_READ_ONLY,
 			RK3288_CLKGATE_CON(12), 7, GFLAGS),
 	COMPOSITE_NOMUX(0, "pclk_dbg_pre", "armclk", CLK_IGNORE_UNUSED,
 			RK3288_CLKSEL_CON(37), 9, 5, DFLAGS | CLK_DIVIDER_READ_ONLY,
 			RK3288_CLKGATE_CON(12), 8, GFLAGS),
-	GATE(0, "pclk_dbg", "pclk_dbg_pre", CLK_IGNORE_UNUSED,
+	GATE(0, "pclk_dbg", "pclk_dbg_pre", 0,
 			RK3288_CLKGATE_CON(12), 9, GFLAGS),
 	GATE(0, "cs_dbg", "pclk_dbg_pre", CLK_IGNORE_UNUSED,
 			RK3288_CLKGATE_CON(12), 10, GFLAGS),
@@ -647,7 +647,7 @@ static struct rockchip_clk_branch rk3288_clk_branches[] __initdata = {
 	INVERTER(SCLK_HSADC, "sclk_hsadc", "sclk_hsadc_out",
 			RK3288_CLKSEL_CON(22), 7, IFLAGS),
 
-	GATE(0, "jtag", "ext_jtag", CLK_IGNORE_UNUSED,
+	GATE(0, "jtag", "ext_jtag", 0,
 			RK3288_CLKGATE_CON(4), 14, GFLAGS),
 
 	COMPOSITE_NODIV(SCLK_USBPHY480M_SRC, "usbphy480m_src", mux_usbphy480m_p, 0,
@@ -656,7 +656,7 @@ static struct rockchip_clk_branch rk3288_clk_branches[] __initdata = {
 	COMPOSITE_NODIV(SCLK_HSICPHY480M, "sclk_hsicphy480m", mux_hsicphy480m_p, 0,
 			RK3288_CLKSEL_CON(29), 0, 2, MFLAGS,
 			RK3288_CLKGATE_CON(3), 6, GFLAGS),
-	GATE(0, "hsicphy12m_xin12m", "xin12m", CLK_IGNORE_UNUSED,
+	GATE(0, "hsicphy12m_xin12m", "xin12m", 0,
 			RK3288_CLKGATE_CON(13), 9, GFLAGS),
 	DIV(0, "hsicphy12m_usbphy", "sclk_hsicphy480m", 0,
 			RK3288_CLKSEL_CON(11), 8, 6, DFLAGS),
@@ -837,11 +837,6 @@ static const char *const rk3288_critical_clocks[] __initconst = {
 	"pclk_alive_niu",
 	"pclk_pd_pmu",
 	"pclk_pmu_niu",
-	"pclk_core_niu",
-	"pclk_ddrupctl0",
-	"pclk_publ0",
-	"pclk_ddrupctl1",
-	"pclk_publ1",
 	"pmu_hclk_otg0",
 };
 
Signature · 1 line
2.20.1
author maintainer
#151 lore raw

PATCH 151/405 perf/arm-cci: Remove broken race mitigation

In reply to Greg Kroah-Hartman

[ Upstream commit 0d2e2a82d4de298d006bf8eddc86829e3c7da820 ]

Uncore PMU drivers face an awkward cyclic dependency wherein:

- They have to pick a valid online CPU to associate with before registering the PMU device, since it will get exposed to userspace immediately.
- The PMU registration has to be be at least partly complete before hotplug events can be handled, since trying to migrate an uninitialised context would be bad.
- The hotplug handler has to be ready as soon as a CPU is chosen, lest it go offline without the user-visible cpumask value getting updated.

The arm-cci driver has tried to solve this by using get_cpu() to pick the current CPU and prevent it from disappearing while both registrations are performed, but that results in taking mutexes with preemption disabled, which makes certain configurations very unhappy:

[ 1.983337] BUG: sleeping function called from invalid context at kernel/locking/rtmutex.c:2004
[ 1.983340] in_atomic(): 1, irqs_disabled(): 0, pid: 1, name: swapper/0
[ 1.983342] Preemption disabled at:
[ 1.983353] [<ffffff80089801f4>] cci_pmu_probe+0x1dc/0x488
[ 1.983360] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.18.20-rt8-yocto-preempt-rt #1
[ 1.983362] Hardware name: ZynqMP ZCU102 Rev1.0 (DT)
[ 1.983364] Call trace:
[ 1.983369] dump_backtrace+0x0/0x158
[ 1.983372] show_stack+0x24/0x30
[ 1.983378] dump_stack+0x80/0xa4
[ 1.983383] ___might_sleep+0x138/0x160
[ 1.983386] __might_sleep+0x58/0x90
[ 1.983391] __rt_mutex_lock_state+0x30/0xc0
[ 1.983395] _mutex_lock+0x24/0x30
[ 1.983400] perf_pmu_register+0x2c/0x388
[ 1.983404] cci_pmu_probe+0x2bc/0x488
[ 1.983409] platform_drv_probe+0x58/0xa8

It is not feasible to resolve all the possible races outside of the perf core itself, so address the immediate bug by following the example of nearly every other PMU driver and not even trying to do so. Registering the hotplug notifier first should minimise the window in which things can go wrong, so that's about as much as we can reasonably do here. This also revealed an additional race in assigning the global pointer too late relative to the hotplug notifier, which gets fixed in the process.

---
 drivers/perf/arm-cci.c | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)
drivers/perf/arm-cci.c +12 -9
index bfd03e0233084..8f8606b9bc9ee 100644
@@ -1684,21 +1684,24 @@ static int cci_pmu_probe(struct platform_device *pdev)
 	raw_spin_lock_init(&cci_pmu->hw_events.pmu_lock);
 	mutex_init(&cci_pmu->reserve_mutex);
 	atomic_set(&cci_pmu->active_events, 0);
-	cci_pmu->cpu = get_cpu();
-
-	ret = cci_pmu_init(cci_pmu, pdev);
-	if (ret) {
-		put_cpu();
-		return ret;
-	}
 
+	cci_pmu->cpu = raw_smp_processor_id();
+	g_cci_pmu = cci_pmu;
 	cpuhp_setup_state_nocalls(CPUHP_AP_PERF_ARM_CCI_ONLINE,
 				  "perf/arm/cci:online", NULL,
 				  cci_pmu_offline_cpu);
-	put_cpu();
-	g_cci_pmu = cci_pmu;
+
+	ret = cci_pmu_init(cci_pmu, pdev);
+	if (ret)
+		goto error_pmu_init;
+
 	pr_info("ARM %s PMU driver probed", cci_pmu->model->name);
 	return 0;
+
+error_pmu_init:
+	cpuhp_remove_state(CPUHP_AP_PERF_ARM_CCI_ONLINE);
+	g_cci_pmu = NULL;
+	return ret;
 }
 
 static int cci_pmu_remove(struct platform_device *pdev)
Signature · 1 line
2.20.1
author maintainer
#152 lore raw

PATCH 152/405 dmaengine: at_xdmac: remove BUG_ON macro in tasklet

In reply to Greg Kroah-Hartman

[ Upstream commit e2c114c06da2d9ffad5b16690abf008d6696f689 ]

Even if this case shouldn't happen when controller is properly programmed, it's still better to avoid dumping a kernel Oops for this.
As the sequence may happen only for debugging purposes, log the error and just finish the tasklet call.

---
 drivers/dma/at_xdmac.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/dma/at_xdmac.c +5 -1
index fe69dccfa0c05..37a2694204351 100644
@@ -1606,7 +1606,11 @@ static void at_xdmac_tasklet(unsigned long data)
 					struct at_xdmac_desc,
 					xfer_node);
 		dev_vdbg(chan2dev(&atchan->chan), "%s: desc 0x%p\n", __func__, desc);
-		BUG_ON(!desc->active_xfer);
+		if (!desc->active_xfer) {
+			dev_err(chan2dev(&atchan->chan), "Xfer not active: exiting");
+			spin_unlock_bh(&atchan->lock);
+			return;
+		}
 
 		txd = &desc->tx_dma_desc;
 
Signature · 1 line
2.20.1
author maintainer
#153 lore raw

PATCH 153/405 media: coda: clear error return value before picture run

In reply to Greg Kroah-Hartman

[ Upstream commit bbeefa7357a648afe70e7183914c87c3878d528d ]

The error return value is not written by some firmware codecs, such as MPEG-2 decode on CodaHx4. Clear the error return value before starting the picture run to avoid misinterpreting unrelated values returned by sequence initialization as error return value.

---
 drivers/media/platform/coda/coda-bit.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/media/platform/coda/coda-bit.c +3 -0
index b4f396c2e72c7..eaa86737fa04e 100644
@@ -2010,6 +2010,9 @@ static int coda_prepare_decode(struct coda_ctx *ctx)
 	/* Clear decode success flag */
 	coda_write(dev, 0, CODA_RET_DEC_PIC_SUCCESS);
 
+	/* Clear error return value */
+	coda_write(dev, 0, CODA_RET_DEC_PIC_ERR_MB);
+
 	trace_coda_dec_pic_run(ctx, meta);
 
 	coda_command_async(ctx, CODA_COMMAND_PIC_RUN);
Signature · 1 line
2.20.1
author maintainer
#154 lore raw

PATCH 154/405 media: ov6650: Move v4l2_clk_get() to ov6650_video_probe() helper

In reply to Greg Kroah-Hartman

[ Upstream commit ccdd85d518d8b9320ace1d87271f0ba2175f21fa ]

In preparation for adding asynchronous subdevice support to the driver, don't acquire v4l2_clk from the driver .probe() callback as that may fail if the clock is provided by a bridge driver which may be not yet initialized. Move the v4l2_clk_get() to ov6650_video_probe() helper which is going to be converted to v4l2_subdev_internal_ops.registered() callback, executed only when the bridge driver is ready.

---
 drivers/media/i2c/ov6650.c | 25 ++++++++++++++-----------
 1 file changed, 14 insertions(+), 11 deletions(-)
drivers/media/i2c/ov6650.c +14 -11
index f9359b11fa5cb..de7d9790f0542 100644
@@ -810,9 +810,16 @@ static int ov6650_video_probe(struct i2c_client *client)
 	u8		pidh, pidl, midh, midl;
 	int		ret;
 
+	priv->clk = v4l2_clk_get(&client->dev, NULL);
+	if (IS_ERR(priv->clk)) {
+		ret = PTR_ERR(priv->clk);
+		dev_err(&client->dev, "v4l2_clk request err: %d\n", ret);
+		return ret;
+	}
+
 	ret = ov6650_s_power(&priv->subdev, 1);
 	if (ret < 0)
-		return ret;
+		goto eclkput;
 
 	msleep(20);
 
@@ -849,6 +856,11 @@ static int ov6650_video_probe(struct i2c_client *client)
 
 done:
 	ov6650_s_power(&priv->subdev, 0);
+	if (!ret)
+		return 0;
+eclkput:
+	v4l2_clk_put(priv->clk);
+
 	return ret;
 }
 
@@ -991,18 +1003,9 @@ static int ov6650_probe(struct i2c_client *client,
 	priv->code	  = MEDIA_BUS_FMT_YUYV8_2X8;
 	priv->colorspace  = V4L2_COLORSPACE_JPEG;
 
-	priv->clk = v4l2_clk_get(&client->dev, NULL);
-	if (IS_ERR(priv->clk)) {
-		ret = PTR_ERR(priv->clk);
-		goto eclkget;
-	}
-
 	ret = ov6650_video_probe(client);
-	if (ret) {
-		v4l2_clk_put(priv->clk);
-eclkget:
+	if (ret)
 		v4l2_ctrl_handler_free(&priv->hdl);
-	}
 
 	return ret;
 }
Signature · 1 line
2.20.1
author maintainer
#155 lore raw

PATCH 155/405 media: au0828: stop video streaming only when last user stops

In reply to Greg Kroah-Hartman

[ Upstream commit f604f0f5afb88045944567f604409951b5eb6af8 ]

If the application was streaming from both videoX and vbiX, and streaming from videoX was stopped, then the vbi streaming also stopped.

The cause being that stop_streaming for video stopped the subdevs as well, instead of only doing that if dev->streaming_users reached 0.

au0828_stop_vbi_streaming was also wrong since it didn't stop the subdevs at all when dev->streaming_users reached 0.

---
 drivers/media/usb/au0828/au0828-video.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)
drivers/media/usb/au0828/au0828-video.c +8 -5
index 7876c897cc1d6..ad2b1b7ecea4d 100644
@@ -839,9 +839,9 @@ int au0828_start_analog_streaming(struct vb2_queue *vq, unsigned int count)
 			return rc;
 		}
 
+		v4l2_device_call_all(&dev->v4l2_dev, 0, video, s_stream, 1);
+
 		if (vq->type == V4L2_BUF_TYPE_VIDEO_CAPTURE) {
-			v4l2_device_call_all(&dev->v4l2_dev, 0, video,
-						s_stream, 1);
 			dev->vid_timeout_running = 1;
 			mod_timer(&dev->vid_timeout, jiffies + (HZ / 10));
 		} else if (vq->type == V4L2_BUF_TYPE_VBI_CAPTURE) {
@@ -861,10 +861,11 @@ static void au0828_stop_streaming(struct vb2_queue *vq)
 
 	dprintk(1, "au0828_stop_streaming called %d\n", dev->streaming_users);
 
-	if (dev->streaming_users-- == 1)
+	if (dev->streaming_users-- == 1) {
 		au0828_uninit_isoc(dev);
+		v4l2_device_call_all(&dev->v4l2_dev, 0, video, s_stream, 0);
+	}
 
-	v4l2_device_call_all(&dev->v4l2_dev, 0, video, s_stream, 0);
 	dev->vid_timeout_running = 0;
 	del_timer_sync(&dev->vid_timeout);
 
@@ -893,8 +894,10 @@ void au0828_stop_vbi_streaming(struct vb2_queue *vq)
 	dprintk(1, "au0828_stop_vbi_streaming called %d\n",
 		dev->streaming_users);
 
-	if (dev->streaming_users-- == 1)
+	if (dev->streaming_users-- == 1) {
 		au0828_uninit_isoc(dev);
+		v4l2_device_call_all(&dev->v4l2_dev, 0, video, s_stream, 0);
+	}
 
 	spin_lock_irqsave(&dev->slock, flags);
 	if (dev->isoc_ctl.vbi_buf != NULL) {
Signature · 1 line
2.20.1
author maintainer
#156 lore raw

PATCH 156/405 media: ov2659: make S_FMT succeed even if requested format doesnt match

In reply to Greg Kroah-Hartman

[ Upstream commit bccb89cf9cd07a0690d519696a00c00a973b3fe4 ]

This driver returns an error if unsupported media bus pixel code is requested by VIDIOC_SUBDEV_S_FMT.

But according to Documentation/media/uapi/v4l/vidioc-subdev-g-fmt.rst,

Drivers must not return an error solely because the requested format doesn't match the device capabilities. They must instead modify the format to match what the hardware can provide.

So select default format code and return success in that case.

This is detected by v4l2-compliance.

---
 drivers/media/i2c/ov2659.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/media/i2c/ov2659.c +4 -2
index 799acce803fe5..a1e9a980a4459 100644
@@ -1117,8 +1117,10 @@ static int ov2659_set_fmt(struct v4l2_subdev *sd,
 		if (ov2659_formats[index].code == mf->code)
 			break;
 
-	if (index < 0)
-		return -EINVAL;
+	if (index < 0) {
+		index = 0;
+		mf->code = ov2659_formats[index].code;
+	}
 
 	mf->colorspace = V4L2_COLORSPACE_SRGB;
 	mf->field = V4L2_FIELD_NONE;
Signature · 1 line
2.20.1
author maintainer
#157 lore raw

PATCH 157/405 audit: fix a memory leak bug

In reply to Greg Kroah-Hartman

[ Upstream commit 70c4cf17e445264453bc5323db3e50aa0ac9e81f ]

In audit_rule_change(), audit_data_to_entry() is firstly invoked to translate the payload data to the kernel's rule representation. In audit_data_to_entry(), depending on the audit field type, an audit tree may be created in audit_make_tree(), which eventually invokes kmalloc() to allocate the tree. Since this tree is a temporary tree, it will be then freed in the following execution, e.g., audit_add_rule() if the message type is AUDIT_ADD_RULE or audit_del_rule() if the message type is AUDIT_DEL_RULE. However, if the message type is neither AUDIT_ADD_RULE nor AUDIT_DEL_RULE, i.e., the default case of the switch statement, this temporary tree is not freed.

To fix this issue, only allocate the tree when the type is AUDIT_ADD_RULE or AUDIT_DEL_RULE.

---
 kernel/auditfilter.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)
kernel/auditfilter.c +7 -5
index 63f8b3f26fab4..3ac71c4fda49a 100644
@@ -1114,22 +1114,24 @@ int audit_rule_change(int type, int seq, void *data, size_t datasz)
 	int err = 0;
 	struct audit_entry *entry;
 
-	entry = audit_data_to_entry(data, datasz);
-	if (IS_ERR(entry))
-		return PTR_ERR(entry);
-
 	switch (type) {
 	case AUDIT_ADD_RULE:
+		entry = audit_data_to_entry(data, datasz);
+		if (IS_ERR(entry))
+			return PTR_ERR(entry);
 		err = audit_add_rule(entry);
 		audit_log_rule_change("add_rule", &entry->rule, !err);
 		break;
 	case AUDIT_DEL_RULE:
+		entry = audit_data_to_entry(data, datasz);
+		if (IS_ERR(entry))
+			return PTR_ERR(entry);
 		err = audit_del_rule(entry);
 		audit_log_rule_change("remove_rule", &entry->rule, !err);
 		break;
 	default:
-		err = -EINVAL;
 		WARN_ON(1);
+		return -EINVAL;
 	}
 
 	if (err || type == AUDIT_DEL_RULE) {
Signature · 1 line
2.20.1
author maintainer
#158 lore raw

PATCH 158/405 media: stm32-dcmi: fix crash when subdev do not expose any formats

In reply to Greg Kroah-Hartman

[ Upstream commit 33dfeb62e23c31619d2197850f7e8b50e8cc5466 ]

Do not access sd_formats[] if num_of_sd_formats is zero, ie subdev sensor didn't expose any formats.

---
 drivers/media/platform/stm32/stm32-dcmi.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/media/platform/stm32/stm32-dcmi.c +6 -0
index a1f0801081ba9..922855b6025c7 100644
@@ -811,6 +811,9 @@ static int dcmi_try_fmt(struct stm32_dcmi *dcmi, struct v4l2_format *f,
 
 	sd_fmt = find_format_by_fourcc(dcmi, pix->pixelformat);
 	if (!sd_fmt) {
+		if (!dcmi->num_of_sd_formats)
+			return -ENODATA;
+
 		sd_fmt = dcmi->sd_formats[dcmi->num_of_sd_formats - 1];
 		pix->pixelformat = sd_fmt->fourcc;
 	}
@@ -989,6 +992,9 @@ static int dcmi_set_sensor_format(struct stm32_dcmi *dcmi,
 
 	sd_fmt = find_format_by_fourcc(dcmi, pix->pixelformat);
 	if (!sd_fmt) {
+		if (!dcmi->num_of_sd_formats)
+			return -ENODATA;
+
 		sd_fmt = dcmi->sd_formats[dcmi->num_of_sd_formats - 1];
 		pix->pixelformat = sd_fmt->fourcc;
 	}
Signature · 1 line
2.20.1
author maintainer
#159 lore raw

PATCH 159/405 media: au0828: Fix NULL pointer dereference in au0828_analog_stream_enable()

In reply to Greg Kroah-Hartman

[ Upstream commit 898bc40bfcc26abb6e06e960d6d4754c36c58b50 ]

Fix au0828_analog_stream_enable() to check if device is in the right state first. When unbind happens while bind is in progress, usbdev pointer could be invalid in au0828_analog_stream_enable() and a call to usb_ifnum_to_if() will result in the null pointer dereference.

This problem is found with the new media_dev_allocator.sh test.

kernel: [ 590.359623] BUG: unable to handle kernel NULL pointer dereference at 00000000000004e8
kernel: [ 590.359627] #PF error: [normal kernel read fault]
kernel: [ 590.359629] PGD 0 P4D 0
kernel: [ 590.359632] Oops: 0000 [#1] SMP PTI
kernel: [ 590.359634] CPU: 3 PID: 1458 Comm: v4l_id Not tainted 5.1.0-rc2+ #30
kernel: [ 590.359636] Hardware name: Dell Inc. OptiPlex 7 90/0HY9JP, BIOS A18 09/24/2013
kernel: [ 590.359641] RIP: 0010:usb_ifnum_to_if+0x6/0x60
kernel: [ 590.359643] Code: 5d 41 5e 41 5f 5d c3 48 83 c4
10 b8 fa ff ff ff 5b 41 5c 41 5d 41 5e 41 5f 5d c3 b8 fa ff ff ff c3 0f 1f 00 6
6 66 66 66 90 55 <48> 8b 97 e8 04 00 00 48 89 e5 48 85 d2 74 41 0f b6 4a 04 84 c
9 74
kernel: [ 590.359645] RSP: 0018:ffffad3cc3c1fc00 EFLAGS: 00010246
kernel: [ 590.359646] RAX: 0000000000000000 RBX: ffff8ded b1f3c000 RCX: 1f377e4500000000
kernel: [ 590.359648] RDX: ffff8dedfa3a6b50 RSI: 00000000 00000000 RDI: 0000000000000000
kernel: [ 590.359649] RBP: ffffad3cc3c1fc28 R08: 00000000 8574acc2 R09: ffff8dedfa3a6b50
kernel: [ 590.359650] R10: 0000000000000001 R11: 00000000 00000000 R12: 0000000000000000
kernel: [ 590.359652] R13: ffff8dedb1f3f0f0 R14: ffffffff adcf7ec0 R15: 0000000000000000
kernel: [ 590.359654] FS: 00007f7917198540(0000) GS:ffff 8dee258c0000(0000) knlGS:0000000000000000
kernel: [ 590.359655] CS: 0010 DS: 0000 ES: 0000 CR0: 00 00000080050033
kernel: [ 590.359657] CR2: 00000000000004e8 CR3: 00000001 a388e002 CR4: 00000000000606e0
kernel: [ 590.359658] Call Trace:
kernel: [ 590.359664] ? au0828_analog_stream_enable+0x2c/0x180
kernel: [ 590.359666] au0828_v4l2_open+0xa4/0x110
kernel: [ 590.359670] v4l2_open+0x8b/0x120
kernel: [ 590.359674] chrdev_open+0xa6/0x1c0
kernel: [ 590.359676] ? cdev_put.part.3+0x20/0x20
kernel: [ 590.359678] do_dentry_open+0x1f6/0x360
kernel: [ 590.359681] vfs_open+0x2f/0x40
kernel: [ 590.359684] path_openat+0x299/0xc20
kernel: [ 590.359688] do_filp_open+0x9b/0x110
kernel: [ 590.359695] ? _raw_spin_unlock+0x27/0x40
kernel: [ 590.359697] ? __alloc_fd+0xb2/0x160
kernel: [ 590.359700] do_sys_open+0x1ba/0x260
kernel: [ 590.359702] ? do_sys_open+0x1ba/0x260
kernel: [ 590.359712] __x64_sys_openat+0x20/0x30
kernel: [ 590.359715] do_syscall_64+0x5a/0x120
kernel: [ 590.359718] entry_SYSCALL_64_after_hwframe+0x44/0xa9

---
 drivers/media/usb/au0828/au0828-video.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/media/usb/au0828/au0828-video.c +3 -0
index ad2b1b7ecea4d..222723d946e4c 100644
@@ -758,6 +758,9 @@ static int au0828_analog_stream_enable(struct au0828_dev *d)
 
 	dprintk(1, "au0828_analog_stream_enable called\n");
 
+	if (test_bit(DEV_DISCONNECTED, &d->dev_state))
+		return -ENODEV;
+
 	iface = usb_ifnum_to_if(d->usbdev, 0);
 	if (iface && iface->cur_altsetting->desc.bAlternateSetting != 5) {
 		dprintk(1, "Changing intf#0 to alt 5\n");
Signature · 1 line
2.20.1
author maintainer
#160 lore raw

PATCH 160/405 media: pvrusb2: Prevent a buffer overflow

In reply to Greg Kroah-Hartman

[ Upstream commit c1ced46c7b49ad7bc064e68d966e0ad303f917fb ]

The ctrl_check_input() function is called from pvr2_ctrl_range_check(). It's supposed to validate user supplied input and return true or false depending on whether the input is valid or not. The problem is that negative shifts or shifts greater than 31 are undefined in C. In practice with GCC they result in shift wrapping so this function returns true for some inputs which are not valid and this could result in a buffer overflow:

drivers/media/usb/pvrusb2/pvrusb2-ctrl.c:205 pvr2_ctrl_get_valname()
warn: uncapped user index 'names[val]'

The cptr->hdw->input_allowed_mask mask is configured in pvr2_hdw_create() and the highest valid bit is BIT(4).

Fixes 7fb20fa38caa ("V4L/DVB (7299): pvrusb2: Improve logic which handles input choice availability")
---
 drivers/media/usb/pvrusb2/pvrusb2-hdw.c | 2 ++
 drivers/media/usb/pvrusb2/pvrusb2-hdw.h | 1 +
 2 files changed, 3 insertions(+)
drivers/media/usb/pvrusb2/pvrusb2-hdw.c +2 -0
index 446a999dd2ce1..2bab4713bc5b9 100644
@@ -666,6 +666,8 @@ static int ctrl_get_input(struct pvr2_ctrl *cptr,int *vp)
 
 static int ctrl_check_input(struct pvr2_ctrl *cptr,int v)
 {
+	if (v < 0 || v > PVR2_CVAL_INPUT_MAX)
+		return 0;
 	return ((1 << v) & cptr->hdw->input_allowed_mask) != 0;
 }
 
drivers/media/usb/pvrusb2/pvrusb2-hdw.h +1 -0
index 25648add77e58..bd2b7a67b7322 100644
@@ -50,6 +50,7 @@
 #define PVR2_CVAL_INPUT_COMPOSITE 2
 #define PVR2_CVAL_INPUT_SVIDEO 3
 #define PVR2_CVAL_INPUT_RADIO 4
+#define PVR2_CVAL_INPUT_MAX PVR2_CVAL_INPUT_RADIO
 
 enum pvr2_config {
 	pvr2_config_empty,    /* No configuration */
Signature · 1 line
2.20.1
author maintainer
#161 lore raw

PATCH 161/405 iio: adc: stm32-dfsdm: fix unmet direct dependencies detected

In reply to Greg Kroah-Hartman

[ Upstream commit ba7ecfe43d6bf12e2aa76705c45f7d187ae3d7c0 ]

This fixes unmet direct dependencies seen when CONFIG_STM32_DFSDM_ADC is selected:

WARNING: unmet direct dependencies detected for IIO_BUFFER_HW_CONSUMER
Depends on [n]: IIO [=y] && IIO_BUFFER [=n]
Selected by [y]:
- STM32_DFSDM_ADC [=y] && IIO [=y] && (ARCH_STM32 [=y] && OF [=y] ||

    COMPILE_TEST [=n])
Fixes e2e6771c6462 ("IIO: ADC: add STM32 DFSDM sigma delta ADC support")
---
 drivers/iio/adc/Kconfig | 1 +
 1 file changed, 1 insertion(+)
drivers/iio/adc/Kconfig +1 -0
index 76db6e5cc2961..9ca21a8dfcd71 100644
@@ -809,6 +809,7 @@ config STM32_DFSDM_ADC
 	depends on (ARCH_STM32 && OF) || COMPILE_TEST
 	select STM32_DFSDM_CORE
 	select REGMAP_MMIO
+	select IIO_BUFFER
 	select IIO_BUFFER_HW_CONSUMER
 	help
 	  Select this option to support ADCSigma delta modulator for
Signature · 1 line
2.20.1
author maintainer
#162 lore raw

PATCH 162/405 block: fix use-after-free on gendisk

In reply to Greg Kroah-Hartman

[ Upstream commit 2c88e3c7ec32d7a40cc7c9b4a487cf90e4671bdd ]

commit 2da78092dda "block: Fix dev_t minor allocation lifetime" specifically moved blk_free_devt(dev->devt) call to part_release() to avoid reallocating device number before the device is fully shutdown.

However, it can cause use-after-free on gendisk in get_gendisk(). We use md device as example to show the race scenes:

Process1 Worker Process2
md_free
blkdev_open
del_gendisk
add delete_partition_work_fn() to wq
__blkdev_get get_gendisk
put_disk
disk_release
kfree(disk)
find part from ext_devt_idr get_disk_and_module(disk) cause use after free

delete_partition_work_fn put_device(part) part_release
remove part from ext_devt_idr

Before <devt, hd_struct pointer> is removed from ext_devt_idr by delete_partition_work_fn(), we can find the devt and then access gendisk by hd_struct pointer. But, if we access the gendisk after it have been freed, it can cause in use-after-freeon gendisk in get_gendisk().

We fix this by adding a new helper blk_invalidate_devt() in delete_partition() and del_gendisk(). It replaces hd_struct pointer in idr with value 'NULL', and deletes the entry from idr in part_release() as we do now.

Thanks to Jan Kara for providing the solution and more clear comments for the code.

Fixes 2da78092dda1 ("block: Fix dev_t minor allocation lifetime")
Cc Al Viro <[email protected]>
---
 block/genhd.c             | 19 +++++++++++++++++++
 block/partition-generic.c |  7 +++++++
 include/linux/genhd.h     |  1 +
 3 files changed, 27 insertions(+)
block/genhd.c +19 -0
index 703267865f14d..d8dff0b21f7d1 100644
@@ -531,6 +531,18 @@ void blk_free_devt(dev_t devt)
 	}
 }
 
+/**
+ *	We invalidate devt by assigning NULL pointer for devt in idr.
+ */
+void blk_invalidate_devt(dev_t devt)
+{
+	if (MAJOR(devt) == BLOCK_EXT_MAJOR) {
+		spin_lock_bh(&ext_devt_lock);
+		idr_replace(&ext_devt_idr, NULL, blk_mangle_minor(MINOR(devt)));
+		spin_unlock_bh(&ext_devt_lock);
+	}
+}
+
 static char *bdevt_str(dev_t devt, char *buf)
 {
 	if (MAJOR(devt) <= 0xff && MINOR(devt) <= 0xff) {
@@ -793,6 +805,13 @@ void del_gendisk(struct gendisk *disk)
 
 	if (!(disk->flags & GENHD_FL_HIDDEN))
 		blk_unregister_region(disk_devt(disk), disk->minors);
+	/*
+	 * Remove gendisk pointer from idr so that it cannot be looked up
+	 * while RCU period before freeing gendisk is running to prevent
+	 * use-after-free issues. Note that the device number stays
+	 * "in-use" until we really free the gendisk.
+	 */
+	blk_invalidate_devt(disk_devt(disk));
 
 	kobject_put(disk->part0.holder_dir);
 	kobject_put(disk->slave_dir);
block/partition-generic.c +7 -0
index 8e596a8dff321..aee643ce13d15 100644
@@ -285,6 +285,13 @@ void delete_partition(struct gendisk *disk, int partno)
 	kobject_put(part->holder_dir);
 	device_del(part_to_dev(part));
 
+	/*
+	 * Remove gendisk pointer from idr so that it cannot be looked up
+	 * while RCU period before freeing gendisk is running to prevent
+	 * use-after-free issues. Note that the device number stays
+	 * "in-use" until we really free the gendisk.
+	 */
+	blk_invalidate_devt(part_devt(part));
 	hd_struct_kill(part);
 }
 
include/linux/genhd.h +1 -0
index 06c0fd594097d..69db1affedb0b 100644
@@ -610,6 +610,7 @@ struct unixware_disklabel {
 
 extern int blk_alloc_devt(struct hd_struct *part, dev_t *devt);
 extern void blk_free_devt(dev_t devt);
+extern void blk_invalidate_devt(dev_t devt);
 extern dev_t blk_lookup_devt(const char *name, int partno);
 extern char *disk_name (struct gendisk *hd, int partno, char *buf);
 
Signature · 1 line
2.20.1
author maintainer
#163 lore raw

PATCH 163/405 powerpc/numa: improve control of topology updates

In reply to Greg Kroah-Hartman

[ Upstream commit 2d4d9b308f8f8dec68f6dbbff18c68ec7c6bd26f ]

When booted with "topology_updates=no", or when "off" is written to /proc/powerpc/topology_updates, NUMA reassignments are inhibited for PRRN and VPHN events. However, migration and suspend unconditionally re-enable reassignments via start_topology_update(). This is incoherent.

Check the topology_updates_enabled flag in start/stop_topology_update() so that callers of those APIs need not be aware of whether reassignments are enabled. This allows the administrative decision on reassignments to remain in force across migrations and suspensions.

---
 arch/powerpc/mm/numa.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)
arch/powerpc/mm/numa.c +12 -6
index f976676004ad0..48c9a97eb2c33 100644
@@ -1498,6 +1498,9 @@ int start_topology_update(void)
 {
 	int rc = 0;
 
+	if (!topology_updates_enabled)
+		return 0;
+
 	if (firmware_has_feature(FW_FEATURE_PRRN)) {
 		if (!prrn_enabled) {
 			prrn_enabled = 1;
@@ -1531,6 +1534,9 @@ int stop_topology_update(void)
 {
 	int rc = 0;
 
+	if (!topology_updates_enabled)
+		return 0;
+
 	if (prrn_enabled) {
 		prrn_enabled = 0;
 #ifdef CONFIG_SMP
@@ -1588,11 +1594,13 @@ static ssize_t topology_write(struct file *file, const char __user *buf,
 
 	kbuf[read_len] = '\0';
 
-	if (!strncmp(kbuf, "on", 2))
+	if (!strncmp(kbuf, "on", 2)) {
+		topology_updates_enabled = true;
 		start_topology_update();
-	else if (!strncmp(kbuf, "off", 3))
+	} else if (!strncmp(kbuf, "off", 3)) {
 		stop_topology_update();
-	else
+		topology_updates_enabled = false;
+	} else
 		return -EINVAL;
 
 	return count;
@@ -1607,9 +1615,7 @@ static const struct file_operations topology_ops = {
 
 static int topology_update_init(void)
 {
-	/* Do not poll for changes if disabled at boot */
-	if (topology_updates_enabled)
-		start_topology_update();
+	start_topology_update();
 
 	if (vphn_enabled)
 		topology_schedule_update();
Signature · 1 line
2.20.1
author maintainer
#164 lore raw

PATCH 164/405 powerpc/64: Fix booting large kernels with STRICT_KERNEL_RWX

In reply to Greg Kroah-Hartman

[ Upstream commit 56c46bba9bbfe229b4472a5be313c44c5b714a39 ]

With STRICT_KERNEL_RWX enabled anything marked __init is placed at a 16M boundary. This is necessary so that it can be repurposed later with different permissions. However, in kernels with text larger than 16M, this pushes early_setup past 32M, incapable of being reached by the branch instruction.

Fix this by setting the CTR and branching there instead.

Fixes 1e0fc9d1eb2b ("powerpc/Kconfig: Enable STRICT_KERNEL_RWX for some configs")
[mpe: Fix it to work on BE by using DOTSYM()]
---
 arch/powerpc/kernel/head_64.S | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
arch/powerpc/kernel/head_64.S +3 -1
index 3fad8d499767c..5321a11c28358 100644
@@ -968,7 +968,9 @@ start_here_multiplatform:
 
 	/* Restore parameters passed from prom_init/kexec */
 	mr	r3,r31
-	bl	early_setup		/* also sets r13 and SPRG_PACA */
+	LOAD_REG_ADDR(r12, DOTSYM(early_setup))
+	mtctr	r12
+	bctrl		/* also sets r13 and SPRG_PACA */
 
 	LOAD_REG_ADDR(r3, start_here_common)
 	ld	r4,PACAKMSR(r13)
Signature · 1 line
2.20.1
author maintainer
#165 lore raw

PATCH 165/405 random: fix CRNG initialization when random.trust_cpu=1

In reply to Greg Kroah-Hartman

[ Upstream commit fe6f1a6a8eedc1aa538fee0baa612b6a59639cf8 ]

When the system boots with random.trust_cpu=1 it doesn't initialize the per-NUMA CRNGs because it skips the rest of the CRNG startup code. This means that the code from 1e7f583af67b ("random: make /dev/urandom scalable for silly userspace programs") is not used when random.trust_cpu=1.

crash> dmesg | grep random:

[    0.000000] random: get_random_bytes called from start_kernel+0x94/0x530 with crng_init=0
[    0.314029] random: crng done (trusting CPU's manufacturer)

crash> print crng_node_pool
$6 = (struct crng_state **) 0x0

After adding the missing call to numa_crng_init() the per-NUMA CRNGs are initialized again:

crash> dmesg | grep random:

[    0.000000] random: get_random_bytes called from start_kernel+0x94/0x530 with crng_init=0
[    0.314031] random: crng done (trusting CPU's manufacturer)

crash> print crng_node_pool
$1 = (struct crng_state **) 0xffff9a915f4014a0

The call to invalidate_batched_entropy() was also missing. This is important for architectures like PPC and S390 which only have the arch_get_random_seed_* functions.

Fixes 39a8883a2b98 ("random: add a config option to trust the CPU's hwrng")
---
 drivers/char/random.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/char/random.c +4 -1
index 38c6d1af6d1c0..d4d45ccfeefc0 100644
@@ -777,6 +777,7 @@ static struct crng_state **crng_node_pool __read_mostly;
 #endif
 
 static void invalidate_batched_entropy(void);
+static void numa_crng_init(void);
 
 static bool trust_cpu __ro_after_init = IS_ENABLED(CONFIG_RANDOM_TRUST_CPU);
 static int __init parse_trust_cpu(char *arg)
@@ -805,7 +806,9 @@ static void crng_initialize(struct crng_state *crng)
 		}
 		crng->state[i] ^= rv;
 	}
-	if (trust_cpu && arch_init) {
+	if (trust_cpu && arch_init && crng == &primary_crng) {
+		invalidate_batched_entropy();
+		numa_crng_init();
 		crng_init = 2;
 		pr_notice("random: crng done (trusting CPU's manufacturer)\n");
 	}
Signature · 1 line
2.20.1
author maintainer
#166 lore raw

PATCH 166/405 random: add a spinlock_t to struct batched_entropy

In reply to Greg Kroah-Hartman

[ Upstream commit b7d5dc21072cda7124d13eae2aefb7343ef94197 ]

The per-CPU variable batched_entropy_uXX is protected by get_cpu_var(). This is just a preempt_disable() which ensures that the variable is only from the local CPU. It does not protect against users on the same CPU from another context. It is possible that a preemptible context reads slot 0 and then an interrupt occurs and the same value is read again.

The above scenario is confirmed by lockdep if we add a spinlock:

| ================================
| WARNING: inconsistent lock state
| 5.1.0-rc3+ #42 Not tainted
| --------------------------------
| inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage.
| ksoftirqd/9/56 [HC0[0]:SC1[1]:HE0:SE0] takes:
| (____ptrval____) (batched_entropy_u32.lock){+.?.}, at: get_random_u32+0x3e/0xe0
| {SOFTIRQ-ON-W} state was registered at:
|   _raw_spin_lock+0x2a/0x40
|   get_random_u32+0x3e/0xe0
|   new_slab+0x15c/0x7b0
|   ___slab_alloc+0x492/0x620
|   __slab_alloc.isra.73+0x53/0xa0
|   kmem_cache_alloc_node+0xaf/0x2a0
|   copy_process.part.41+0x1e1/0x2370
|   _do_fork+0xdb/0x6d0
|   kernel_thread+0x20/0x30
|   kthreadd+0x1ba/0x220
|   ret_from_fork+0x3a/0x50
…
| other info that might help us debug this:
|  Possible unsafe locking scenario:
|
|        CPU0
|        ----
|   lock(batched_entropy_u32.lock);
|   <Interrupt>
|     lock(batched_entropy_u32.lock);
|
|  *** DEADLOCK ***
|
| stack backtrace:
| Call Trace:
…
|  kmem_cache_alloc_trace+0x20e/0x270
|  ipmi_alloc_recv_msg+0x16/0x40
…
|  __do_softirq+0xec/0x48d
|  run_ksoftirqd+0x37/0x60
|  smpboot_thread_fn+0x191/0x290
|  kthread+0xfe/0x130
|  ret_from_fork+0x3a/0x50

Add a spinlock_t to the batched_entropy data structure and acquire the lock while accessing it. Acquire the lock with disabled interrupts because this function may be used from interrupt context.

Remove the batched_entropy_reset_lock lock. Now that we have a lock for the data scructure, we can access it from a remote CPU.

---
 drivers/char/random.c | 52 ++++++++++++++++++++++---------------------
 1 file changed, 27 insertions(+), 25 deletions(-)
drivers/char/random.c +27 -25
index d4d45ccfeefc0..af6e240f98ff4 100644
@@ -2214,8 +2214,8 @@ struct batched_entropy {
 		u32 entropy_u32[CHACHA_BLOCK_SIZE / sizeof(u32)];
 	};
 	unsigned int position;
+	spinlock_t batch_lock;
 };
-static rwlock_t batched_entropy_reset_lock = __RW_LOCK_UNLOCKED(batched_entropy_reset_lock);
 
 /*
  * Get a random word for internal kernel use only. The quality of the random
@@ -2225,12 +2225,14 @@ static rwlock_t batched_entropy_reset_lock = __RW_LOCK_UNLOCKED(batched_entropy_
  * wait_for_random_bytes() should be called and return 0 at least once
  * at any point prior.
  */
-static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u64);
+static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u64) = {
+	.batch_lock	= __SPIN_LOCK_UNLOCKED(batched_entropy_u64.lock),
+};
+
 u64 get_random_u64(void)
 {
 	u64 ret;
-	bool use_lock;
-	unsigned long flags = 0;
+	unsigned long flags;
 	struct batched_entropy *batch;
 	static void *previous;
 
@@ -2245,28 +2247,25 @@ u64 get_random_u64(void)
 
 	warn_unseeded_randomness(&previous);
 
-	use_lock = READ_ONCE(crng_init) < 2;
-	batch = &get_cpu_var(batched_entropy_u64);
-	if (use_lock)
-		read_lock_irqsave(&batched_entropy_reset_lock, flags);
+	batch = raw_cpu_ptr(&batched_entropy_u64);
+	spin_lock_irqsave(&batch->batch_lock, flags);
 	if (batch->position % ARRAY_SIZE(batch->entropy_u64) == 0) {
 		extract_crng((u8 *)batch->entropy_u64);
 		batch->position = 0;
 	}
 	ret = batch->entropy_u64[batch->position++];
-	if (use_lock)
-		read_unlock_irqrestore(&batched_entropy_reset_lock, flags);
-	put_cpu_var(batched_entropy_u64);
+	spin_unlock_irqrestore(&batch->batch_lock, flags);
 	return ret;
 }
 EXPORT_SYMBOL(get_random_u64);
 
-static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u32);
+static DEFINE_PER_CPU(struct batched_entropy, batched_entropy_u32) = {
+	.batch_lock	= __SPIN_LOCK_UNLOCKED(batched_entropy_u32.lock),
+};
 u32 get_random_u32(void)
 {
 	u32 ret;
-	bool use_lock;
-	unsigned long flags = 0;
+	unsigned long flags;
 	struct batched_entropy *batch;
 	static void *previous;
 
@@ -2275,18 +2274,14 @@ u32 get_random_u32(void)
 
 	warn_unseeded_randomness(&previous);
 
-	use_lock = READ_ONCE(crng_init) < 2;
-	batch = &get_cpu_var(batched_entropy_u32);
-	if (use_lock)
-		read_lock_irqsave(&batched_entropy_reset_lock, flags);
+	batch = raw_cpu_ptr(&batched_entropy_u32);
+	spin_lock_irqsave(&batch->batch_lock, flags);
 	if (batch->position % ARRAY_SIZE(batch->entropy_u32) == 0) {
 		extract_crng((u8 *)batch->entropy_u32);
 		batch->position = 0;
 	}
 	ret = batch->entropy_u32[batch->position++];
-	if (use_lock)
-		read_unlock_irqrestore(&batched_entropy_reset_lock, flags);
-	put_cpu_var(batched_entropy_u32);
+	spin_unlock_irqrestore(&batch->batch_lock, flags);
 	return ret;
 }
 EXPORT_SYMBOL(get_random_u32);
@@ -2300,12 +2295,19 @@ static void invalidate_batched_entropy(void)
 	int cpu;
 	unsigned long flags;
 
-	write_lock_irqsave(&batched_entropy_reset_lock, flags);
 	for_each_possible_cpu (cpu) {
-		per_cpu_ptr(&batched_entropy_u32, cpu)->position = 0;
-		per_cpu_ptr(&batched_entropy_u64, cpu)->position = 0;
+		struct batched_entropy *batched_entropy;
+
+		batched_entropy = per_cpu_ptr(&batched_entropy_u32, cpu);
+		spin_lock_irqsave(&batched_entropy->batch_lock, flags);
+		batched_entropy->position = 0;
+		spin_unlock(&batched_entropy->batch_lock);
+
+		batched_entropy = per_cpu_ptr(&batched_entropy_u64, cpu);
+		spin_lock(&batched_entropy->batch_lock);
+		batched_entropy->position = 0;
+		spin_unlock_irqrestore(&batched_entropy->batch_lock, flags);
 	}
-	write_unlock_irqrestore(&batched_entropy_reset_lock, flags);
 }
 
 /**
Signature · 1 line
2.20.1
author maintainer
#167 lore raw

PATCH 167/405 cgroup: protect cgroup->nr_(dying_)descendants by css_set_lock

In reply to Greg Kroah-Hartman

[ Upstream commit 4dcabece4c3a9f9522127be12cc12cc120399b2f ]

The number of descendant cgroups and the number of dying descendant cgroups are currently synchronized using the cgroup_mutex.

The number of descendant cgroups will be required by the cgroup v2 freezer, which will use it to determine if a cgroup is frozen (depending on total number of descendants and number of frozen descendants). It's not always acceptable to grab the cgroup_mutex, especially from quite hot paths (e.g. exit()).

To avoid this, let's additionally synchronize these counters using the css_set_lock.

So, it's safe to read these counters with either cgroup_mutex or css_set_lock locked, and for changing both locks should be acquired.

---
 include/linux/cgroup-defs.h | 5 +++++
 kernel/cgroup/cgroup.c      | 6 ++++++
 2 files changed, 11 insertions(+)
include/linux/cgroup-defs.h +5 -0
index 1c70803e9f770..7d57890cec671 100644
@@ -349,6 +349,11 @@ struct cgroup {
 	 * Dying cgroups are cgroups which were deleted by a user,
 	 * but are still existing because someone else is holding a reference.
 	 * max_descendants is a maximum allowed number of descent cgroups.
+	 *
+	 * nr_descendants and nr_dying_descendants are protected
+	 * by cgroup_mutex and css_set_lock. It's fine to read them holding
+	 * any of cgroup_mutex and css_set_lock; for writing both locks
+	 * should be held.
 	 */
 	int nr_descendants;
 	int nr_dying_descendants;
kernel/cgroup/cgroup.c +6 -0
index 3f2b4bde0f9c3..9fcf6338ea5f9 100644
@@ -4781,9 +4781,11 @@ static void css_release_work_fn(struct work_struct *work)
 		if (cgroup_on_dfl(cgrp))
 			cgroup_rstat_flush(cgrp);
 
+		spin_lock_irq(&css_set_lock);
 		for (tcgrp = cgroup_parent(cgrp); tcgrp;
 		     tcgrp = cgroup_parent(tcgrp))
 			tcgrp->nr_dying_descendants--;
+		spin_unlock_irq(&css_set_lock);
 
 		cgroup_idr_remove(&cgrp->root->cgroup_idr, cgrp->id);
 		cgrp->id = -1;
@@ -5001,12 +5003,14 @@ static struct cgroup *cgroup_create(struct cgroup *parent)
 	if (ret)
 		goto out_psi_free;
 
+	spin_lock_irq(&css_set_lock);
 	for (tcgrp = cgrp; tcgrp; tcgrp = cgroup_parent(tcgrp)) {
 		cgrp->ancestor_ids[tcgrp->level] = tcgrp->id;
 
 		if (tcgrp != cgrp)
 			tcgrp->nr_descendants++;
 	}
+	spin_unlock_irq(&css_set_lock);
 
 	if (notify_on_release(parent))
 		set_bit(CGRP_NOTIFY_ON_RELEASE, &cgrp->flags);
@@ -5291,10 +5295,12 @@ static int cgroup_destroy_locked(struct cgroup *cgrp)
 	if (parent && cgroup_is_threaded(cgrp))
 		parent->nr_threaded_children--;
 
+	spin_lock_irq(&css_set_lock);
 	for (tcgrp = cgroup_parent(cgrp); tcgrp; tcgrp = cgroup_parent(tcgrp)) {
 		tcgrp->nr_descendants--;
 		tcgrp->nr_dying_descendants++;
 	}
+	spin_unlock_irq(&css_set_lock);
 
 	cgroup1_check_for_release(parent);
 
Signature · 1 line
2.20.1
author maintainer
#168 lore raw

PATCH 168/405 sched/core: Check quota and period overflow at usec to nsec conversion

In reply to Greg Kroah-Hartman

[ Upstream commit 1a8b4540db732ca16c9e43ac7c08b1b8f0b252d8 ]

Large values could overflow u64 and pass following sanity checks.

 # echo 18446744073750000 > cpu.cfs_period_us
 # cat cpu.cfs_period_us
 40448
 # echo 18446744073750000 > cpu.cfs_quota_us
 # cat cpu.cfs_quota_us
 40448

After this patch they will fail with -EINVAL.

---
 kernel/sched/core.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
kernel/sched/core.c +6 -1
index 4778c48a7fda4..89c9c1d7d22c5 100644
@@ -6661,8 +6661,10 @@ int tg_set_cfs_quota(struct task_group *tg, long cfs_quota_us)
 	period = ktime_to_ns(tg->cfs_bandwidth.period);
 	if (cfs_quota_us < 0)
 		quota = RUNTIME_INF;
-	else
+	else if ((u64)cfs_quota_us <= U64_MAX / NSEC_PER_USEC)
 		quota = (u64)cfs_quota_us * NSEC_PER_USEC;
+	else
+		return -EINVAL;
 
 	return tg_set_cfs_bandwidth(tg, period, quota);
 }
@@ -6684,6 +6686,9 @@ int tg_set_cfs_period(struct task_group *tg, long cfs_period_us)
 {
 	u64 quota, period;
 
+	if ((u64)cfs_period_us > U64_MAX / NSEC_PER_USEC)
+		return -EINVAL;
+
 	period = (u64)cfs_period_us * NSEC_PER_USEC;
 	quota = tg->cfs_bandwidth.quota;
 
Signature · 1 line
2.20.1
author maintainer
#169 lore raw

PATCH 169/405 sched/rt: Check integer overflow at usec to nsec conversion

In reply to Greg Kroah-Hartman

[ Upstream commit 1a010e29cfa00fee2888fd2fd4983f848cbafb58 ]

Example of unhandled overflows:

 # echo 18446744073709651 > cpu.rt_runtime_us
 # cat cpu.rt_runtime_us
 99
 # echo 18446744073709900 > cpu.rt_period_us
 # cat cpu.rt_period_us
 348

After this patch they will fail with -EINVAL.

---
 kernel/sched/rt.c | 5 +++++
 1 file changed, 5 insertions(+)
kernel/sched/rt.c +5 -0
index 90fa23d36565d..1e6b909dca367 100644
@@ -2555,6 +2555,8 @@ int sched_group_set_rt_runtime(struct task_group *tg, long rt_runtime_us)
 	rt_runtime = (u64)rt_runtime_us * NSEC_PER_USEC;
 	if (rt_runtime_us < 0)
 		rt_runtime = RUNTIME_INF;
+	else if ((u64)rt_runtime_us > U64_MAX / NSEC_PER_USEC)
+		return -EINVAL;
 
 	return tg_set_rt_bandwidth(tg, rt_period, rt_runtime);
 }
@@ -2575,6 +2577,9 @@ int sched_group_set_rt_period(struct task_group *tg, u64 rt_period_us)
 {
 	u64 rt_runtime, rt_period;
 
+	if (rt_period_us > U64_MAX / NSEC_PER_USEC)
+		return -EINVAL;
+
 	rt_period = rt_period_us * NSEC_PER_USEC;
 	rt_runtime = tg->rt_bandwidth.rt_runtime;
 
Signature · 1 line
2.20.1
author maintainer
#170 lore raw

PATCH 170/405 sched/core: Handle overflow in cpu_shares_write_u64

In reply to Greg Kroah-Hartman

[ Upstream commit 5b61d50ab4ef590f5e1d4df15cd2cea5f5715308 ]

Bit shift in scale_load() could overflow shares. This patch saturates it to MAX_SHARES like following sched_group_set_shares().

Example:

 # echo 9223372036854776832 > cpu.shares
 # cat cpu.shares

Before patch: 1024
After pattch: 262144

---
 kernel/sched/core.c | 2 ++
 1 file changed, 2 insertions(+)
kernel/sched/core.c +2 -0
index 89c9c1d7d22c5..a75ad50b5e2ff 100644
@@ -6559,6 +6559,8 @@ static void cpu_cgroup_attach(struct cgroup_taskset *tset)
 static int cpu_shares_write_u64(struct cgroup_subsys_state *css,
 				struct cftype *cftype, u64 shareval)
 {
+	if (shareval > scale_load_down(ULONG_MAX))
+		shareval = MAX_SHARES;
 	return sched_group_set_shares(css_tg(css), scale_load(shareval));
 }
 
Signature · 1 line
2.20.1
author maintainer
#171 lore raw

PATCH 171/405 staging: vc04_services: handle kzalloc failure

In reply to Greg Kroah-Hartman

[ Upstream commit a5112277872a56017b777770e2fd4324d4a6c866 ]

The kzalloc here was being used without checking the return - if the kzalloc fails return VCHIQ_ERROR. The call-site of vchiq_platform_init_state() vchiq_init_state() was not responding to an allocation failure so checks for != VCHIQ_SUCCESS and pass VCHIQ_ERROR up to vchiq_platform_init() which then will fail with -EINVAL.

---
 .../staging/vc04_services/interface/vchiq_arm/vchiq_2835_arm.c | 3 +++
 drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.c | 2 ++
 2 files changed, 5 insertions(+)
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_2835_arm.c +3 -0
index dd4898861b833..eb1e5dcb0d529 100644
@@ -209,6 +209,9 @@ vchiq_platform_init_state(struct vchiq_state *state)
 	struct vchiq_2835_state *platform_state;
 
 	state->platform_state = kzalloc(sizeof(*platform_state), GFP_KERNEL);
+	if (!state->platform_state)
+		return VCHIQ_ERROR;
+
 	platform_state = (struct vchiq_2835_state *)state->platform_state;
 
 	platform_state->inited = 1;
drivers/staging/vc04_services/interface/vchiq_arm/vchiq_core.c +2 -0
index 53f5a1cb4636e..819813e742d8a 100644
@@ -2239,6 +2239,8 @@ vchiq_init_state(struct vchiq_state *state, struct vchiq_slot_zero *slot_zero)
 	local->debug[DEBUG_ENTRIES] = DEBUG_MAX;
 
 	status = vchiq_platform_init_state(state);
+	if (status != VCHIQ_SUCCESS)
+		return VCHIQ_ERROR;
 
 	/*
 		bring up slot handler thread
Signature · 1 line
2.20.1
author maintainer
#172 lore raw

PATCH 172/405 drm/msm/dpu: release resources on modeset failure

In reply to Greg Kroah-Hartman

[ Upstream commit a7fcc3237f31a4e206953bb73cf41bd429442f09 ]

release resources allocated in mode_set if any of the hw check fails. Most of these checks are not necessary and they will be removed in the follow up patches with state based resource allocations.

---
 drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c +5 -2
index 5aa3307f3f0c5..f59c00191a2a2 100644
@@ -1023,13 +1023,13 @@ static void dpu_encoder_virt_mode_set(struct drm_encoder *drm_enc,
 			if (!dpu_enc->hw_pp[i]) {
 				DPU_ERROR_ENC(dpu_enc, "no pp block assigned"
 					     "at idx: %d\n", i);
-				return;
+				goto error;
 			}
 
 			if (!hw_ctl[i]) {
 				DPU_ERROR_ENC(dpu_enc, "no ctl block assigned"
 					     "at idx: %d\n", i);
-				return;
+				goto error;
 			}
 
 			phys->hw_pp = dpu_enc->hw_pp[i];
@@ -1042,6 +1042,9 @@ static void dpu_encoder_virt_mode_set(struct drm_encoder *drm_enc,
 	}
 
 	dpu_enc->mode_set_complete = true;
+
+error:
+	dpu_rm_release(&dpu_kms->rm, drm_enc);
 }
 
 static void _dpu_encoder_virt_enable_helper(struct drm_encoder *drm_enc)
Signature · 1 line
2.20.1
author maintainer
#173 lore raw

PATCH 173/405 drm/msm: a5xx: fix possible object reference leak

In reply to Greg Kroah-Hartman

[ Upstream commit 6cd5235c3135ea84b32469ea51b2aae384eda8af ]

The call to of_get_child_by_name returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
drivers/gpu/drm/msm/adreno/a5xx_gpu.c:57:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 47, but without a corresponding object release within this function. drivers/gpu/drm/msm/adreno/a5xx_gpu.c:66:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 47, but without a corresponding object release within this function. drivers/gpu/drm/msm/adreno/a5xx_gpu.c:118:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 47, but without a corresponding object release within this function. drivers/gpu/drm/msm/adreno/a5xx_gpu.c:57:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 51, but without a corresponding object release within this function. drivers/gpu/drm/msm/adreno/a5xx_gpu.c:66:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 51, but without a corresponding object release within this function. drivers/gpu/drm/msm/adreno/a5xx_gpu.c:118:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 51, but without a corresponding object release within this function.

---
 drivers/gpu/drm/msm/adreno/a5xx_gpu.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)
drivers/gpu/drm/msm/adreno/a5xx_gpu.c +6 -4
index d5f5e56422f57..270da14cba673 100644
@@ -34,7 +34,7 @@ static int zap_shader_load_mdt(struct msm_gpu *gpu, const char *fwname)
 {
 	struct device *dev = &gpu->pdev->dev;
 	const struct firmware *fw;
-	struct device_node *np;
+	struct device_node *np, *mem_np;
 	struct resource r;
 	phys_addr_t mem_phys;
 	ssize_t mem_size;
@@ -48,11 +48,13 @@ static int zap_shader_load_mdt(struct msm_gpu *gpu, const char *fwname)
 	if (!np)
 		return -ENODEV;
 
-	np = of_parse_phandle(np, "memory-region", 0);
-	if (!np)
+	mem_np = of_parse_phandle(np, "memory-region", 0);
+	of_node_put(np);
+	if (!mem_np)
 		return -EINVAL;
 
-	ret = of_address_to_resource(np, 0, &r);
+	ret = of_address_to_resource(mem_np, 0, &r);
+	of_node_put(mem_np);
 	if (ret)
 		return ret;
 
Signature · 8 lines
2.20.1



_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
author maintainer
#174 lore raw

PATCH 174/405 drm/msm: dpu: Dont set frame_busy_mask for async updates

In reply to Greg Kroah-Hartman

[ Upstream commit f98baa3109cea46083d2361ab14a0207d1b1bd16 ]

The frame_busy mask is used in frame_done event handling, which is not invoked for async commits. So an async commit will leave the frame_busy mask populated after it completes and future commits will start with the busy mask incorrect.

This showed up on disable after cursor move. I was hitting the "this should not happen" comment in the frame event worker since frame_busy was set, we queued the event, but there were no frames pending (since async also doesn't set that).

---
 drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/gpu/drm/msm/disp/dpu1/dpu_encoder.c +7 -1
index f59c00191a2a2..dd2c4d11d0e1d 100644
@@ -1550,8 +1550,14 @@ static void _dpu_encoder_kickoff_phys(struct dpu_encoder_virt *dpu_enc,
 		if (!ctl)
 			continue;
 
-		if (phys->split_role != ENC_ROLE_SLAVE)
+		/*
+		 * This is cleared in frame_done worker, which isn't invoked
+		 * for async commits. So don't set this for async, since it'll
+		 * roll over to the next commit.
+		 */
+		if (!async && phys->split_role != ENC_ROLE_SLAVE)
 			set_bit(i, dpu_enc->frame_busy_mask);
+
 		if (!phys->ops.needs_single_flush ||
 				!phys->ops.needs_single_flush(phys))
 			_dpu_encoder_trigger_flush(&dpu_enc->base, phys, 0x0,
Signature · 1 line
2.20.1
author maintainer
#175 lore raw

PATCH 175/405 drm/msm: Fix NULL pointer dereference

In reply to Greg Kroah-Hartman

[ Upstream commit 7603df38cc8c1e5d540b18ec9eb9d62d823197d0 ]

[    3.707412] Unable to handle kernel NULL pointer dereference at virtual address 0000009c
[    3.714511] pgd = (ptrval)
[    3.722742] [0000009c] *pgd=00000000
[    3.725238] Internal error: Oops: 5 [#1] PREEMPT SMP ARM
[    3.728968] Modules linked in:
[    3.734265] CPU: 3 PID: 112 Comm: kworker/3:2 Tainted: G        W         5.0.0-rc7-00183-g06a1c31df9eb #4
[    3.737142] Hardware name: Generic DT based system
[    3.746778] Workqueue: events deferred_probe_work_func
[    3.751542] PC is at msm_gem_map_vma+0x3c/0xac
[    3.756669] LR is at msm_gem_get_and_pin_iova+0xd8/0x134
[    3.761086] pc : [<c07d3b7c>]    lr : [<c07d14f8>]    psr: 60000013
[    3.766560] sp : ee297be8  ip : ed9ab1c0  fp : ed93b800
[    3.772546] r10: ee35e180  r9 : 00000000  r8 : ee297c80
[    3.777752] r7 : 00000000  r6 : 7c100000  r5 : 00000000  r4 : ee35e180
[    3.782968] r3 : 00000001  r2 : 00000003  r1 : ee35e180  r0 : 00000000
[    3.789562] Flags: nZCv  IRQs on  FIQs on  Mode SVC_32  ISA ARM  Segment none
[    3.796079] Control: 10c5787d  Table: 2e3a806a  DAC: 00000051
[    3.803282] Process kworker/3:2 (pid: 112, stack limit = 0x(ptrval))
[    3.809006] Stack: (0xee297be8 to 0xee298000)
[    3.815445] 7be0:                   00000000 c1108c48 eda8c000 00000003 eda8c0fc c1108c48
[    3.819715] 7c00: eda8c000 00000003 eda8c0fc c07d14f8 00000001 c07d1100 7c100000 00000000
[    3.827873] 7c20: eda8c000 bb7ffb78 00000000 eda8c000 00000000 00000000 c0c8b1d4 ee3bfa00
[    3.836037] 7c40: ee3b9800 c07d1684 00000000 c1108c48 ee0d7810 ee3b9800 c0c8b1d4 c07d222c
[    3.844193] 7c60: ee3bfd84 ee297c80 00000000 c0b1d5b0 ee3bfc40 c07dcfd8 ee3bfd84 ee297c80
[    3.852357] 7c80: 0000006d ee3bfc40 ee0d7810 bb7ffb78 c0c8b1d4 00000000 ee3bfc40 c07ddb48
[    3.860516] 7ca0: 00002004 c0eba384 ee3bfc40 c079eba0 ee3bd040 ee3b9800 00000001 ed93b800
[    3.868673] 7cc0: ed9aa100 c07db7e8 ee3bf240 ed9a6500 00000001 ee3b9800 ee3bf2d4 c07a0a30
[    3.876834] 7ce0: ed93b800 7d100000 c1108c48 ee0d7610 ee3b9800 ed93b800 c1108c48 00000000
[    3.884991] 7d00: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[    3.893151] 7d20: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 bb7ffb78
[    3.901310] 7d40: c12113c4 ed93b800 ee3b9800 c1108c48 ee9eec10 00000000 ed93b800 7d100000
[    3.909472] 7d60: eff7b000 c07cf748 7d100000 00000000 c0e9a350 c0b1d5b0 c12113c4 c0961e40
[    3.917633] 7d80: c12113c4 40000113 eeff4bec c0ebe004 00000019 c0b1d230 ee9eeda8 60000113
[    3.925791] 7da0: ee35d300 ee9eeda8 c07ce260 bb7ffb78 c07ce260 ee35d2c0 00000028 00000002
[    3.933950] 7dc0: eeb76280 c118f884 ee0be640 c11c6128 c07ce260 c07ea4ac 00000000 c0962b48
[    3.942108] 7de0: c118f868 00000001 c0ebbc98 ee35d2c0 00000000 eeb76280 00000000 c118f87c
[    3.950270] 7e00: ee35d2c0 00000000 c11c63e0 c118f694 00000019 c07ea5d0 ee0d7810 00000000
[    3.958430] 7e20: c118f694 00000000 00000000 c07f2b0c c120f55c ee0d7810 c120f560 00000000
[    3.966590] 7e40: 00000000 c07f08c4 c07f0e8c ee0d7810 c11ba3d0 ee0d7810 c118f694 c07f0e8c
[    3.974748] 7e60: c1108c48 00000001 c0ebc3cc c11c63f8 c11ba3d0 c07f0c08 00000001 c07f2f8c
[    3.982908] 7e80: c118f694 00000000 ee297ed4 c07f0e8c c1108c48 00000001 c0ebc3cc c11c63f8
[    3.991068] 7ea0: c11ba3d0 c07ee8a0 c11ba3d0 ee82686c ee0baf38 bb7ffb78 ee0d7810 ee0d7810
[    3.999227] 7ec0: c1108c48 ee0d7844 c118faac c07f05b0 ee0d7810 ee0d7810 00000001 bb7ffb78
[    4.007389] 7ee0: ee0d7810 ee0d7810 c118fd18 c118faac c11c63e0 c07ef7d0 ee0d7810 c118fa90
[    4.015548] 7f00: c118fa90 c07efd68 c118fac8 ee27fe00 eefd9c80 eefdcd00 00000000 c118facc
[    4.023708] 7f20: 00000000 c033c038 eefd9c80 eefd9c80 00000008 ee27fe00 ee27fe14 eefd9c80
[    4.031866] 7f40: 00000008 c1103d00 eefd9c98 ee296000 eefd9c80 c033ce54 ee907eac c0b1d230
[    4.040026] 7f60: ee907eac eea24440 ee285000 00000000 ee296000 ee27fe00 c033ce24 eea2445c
[    4.048188] 7f80: ee907eac c0341db0 00000000 ee285000 c0341c8c 00000000 00000000 00000000
[    4.056346] 7fa0: 00000000 00000000 00000000 c03010e8 00000000 00000000 00000000 00000000
[    4.064505] 7fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[    4.072665] 7fe0: 00000000 00000000 00000000 00000000 00000013 00000000 00000000 00000000
[    4.080828] [<c07d3b7c>] (msm_gem_map_vma) from [<c07d14f8>] (msm_gem_get_and_pin_iova+0xd8/0x134)
[    4.088983] [<c07d14f8>] (msm_gem_get_and_pin_iova) from [<c07d1684>] (_msm_gem_kernel_new+0x38/0xac)
[    4.097839] [<c07d1684>] (_msm_gem_kernel_new) from [<c07d222c>] (msm_gem_kernel_new+0x24/0x2c)
[    4.107130] [<c07d222c>] (msm_gem_kernel_new) from [<c07dcfd8>] (dsi_tx_buf_alloc_6g+0x44/0x90)
[    4.115631] [<c07dcfd8>] (dsi_tx_buf_alloc_6g) from [<c07ddb48>] (msm_dsi_host_modeset_init+0x80/0x104)
[    4.124313] [<c07ddb48>] (msm_dsi_host_modeset_init) from [<c07db7e8>] (msm_dsi_modeset_init+0x34/0x1c0)
[    4.133691] [<c07db7e8>] (msm_dsi_modeset_init) from [<c07a0a30>] (mdp5_kms_init+0x764/0x7e0)
[    4.143409] [<c07a0a30>] (mdp5_kms_init) from [<c07cf748>] (msm_drm_bind+0x56c/0x740)
[    4.151824] [<c07cf748>] (msm_drm_bind) from [<c07ea4ac>] (try_to_bring_up_master+0x238/0x2b4)
[    4.159636] [<c07ea4ac>] (try_to_bring_up_master) from [<c07ea5d0>] (component_add+0xa8/0x170)
[    4.168146] [<c07ea5d0>] (component_add) from [<c07f2b0c>] (platform_drv_probe+0x48/0x9c)
[    4.176737] [<c07f2b0c>] (platform_drv_probe) from [<c07f08c4>] (really_probe+0x278/0x404)
[    4.184981] [<c07f08c4>] (really_probe) from [<c07f0c08>] (driver_probe_device+0x78/0x1c0)
[    4.193147] [<c07f0c08>] (driver_probe_device) from [<c07ee8a0>] (bus_for_each_drv+0x74/0xb8)
[    4.201389] [<c07ee8a0>] (bus_for_each_drv) from [<c07f05b0>] (__device_attach+0xd0/0x164)
[    4.209984] [<c07f05b0>] (__device_attach) from [<c07ef7d0>] (bus_probe_device+0x84/0x8c)
[    4.218143] [<c07ef7d0>] (bus_probe_device) from [<c07efd68>] (deferred_probe_work_func+0x48/0xc4)
[    4.226398] [<c07efd68>] (deferred_probe_work_func) from [<c033c038>] (process_one_work+0x204/0x574)
[    4.235254] [<c033c038>] (process_one_work) from [<c033ce54>] (worker_thread+0x30/0x560)
[    4.244534] [<c033ce54>] (worker_thread) from [<c0341db0>] (kthread+0x124/0x154)
[    4.252606] [<c0341db0>] (kthread) from [<c03010e8>] (ret_from_fork+0x14/0x2c)
[    4.259966] Exception stack(0xee297fb0 to 0xee297ff8)
[    4.266998] 7fa0:                                     00000000 00000000 00000000 00000000
[    4.272143] 7fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[    4.280297] 7fe0: 00000000 00000000 00000000 00000000 00000013 00000000
[    4.288451] Code: e5813080 1a000013 e3a03001 e5c4307c (e590009c)
[    4.294933] ---[ end trace 18729cc2bca2b4b3 ]---
---
 drivers/gpu/drm/msm/msm_gem_vma.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/msm/msm_gem_vma.c +1 -1
index 49c04829cf344..fcf7a83f0e6fe 100644
@@ -85,7 +85,7 @@ msm_gem_map_vma(struct msm_gem_address_space *aspace,
 
 	vma->mapped = true;
 
-	if (aspace->mmu)
+	if (aspace && aspace->mmu)
 		ret = aspace->mmu->funcs->map(aspace->mmu, vma->iova, sgt,
 				size, prot);
 
Signature · 1 line
2.20.1
author maintainer
#176 lore raw

PATCH 176/405 irq_work: Do not raise an IPI when queueing work on the local CPU

In reply to Greg Kroah-Hartman

[ Upstream commit 471ba0e686cb13752bc1ff3216c54b69a2d250ea ]

The QEMU PowerPC/PSeries machine model was not expecting a self-IPI, and it may be a bit surprising thing to do, so have irq_work_queue_on do local queueing when target is the current CPU.

[ Simplified the preprocessor comments.
Fixed unbalanced curly brackets pointed out by Thomas. ]

---
 kernel/irq_work.c | 75 ++++++++++++++++++++++++++---------------------
 1 file changed, 42 insertions(+), 33 deletions(-)
kernel/irq_work.c +42 -33
index 6b7cdf17ccf89..73288914ed5e7 100644
@@ -56,61 +56,70 @@ void __weak arch_irq_work_raise(void)
 	 */
 }
 
-/*
- * Enqueue the irq_work @work on @cpu unless it's already pending
- * somewhere.
- *
- * Can be re-enqueued while the callback is still in progress.
- */
-bool irq_work_queue_on(struct irq_work *work, int cpu)
+/* Enqueue on current CPU, work must already be claimed and preempt disabled */
+static void __irq_work_queue_local(struct irq_work *work)
 {
-	/* All work should have been flushed before going offline */
-	WARN_ON_ONCE(cpu_is_offline(cpu));
-
-#ifdef CONFIG_SMP
-
-	/* Arch remote IPI send/receive backend aren't NMI safe */
-	WARN_ON_ONCE(in_nmi());
+	/* If the work is "lazy", handle it from next tick if any */
+	if (work->flags & IRQ_WORK_LAZY) {
+		if (llist_add(&work->llnode, this_cpu_ptr(&lazy_list)) &&
+		    tick_nohz_tick_stopped())
+			arch_irq_work_raise();
+	} else {
+		if (llist_add(&work->llnode, this_cpu_ptr(&raised_list)))
+			arch_irq_work_raise();
+	}
+}
 
+/* Enqueue the irq work @work on the current CPU */
+bool irq_work_queue(struct irq_work *work)
+{
 	/* Only queue if not already pending */
 	if (!irq_work_claim(work))
 		return false;
 
-	if (llist_add(&work->llnode, &per_cpu(raised_list, cpu)))
-		arch_send_call_function_single_ipi(cpu);
-
-#else /* #ifdef CONFIG_SMP */
-	irq_work_queue(work);
-#endif /* #else #ifdef CONFIG_SMP */
+	/* Queue the entry and raise the IPI if needed. */
+	preempt_disable();
+	__irq_work_queue_local(work);
+	preempt_enable();
 
 	return true;
 }
+EXPORT_SYMBOL_GPL(irq_work_queue);
 
-/* Enqueue the irq work @work on the current CPU */
-bool irq_work_queue(struct irq_work *work)
+/*
+ * Enqueue the irq_work @work on @cpu unless it's already pending
+ * somewhere.
+ *
+ * Can be re-enqueued while the callback is still in progress.
+ */
+bool irq_work_queue_on(struct irq_work *work, int cpu)
 {
+#ifndef CONFIG_SMP
+	return irq_work_queue(work);
+
+#else /* CONFIG_SMP: */
+	/* All work should have been flushed before going offline */
+	WARN_ON_ONCE(cpu_is_offline(cpu));
+
 	/* Only queue if not already pending */
 	if (!irq_work_claim(work))
 		return false;
 
-	/* Queue the entry and raise the IPI if needed. */
 	preempt_disable();
-
-	/* If the work is "lazy", handle it from next tick if any */
-	if (work->flags & IRQ_WORK_LAZY) {
-		if (llist_add(&work->llnode, this_cpu_ptr(&lazy_list)) &&
-		    tick_nohz_tick_stopped())
-			arch_irq_work_raise();
+	if (cpu != smp_processor_id()) {
+		/* Arch remote IPI send/receive backend aren't NMI safe */
+		WARN_ON_ONCE(in_nmi());
+		if (llist_add(&work->llnode, &per_cpu(raised_list, cpu)))
+			arch_send_call_function_single_ipi(cpu);
 	} else {
-		if (llist_add(&work->llnode, this_cpu_ptr(&raised_list)))
-			arch_irq_work_raise();
+		__irq_work_queue_local(work);
 	}
-
 	preempt_enable();
 
 	return true;
+#endif /* CONFIG_SMP */
 }
-EXPORT_SYMBOL_GPL(irq_work_queue);
+
 
 bool irq_work_needs_cpu(void)
 {
Signature · 1 line
2.20.1
author maintainer
#177 lore raw

PATCH 177/405 thunderbolt: Take domain lock in switch sysfs attribute callbacks

In reply to Greg Kroah-Hartman

[ Upstream commit 09f11b6c99feaf86a26444bca85dc693b3f58f8b ]

switch_lock was introduced because it allowed serialization of device authorization requests from userspace without need to take the big domain lock (tb->lock). This was fine because device authorization with ICM is just one command that is sent to the firmware. Now that we start to handle all tunneling in the driver switch_lock is not enough because we need to walk over the topology to establish paths.

For this reason drop switch_lock from the driver completely in favour of big domain lock.

There is one complication, though. If userspace is waiting for the lock in tb_switch_set_authorized(), it keeps the device_del() from removing the sysfs attribute because it waits for active users to release the attribute first which leads into following splat:

    INFO: task kworker/u8:3:73 blocked for more than 61 seconds.
          Tainted: G        W         5.1.0-rc1+ #244
    "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
    kworker/u8:3    D12976    73      2 0x80000000
    Workqueue: thunderbolt0 tb_handle_hotplug [thunderbolt]
    Call Trace:
     ? __schedule+0x2e5/0x740
     ? _raw_spin_lock_irqsave+0x12/0x40
     ? prepare_to_wait_event+0xc5/0x160
     schedule+0x2d/0x80
     __kernfs_remove.part.17+0x183/0x1f0
     ? finish_wait+0x80/0x80
     kernfs_remove_by_name_ns+0x4a/0x90
     remove_files.isra.1+0x2b/0x60
     sysfs_remove_group+0x38/0x80
     sysfs_remove_groups+0x24/0x40
     device_remove_attrs+0x3d/0x70
     device_del+0x14c/0x360
     device_unregister+0x15/0x50
     tb_switch_remove+0x9e/0x1d0 [thunderbolt]
     tb_handle_hotplug+0x119/0x5a0 [thunderbolt]
     ? process_one_work+0x1b7/0x420
     process_one_work+0x1b7/0x420
     worker_thread+0x37/0x380
     ? _raw_spin_unlock_irqrestore+0xf/0x30
     ? process_one_work+0x420/0x420
     kthread+0x118/0x130
     ? kthread_create_on_node+0x60/0x60
     ret_from_fork+0x35/0x40

We deal this by following what network stack did for some of their attributes and use mutex_trylock() with restart_syscall(). This makes userspace release the attribute allowing sysfs attribute removal to progress before the write is restarted and eventually fail when the attribute is removed.

---
 drivers/thunderbolt/switch.c | 45 +++++++++++++++---------------------
 drivers/thunderbolt/tb.h     |  3 +--
 2 files changed, 20 insertions(+), 28 deletions(-)
drivers/thunderbolt/switch.c +19 -26
index cd96994dc0947..32e012713dbeb 100644
@@ -10,15 +10,13 @@
 #include <linux/idr.h>
 #include <linux/nvmem-provider.h>
 #include <linux/pm_runtime.h>
+#include <linux/sched/signal.h>
 #include <linux/sizes.h>
 #include <linux/slab.h>
 #include <linux/vmalloc.h>
 
 #include "tb.h"
 
-/* Switch authorization from userspace is serialized by this lock */
-static DEFINE_MUTEX(switch_lock);
-
 /* Switch NVM support */
 
 #define NVM_DEVID		0x05
@@ -254,8 +252,8 @@ static int tb_switch_nvm_write(void *priv, unsigned int offset, void *val,
 	struct tb_switch *sw = priv;
 	int ret = 0;
 
-	if (mutex_lock_interruptible(&switch_lock))
-		return -ERESTARTSYS;
+	if (!mutex_trylock(&sw->tb->lock))
+		return restart_syscall();
 
 	/*
 	 * Since writing the NVM image might require some special steps,
@@ -275,7 +273,7 @@ static int tb_switch_nvm_write(void *priv, unsigned int offset, void *val,
 	memcpy(sw->nvm->buf + offset, val, bytes);
 
 unlock:
-	mutex_unlock(&switch_lock);
+	mutex_unlock(&sw->tb->lock);
 
 	return ret;
 }
@@ -364,10 +362,7 @@ static int tb_switch_nvm_add(struct tb_switch *sw)
 	}
 	nvm->non_active = nvm_dev;
 
-	mutex_lock(&switch_lock);
 	sw->nvm = nvm;
-	mutex_unlock(&switch_lock);
-
 	return 0;
 
 err_nvm_active:
@@ -384,10 +379,8 @@ static void tb_switch_nvm_remove(struct tb_switch *sw)
 {
 	struct tb_switch_nvm *nvm;
 
-	mutex_lock(&switch_lock);
 	nvm = sw->nvm;
 	sw->nvm = NULL;
-	mutex_unlock(&switch_lock);
 
 	if (!nvm)
 		return;
@@ -716,8 +709,8 @@ static int tb_switch_set_authorized(struct tb_switch *sw, unsigned int val)
 {
 	int ret = -EINVAL;
 
-	if (mutex_lock_interruptible(&switch_lock))
-		return -ERESTARTSYS;
+	if (!mutex_trylock(&sw->tb->lock))
+		return restart_syscall();
 
 	if (sw->authorized)
 		goto unlock;
@@ -760,7 +753,7 @@ static int tb_switch_set_authorized(struct tb_switch *sw, unsigned int val)
 	}
 
 unlock:
-	mutex_unlock(&switch_lock);
+	mutex_unlock(&sw->tb->lock);
 	return ret;
 }
 
@@ -817,15 +810,15 @@ static ssize_t key_show(struct device *dev, struct device_attribute *attr,
 	struct tb_switch *sw = tb_to_switch(dev);
 	ssize_t ret;
 
-	if (mutex_lock_interruptible(&switch_lock))
-		return -ERESTARTSYS;
+	if (!mutex_trylock(&sw->tb->lock))
+		return restart_syscall();
 
 	if (sw->key)
 		ret = sprintf(buf, "%*phN\n", TB_SWITCH_KEY_SIZE, sw->key);
 	else
 		ret = sprintf(buf, "\n");
 
-	mutex_unlock(&switch_lock);
+	mutex_unlock(&sw->tb->lock);
 	return ret;
 }
 
@@ -842,8 +835,8 @@ static ssize_t key_store(struct device *dev, struct device_attribute *attr,
 	else if (hex2bin(key, buf, sizeof(key)))
 		return -EINVAL;
 
-	if (mutex_lock_interruptible(&switch_lock))
-		return -ERESTARTSYS;
+	if (!mutex_trylock(&sw->tb->lock))
+		return restart_syscall();
 
 	if (sw->authorized) {
 		ret = -EBUSY;
@@ -858,7 +851,7 @@ static ssize_t key_store(struct device *dev, struct device_attribute *attr,
 		}
 	}
 
-	mutex_unlock(&switch_lock);
+	mutex_unlock(&sw->tb->lock);
 	return ret;
 }
 static DEVICE_ATTR(key, 0600, key_show, key_store);
@@ -904,8 +897,8 @@ static ssize_t nvm_authenticate_store(struct device *dev,
 	bool val;
 	int ret;
 
-	if (mutex_lock_interruptible(&switch_lock))
-		return -ERESTARTSYS;
+	if (!mutex_trylock(&sw->tb->lock))
+		return restart_syscall();
 
 	/* If NVMem devices are not yet added */
 	if (!sw->nvm) {
@@ -953,7 +946,7 @@ static ssize_t nvm_authenticate_store(struct device *dev,
 	}
 
 exit_unlock:
-	mutex_unlock(&switch_lock);
+	mutex_unlock(&sw->tb->lock);
 
 	if (ret)
 		return ret;
@@ -967,8 +960,8 @@ static ssize_t nvm_version_show(struct device *dev,
 	struct tb_switch *sw = tb_to_switch(dev);
 	int ret;
 
-	if (mutex_lock_interruptible(&switch_lock))
-		return -ERESTARTSYS;
+	if (!mutex_trylock(&sw->tb->lock))
+		return restart_syscall();
 
 	if (sw->safe_mode)
 		ret = -ENODATA;
@@ -977,7 +970,7 @@ static ssize_t nvm_version_show(struct device *dev,
 	else
 		ret = sprintf(buf, "%x.%x\n", sw->nvm->major, sw->nvm->minor);
 
-	mutex_unlock(&switch_lock);
+	mutex_unlock(&sw->tb->lock);
 
 	return ret;
 }
drivers/thunderbolt/tb.h +1 -2
index 52584c4003e3a..f5e0282225d1c 100644
@@ -80,8 +80,7 @@ struct tb_switch_nvm {
  * @depth: Depth in the chain this switch is connected (ICM only)
  *
  * When the switch is being added or removed to the domain (other
- * switches) you need to have domain lock held. For switch authorization
- * internal switch_lock is enough.
+ * switches) you need to have domain lock held.
  */
 struct tb_switch {
 	struct device dev;
Signature · 1 line
2.20.1
author maintainer
#178 lore raw

PATCH 178/405 s390/qeth: handle error from qeth_update_from_chp_desc()

In reply to Greg Kroah-Hartman

[ Upstream commit a4cdc9baee0740748f16e50cd70c2607510df492 ]

Subsequent code relies on the values that qeth_update_from_chp_desc() reads from the CHP descriptor. Rather than dealing with weird errors later on, just handle it properly here.

---
 drivers/s390/net/qeth_core_main.c | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)
drivers/s390/net/qeth_core_main.c +10 -4
index 44bd6f04c145d..8c73a99daff3e 100644
@@ -1308,7 +1308,7 @@ static void qeth_set_multiple_write_queues(struct qeth_card *card)
 	card->qdio.no_out_queues = 4;
 }
 
-static void qeth_update_from_chp_desc(struct qeth_card *card)
+static int qeth_update_from_chp_desc(struct qeth_card *card)
 {
 	struct ccw_device *ccwdev;
 	struct channel_path_desc_fmt0 *chp_dsc;
@@ -1318,7 +1318,7 @@ static void qeth_update_from_chp_desc(struct qeth_card *card)
 	ccwdev = card->data.ccwdev;
 	chp_dsc = ccw_device_get_chp_desc(ccwdev, 0);
 	if (!chp_dsc)
-		goto out;
+		return -ENOMEM;
 
 	card->info.func_level = 0x4100 + chp_dsc->desc;
 	if (card->info.type == QETH_CARD_TYPE_IQD)
@@ -1333,6 +1333,7 @@ static void qeth_update_from_chp_desc(struct qeth_card *card)
 	kfree(chp_dsc);
 	QETH_DBF_TEXT_(SETUP, 2, "nr:%x", card->qdio.no_out_queues);
 	QETH_DBF_TEXT_(SETUP, 2, "lvl:%02x", card->info.func_level);
+	return 0;
 }
 
 static void qeth_init_qdio_info(struct qeth_card *card)
@@ -4986,7 +4987,9 @@ int qeth_core_hardsetup_card(struct qeth_card *card, bool *carrier_ok)
 
 	QETH_DBF_TEXT(SETUP, 2, "hrdsetup");
 	atomic_set(&card->force_alloc_skb, 0);
-	qeth_update_from_chp_desc(card);
+	rc = qeth_update_from_chp_desc(card);
+	if (rc)
+		return rc;
 retry:
 	if (retries < 3)
 		QETH_DBF_MESSAGE(2, "Retrying to do IDX activates on device %x.\n",
@@ -5641,7 +5644,9 @@ static int qeth_core_probe_device(struct ccwgroup_device *gdev)
 	}
 
 	qeth_setup_card(card);
-	qeth_update_from_chp_desc(card);
+	rc = qeth_update_from_chp_desc(card);
+	if (rc)
+		goto err_chp_desc;
 
 	card->dev = qeth_alloc_netdev(card);
 	if (!card->dev) {
@@ -5676,6 +5681,7 @@ static int qeth_core_probe_device(struct ccwgroup_device *gdev)
 	qeth_core_free_discipline(card);
 err_load:
 	free_netdev(card->dev);
+err_chp_desc:
 err_card:
 	qeth_core_free_card(card);
 err_dev:
Signature · 1 line
2.20.1
author maintainer
#179 lore raw

PATCH 179/405 USB: core: Dont unbind interfaces following device reset failure

In reply to Greg Kroah-Hartman

[ Upstream commit 381419fa720060ba48b7bbc483be787d5b1dca6f ]

The SCSI core does not like to have devices or hosts unregistered while error recovery is in progress. Trying to do so can lead to
self-deadlock: Part of the removal code tries to obtain a lock already held by the error handler.

This can cause problems for the usb-storage and uas drivers, because their error handler routines perform a USB reset, and if the reset fails then the USB core automatically goes on to unbind all drivers from the device's interfaces -- all while still in the context of the SCSI error handler.

As it turns out, practically all the scenarios leading to a USB reset failure end up causing a device disconnect (the main error pathway in usb_reset_and_verify_device(), at the end of the routine, calls hub_port_logical_disconnect() before returning). As a result, the hub_wq thread will soon become aware of the problem and will unbind all the device's drivers in its own context, not in the error-handler's context.

This means that usb_reset_device() does not need to call usb_unbind_and_rebind_marked_interfaces() in cases where usb_reset_and_verify_device() has returned an error, because hub_wq will take care of everything anyway.

This particular problem was observed in somewhat artificial circumstances, by using usbfs to tell a hub to power-down a port connected to a USB-3 mass storage device using the UAS protocol. With the port turned off, the currently executing command timed out and the error handler started running. The USB reset naturally failed, because the hub port was off, and the error handler deadlocked as described above. Not carrying out the call to usb_unbind_and_rebind_marked_interfaces() fixes this issue.

---
 drivers/usb/core/hub.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/usb/core/hub.c +4 -1
index 8d4631c81b9f0..310eef451db82 100644
@@ -5902,7 +5902,10 @@ int usb_reset_device(struct usb_device *udev)
 					cintf->needs_binding = 1;
 			}
 		}
-		usb_unbind_and_rebind_marked_interfaces(udev);
+
+		/* If the reset failed, hub_wq will unbind drivers later */
+		if (ret == 0)
+			usb_unbind_and_rebind_marked_interfaces(udev);
 	}
 
 	usb_autosuspend_device(udev);
Signature · 1 line
2.20.1
author maintainer
#180 lore raw

PATCH 180/405 x86/irq/64: Limit IST stack overflow check to #DB stack

In reply to Greg Kroah-Hartman

[ Upstream commit 7dbcf2b0b770eeb803a416ee8dcbef78e6389d40 ]

Commit

37fe6a42b343 ("x86: Check stack overflow in detail")

added a broad check for the full exception stack area, i.e. it considers the full exception stack area as valid.

That's wrong in two aspects:

1) It does not check the individual areas one by one

2) #DF, NMI and #MCE are not enabling interrupts which means that a regular device interrupt cannot happen in their context. In fact if a device interrupt hits one of those IST stacks that's a bug because some code path enabled interrupts while handling the exception.

Limit the check to the #DB stack and consider all other IST stacks as 'overflow' or invalid.

Cc Andy Lutomirski <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Josh Poimboeuf <[email protected]>
Cc Mitsuo Hayasaka <[email protected]>
Cc Nicolai Stange <[email protected]>
Cc Sean Christopherson <[email protected]>
Cc x86-ml <[email protected]>
---
 arch/x86/kernel/irq_64.c | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)
arch/x86/kernel/irq_64.c +14 -5
index 0469cd078db15..b50ac9c7397bb 100644
@@ -26,9 +26,18 @@ int sysctl_panic_on_stackoverflow;
 /*
  * Probabilistic stack overflow check:
  *
- * Only check the stack in process context, because everything else
- * runs on the big interrupt stacks. Checking reliably is too expensive,
- * so we just check from interrupts.
+ * Regular device interrupts can enter on the following stacks:
+ *
+ * - User stack
+ *
+ * - Kernel task stack
+ *
+ * - Interrupt stack if a device driver reenables interrupts
+ *   which should only happen in really old drivers.
+ *
+ * - Debug IST stack
+ *
+ * All other contexts are invalid.
  */
 static inline void stack_overflow_check(struct pt_regs *regs)
 {
@@ -53,8 +62,8 @@ static inline void stack_overflow_check(struct pt_regs *regs)
 		return;
 
 	oist = this_cpu_ptr(&orig_ist);
-	estack_top = (u64)oist->ist[0] - EXCEPTION_STKSZ + STACK_TOP_MARGIN;
-	estack_bottom = (u64)oist->ist[N_EXCEPTION_STACKS - 1];
+	estack_bottom = (u64)oist->ist[DEBUG_STACK];
+	estack_top = estack_bottom - DEBUG_STKSZ + STACK_TOP_MARGIN;
 	if (regs->sp >= estack_top && regs->sp <= estack_bottom)
 		return;
 
Signature · 1 line
2.20.1
author maintainer
#181 lore raw

PATCH 181/405 drm: etnaviv: avoid DMA API warning when importing buffers

In reply to Greg Kroah-Hartman

[ Upstream commit 1262cc8893ecb0eb2c21e042d0d268cc180edb61 ]

During boot, I get this kernel warning:

WARNING: CPU: 0 PID: 19001 at kernel/dma/debug.c:1301 debug_dma_map_sg+0x284/0x3dc
etnaviv etnaviv: DMA-API: mapping sg segment longer than device claims to support [len=3145728] [max=65536]
Modules linked in: ip6t_REJECT nf_reject_ipv6 ip6t_rpfilter xt_tcpudp ipt_REJECT nf_reject_ipv4 xt_conntrack ip_set nfnetlink ebtable_broute ebtable_nat ip6table_raw ip6table_nat nf_nat_ipv6 ip6table_mangle iptable_raw iptable_nat nf_nat_ipv4 nf_nat nf_conntrack nf_defrag_ipv4 nf_defrag_ipv6 libcrc32c iptable_mangle ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter caam_jr error snd_soc_imx_spdif imx_thermal snd_soc_imx_audmux nvmem_imx_ocotp snd_soc_sgtl5000
caam imx_sdma virt_dma coda rc_cec v4l2_mem2mem snd_soc_fsl_ssi snd_soc_fsl_spdif imx_vdoa imx_pcm_dma videobuf2_dma_contig etnaviv dw_hdmi_cec gpu_sched dw_hdmi_ahb_audio imx6q_cpufreq nfsd sch_fq_codel ip_tables x_tables
CPU: 0 PID: 19001 Comm: Xorg Not tainted 4.20.0+ #307
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
[<c0019658>] (unwind_backtrace) from [<c001489c>] (show_stack+0x10/0x14)
[<c001489c>] (show_stack) from [<c07fb420>] (dump_stack+0x9c/0xd4)
[<c07fb420>] (dump_stack) from [<c00312dc>] (__warn+0xf8/0x124)
[<c00312dc>] (__warn) from [<c00313d0>] (warn_slowpath_fmt+0x38/0x48)
[<c00313d0>] (warn_slowpath_fmt) from [<c00b14e8>] (debug_dma_map_sg+0x284/0x3dc)
[<c00b14e8>] (debug_dma_map_sg) from [<c046eb40>] (drm_gem_map_dma_buf+0xc4/0x13c)
[<c046eb40>] (drm_gem_map_dma_buf) from [<c04c3314>] (dma_buf_map_attachment+0x38/0x5c)
[<c04c3314>] (dma_buf_map_attachment) from [<c046e728>] (drm_gem_prime_import_dev+0x74/0x104)
[<c046e728>] (drm_gem_prime_import_dev) from [<c046e5bc>] (drm_gem_prime_fd_to_handle+0x84/0x17c)
[<c046e5bc>] (drm_gem_prime_fd_to_handle) from [<c046edd0>] (drm_prime_fd_to_handle_ioctl+0x38/0x4c)
[<c046edd0>] (drm_prime_fd_to_handle_ioctl) from [<c0460efc>] (drm_ioctl_kernel+0x90/0xc8)
[<c0460efc>] (drm_ioctl_kernel) from [<c0461114>] (drm_ioctl+0x1e0/0x3b0)
[<c0461114>] (drm_ioctl) from [<c01cae20>] (do_vfs_ioctl+0x90/0xa48)
[<c01cae20>] (do_vfs_ioctl) from [<c01cb80c>] (ksys_ioctl+0x34/0x60)
[<c01cb80c>] (ksys_ioctl) from [<c0009000>] (ret_fast_syscall+0x0/0x28)
Exception stack(0xd81a9fa8 to 0xd81a9ff0)
9fa0:                   b6c69c88 bec613f8 00000009 c00c642e bec613f8 b86c4600
9fc0: b6c69c88 bec613f8 c00c642e 00000036 012762e0 01276348 00000300 012d91f8
9fe0: b6989f18 bec613dc b697185c b667be5c
irq event stamp: 47905
hardirqs last  enabled at (47913): [<c0098824>] console_unlock+0x46c/0x680
hardirqs last disabled at (47922): [<c0098470>] console_unlock+0xb8/0x680
softirqs last  enabled at (47754): [<c000a484>] __do_softirq+0x344/0x540
softirqs last disabled at (47701): [<c0038700>] irq_exit+0x124/0x144
---[ end trace af477747acbcc642 ]---

The reason is the contiguous buffer exceeds the default maximum segment size of 64K as specified by dma_get_max_seg_size() in linux/dma-mapping.h. Fix this by providing our own segment size, which is set to 2GiB to cover the window found in MMUv1 GPUs.

---
 drivers/gpu/drm/etnaviv/etnaviv_drv.c | 5 +++++
 drivers/gpu/drm/etnaviv/etnaviv_drv.h | 1 +
 2 files changed, 6 insertions(+)
drivers/gpu/drm/etnaviv/etnaviv_drv.c +5 -0
index 18c27f795cf61..3156450723bad 100644
@@ -515,6 +515,9 @@ static int etnaviv_bind(struct device *dev)
 	}
 	drm->dev_private = priv;
 
+	dev->dma_parms = &priv->dma_parms;
+	dma_set_max_seg_size(dev, SZ_2G);
+
 	mutex_init(&priv->gem_lock);
 	INIT_LIST_HEAD(&priv->gem_list);
 	priv->num_gpus = 0;
@@ -552,6 +555,8 @@ static void etnaviv_unbind(struct device *dev)
 
 	component_unbind_all(dev, drm);
 
+	dev->dma_parms = NULL;
+
 	drm->dev_private = NULL;
 	kfree(priv);
 
drivers/gpu/drm/etnaviv/etnaviv_drv.h +1 -0
index a6a7ded37ef1d..6a4ea127c4f1a 100644
@@ -42,6 +42,7 @@ struct etnaviv_file_private {
 
 struct etnaviv_drm_private {
 	int num_gpus;
+	struct device_dma_parameters dma_parms;
 	struct etnaviv_gpu *gpu[ETNA_MAX_PIPES];
 
 	/* list of GEM objects: */
Signature · 1 line
2.20.1
author maintainer
#182 lore raw

PATCH 182/405 dt-bindings: phy-qcom-qmp: Add UFS PHY reset

In reply to Greg Kroah-Hartman

[ Upstream commit 95cee0b4e30a09a411a17e9a3bc6b72ed92063da ]

Add a required reset to the SDM845 UFS phy to express the PHY reset bit inside the UFS controller register space. Before this change, this reset was not expressed in the DT, and the driver utilized two different callbacks (phy_init and phy_poweron) to implement a two-phase initialization procedure that involved deasserting this reset between init and poweron. This abused the two callbacks and diluted their purpose.

That scheme does not work as regulators cannot be turned off in phy_poweroff because they were turned on in init, rather than poweron. The net result is that regulators are left on in suspend that shouldn't be.

This new scheme gives the UFS reset to the PHY, so that it can fully initialize itself in a single callback. We can then turn regulators on during poweron and off during poweroff.

---
 Documentation/devicetree/bindings/phy/qcom-qmp-phy.txt | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
Documentation/devicetree/bindings/phy/qcom-qmp-phy.txt +4 -2
index 5d181fc3cc182..4a78ba8b85bc0 100644
@@ -59,7 +59,8 @@ Required properties:
 	   one for each entry in reset-names.
  - reset-names: "phy" for reset of phy block,
 		"common" for phy common block reset,
-		"cfg" for phy's ahb cfg block reset.
+		"cfg" for phy's ahb cfg block reset,
+		"ufsphy" for the PHY reset in the UFS controller.
 
 		For "qcom,ipq8074-qmp-pcie-phy" must contain:
 			"phy", "common".
@@ -74,7 +75,8 @@ Required properties:
 			"phy", "common".
 		For "qcom,sdm845-qmp-usb3-uni-phy" must contain:
 			"phy", "common".
-		For "qcom,sdm845-qmp-ufs-phy": no resets are listed.
+		For "qcom,sdm845-qmp-ufs-phy": must contain:
+			"ufsphy".
 
  - vdda-phy-supply: Phandle to a regulator supply to PHY core block.
  - vdda-pll-supply: Phandle to 1.8V regulator supply to PHY refclk pll block.
Signature · 1 line
2.20.1
author maintainer
#183 lore raw

PATCH 183/405 phy: sun4i-usb: Make sure to disable PHY0 passby for peripheral mode

In reply to Greg Kroah-Hartman

[ Upstream commit e6f32efb1b128344a2c7df9875bc1a1abaa1d395 ]

On platforms where the MUSB and HCI controllers share PHY0, PHY passby is required when using the HCI controller with the PHY, but it must be disabled when the MUSB controller is used instead.

Without this, PHY0 passby is always enabled, which results in broken peripheral mode on such platforms (e.g. H3/H5).

Fixes ba4bdc9e1dc0 ("PHY: sunxi: Add driver for sunxi usb phy")
---
 drivers/phy/allwinner/phy-sun4i-usb.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/phy/allwinner/phy-sun4i-usb.c +4 -0
index 4bbd9ede38c83..cc5af961778d6 100644
@@ -554,6 +554,7 @@ static void sun4i_usb_phy0_id_vbus_det_scan(struct work_struct *work)
 	struct sun4i_usb_phy_data *data =
 		container_of(work, struct sun4i_usb_phy_data, detect.work);
 	struct phy *phy0 = data->phys[0].phy;
+	struct sun4i_usb_phy *phy = phy_get_drvdata(phy0);
 	bool force_session_end, id_notify = false, vbus_notify = false;
 	int id_det, vbus_det;
 
@@ -610,6 +611,9 @@ static void sun4i_usb_phy0_id_vbus_det_scan(struct work_struct *work)
 			mutex_unlock(&phy0->mutex);
 		}
 
+		/* Enable PHY0 passby for host mode only. */
+		sun4i_usb_phy_passby(phy, !id_det);
+
 		/* Re-route PHY0 if necessary */
 		if (data->cfg->phy0_dual_route)
 			sun4i_usb_phy0_reroute(data, id_det);
Signature · 1 line
2.20.1
author maintainer
#184 lore raw

PATCH 184/405 phy: mapphone-mdm6600: add gpiolib dependency

In reply to Greg Kroah-Hartman

[ Upstream commit 208d3423ee463ab257908456f6bbca4024ab63f7 ]

gcc points out that when CONFIG_GPIOLIB is disabled, gpiod_get_array_value_cansleep() returns 0 but fails to set its output:

drivers/phy/motorola/phy-mapphone-mdm6600.c: In function 'phy_mdm6600_status':
drivers/phy/motorola/phy-mapphone-mdm6600.c:220:24: error: 'values[0]' is used uninitialized in this function [-Werror=uninitialized]

This could be fixed more generally in gpiolib by returning a failure code, but for this specific case, the easier workaround is to add a gpiolib dependency.

Fixes 5d1ebbda0318 ("phy: mapphone-mdm6600: Add USB PHY driver for MDM6600 on Droid 4")
---
 drivers/phy/motorola/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/phy/motorola/Kconfig +1 -1
index 82651524ffb9c..718f8729701df 100644
@@ -13,7 +13,7 @@ config PHY_CPCAP_USB
 
 config PHY_MAPPHONE_MDM6600
 	tristate "Motorola Mapphone MDM6600 modem USB PHY driver"
-	depends on OF && USB_SUPPORT
+	depends on OF && USB_SUPPORT && GPIOLIB
 	select GENERIC_PHY
 	help
 	  Enable this for MDM6600 USB modem to work on Motorola phones
Signature · 1 line
2.20.1
author maintainer
#185 lore raw

PATCH 185/405 phy: ti: usb2: fix OMAP_CONTROL_PHY dependency

In reply to Greg Kroah-Hartman

[ Upstream commit d41ce98a122c13ea77938af04ef06fb12ae0c69e ]

With randconfig build testing on arm64, we can run into a configuration that has CONFIG_OMAP_CONTROL_PHY=m and CONFIG_OMAP_USB2=y, which in turn causes a link failure:

drivers/phy/ti/phy-omap-usb2.o: In function `omap_usb_phy_power':
phy-omap-usb2.c:(.text+0x17c): undefined reference to `omap_control_phy_power'

I could not come up with a good way to correctly describe the relation of the two symbols, but if we just select CONFIG_OMAP_CONTROL_PHY during compile testing, we can no longer run into the broken configuration.

Fixes 6777cee3a872 ("phy: ti: usb2: Add support for AM654 USB2 PHY")
---
 drivers/phy/ti/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/phy/ti/Kconfig +1 -1
index 103efc456a12e..022ac16f626cf 100644
@@ -37,7 +37,7 @@ config OMAP_USB2
 	depends on USB_SUPPORT
 	select GENERIC_PHY
 	select USB_PHY
-	select OMAP_CONTROL_PHY if ARCH_OMAP2PLUS
+	select OMAP_CONTROL_PHY if ARCH_OMAP2PLUS || COMPILE_TEST
 	help
 	  Enable this to support the transceiver that is part of SOC. This
 	  driver takes care of all the PHY functionality apart from comparator.
Signature · 1 line
2.20.1
author maintainer
#186 lore raw

PATCH 186/405 dpaa2-eth: Fix Rx classification status

In reply to Greg Kroah-Hartman

[ Upstream commit df8e249be866e2f762be11b14a9e7a94752614d4 ]

Set the Rx flow classification enable flag only if key config operation is successful.

Fixes 3f9b5c9 ("dpaa2-eth: Configure Rx flow classification key")

---
 drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c +6 -1
index dc339dc1adb21..57cbaa38d2477 100644
@@ -2796,6 +2796,7 @@ int dpaa2_eth_set_hash(struct net_device *net_dev, u64 flags)
 static int dpaa2_eth_set_cls(struct dpaa2_eth_priv *priv)
 {
 	struct device *dev = priv->net_dev->dev.parent;
+	int err;
 
 	/* Check if we actually support Rx flow classification */
 	if (dpaa2_eth_has_legacy_dist(priv)) {
@@ -2814,9 +2815,13 @@ static int dpaa2_eth_set_cls(struct dpaa2_eth_priv *priv)
 		return -EOPNOTSUPP;
 	}
 
+	err = dpaa2_eth_set_dist_key(priv->net_dev, DPAA2_ETH_RX_DIST_CLS, 0);
+	if (err)
+		return err;
+
 	priv->rx_cls_enabled = 1;
 
-	return dpaa2_eth_set_dist_key(priv->net_dev, DPAA2_ETH_RX_DIST_CLS, 0);
+	return 0;
 }
 
 /* Bind the DPNI to its needed objects and resources: buffer pool, DPIOs,
Signature · 1 line
2.20.1
author maintainer
#187 lore raw

PATCH 187/405 i40e: Able to add up to 16 MAC filters on an untrusted VF

In reply to Greg Kroah-Hartman

[ Upstream commit 06b6e2a2333eb3581567a7ac43ca465ef45f4daa ]

This patch fixes the problem with the driver being able to add only 7 multicast MAC address filters instead of 16. The problem is fixed by changing the maximum number of MAC address filters to 16+1+1 (two extra are needed because the driver uses 1 for unicast MAC address and 1 for broadcast).

---
 drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/net/ethernet/intel/i40e/i40e_virtchnl_pf.c +4 -2
index 0b5b867c9fbcb..2b0362c827e98 100644
@@ -2454,8 +2454,10 @@ static int i40e_vc_get_stats_msg(struct i40e_vf *vf, u8 *msg)
 				      (u8 *)&stats, sizeof(stats));
 }
 
-/* If the VF is not trusted restrict the number of MAC/VLAN it can program */
-#define I40E_VC_MAX_MAC_ADDR_PER_VF 12
+/* If the VF is not trusted restrict the number of MAC/VLAN it can program
+ * MAC filters: 16 for multicast, 1 for MAC, 1 for broadcast
+ */
+#define I40E_VC_MAX_MAC_ADDR_PER_VF (16 + 1 + 1)
 #define I40E_VC_MAX_VLAN_PER_VF 8
 
 /**
Signature · 1 line
2.20.1
author maintainer
#188 lore raw

PATCH 188/405 i40e: dont allow changes to HW VLAN stripping on active port VLANs

In reply to Greg Kroah-Hartman

[ Upstream commit bfb0ebed53857cfc57f11c63fa3689940d71c1c8 ]

Modifying the VLAN stripping options when a port VLAN is configured will break traffic for the VSI, and conceptually doesn't make sense, so don't allow this.

---
 drivers/net/ethernet/intel/i40e/i40e_main.c | 8 ++++++++
 1 file changed, 8 insertions(+)
drivers/net/ethernet/intel/i40e/i40e_main.c +8 -0
index b1c265012c8ad..ac9fcb0976890 100644
@@ -2654,6 +2654,10 @@ void i40e_vlan_stripping_enable(struct i40e_vsi *vsi)
 	struct i40e_vsi_context ctxt;
 	i40e_status ret;
 
+	/* Don't modify stripping options if a port VLAN is active */
+	if (vsi->info.pvid)
+		return;
+
 	if ((vsi->info.valid_sections &
 	     cpu_to_le16(I40E_AQ_VSI_PROP_VLAN_VALID)) &&
 	    ((vsi->info.port_vlan_flags & I40E_AQ_VSI_PVLAN_MODE_MASK) == 0))
@@ -2684,6 +2688,10 @@ void i40e_vlan_stripping_disable(struct i40e_vsi *vsi)
 	struct i40e_vsi_context ctxt;
 	i40e_status ret;
 
+	/* Don't modify stripping options if a port VLAN is active */
+	if (vsi->info.pvid)
+		return;
+
 	if ((vsi->info.valid_sections &
 	     cpu_to_le16(I40E_AQ_VSI_PROP_VLAN_VALID)) &&
 	    ((vsi->info.port_vlan_flags & I40E_AQ_VSI_PVLAN_EMOD_MASK) ==
Signature · 1 line
2.20.1
author maintainer
#189 lore raw

PATCH 189/405 ACPI/IORT: Reject platform device creation on NUMA node mapping failure

In reply to Greg Kroah-Hartman

[ Upstream commit 36a2ba07757df790b4a874efb1a105b9330a9ae7 ]

In a system where, through IORT firmware mappings, the SMMU device is mapped to a NUMA node that is not online, the kernel bootstrap results in the following crash:

  Unable to handle kernel paging request at virtual address 0000000000001388
  Mem abort info:
    ESR = 0x96000004
    Exception class = DABT (current EL), IL = 32 bits
    SET = 0, FnV = 0
    EA = 0, S1PTW = 0
  Data abort info:
    ISV = 0, ISS = 0x00000004
    CM = 0, WnR = 0
  [0000000000001388] user address but active_mm is swapper
  Internal error: Oops: 96000004 [#1] SMP
  Modules linked in:
  CPU: 5 PID: 1 Comm: swapper/0 Not tainted 5.0.0 #15
  pstate: 80c00009 (Nzcv daif +PAN +UAO)
  pc : __alloc_pages_nodemask+0x13c/0x1068
  lr : __alloc_pages_nodemask+0xdc/0x1068
  ...
  Process swapper/0 (pid: 1, stack limit = 0x(____ptrval____))
  Call trace:
   __alloc_pages_nodemask+0x13c/0x1068
   new_slab+0xec/0x570
   ___slab_alloc+0x3e0/0x4f8
   __slab_alloc+0x60/0x80
   __kmalloc_node_track_caller+0x10c/0x478
   devm_kmalloc+0x44/0xb0
   pinctrl_bind_pins+0x4c/0x188
   really_probe+0x78/0x2b8
   driver_probe_device+0x64/0x110
   device_driver_attach+0x74/0x98
   __driver_attach+0x9c/0xe8
   bus_for_each_dev+0x84/0xd8
   driver_attach+0x30/0x40
   bus_add_driver+0x170/0x218
   driver_register+0x64/0x118
   __platform_driver_register+0x54/0x60
   arm_smmu_driver_init+0x24/0x2c
   do_one_initcall+0xbc/0x328
   kernel_init_freeable+0x304/0x3ac
   kernel_init+0x18/0x110
   ret_from_fork+0x10/0x1c
  Code: f90013b5 b9410fa1 1a9f0694 b50014c2 (b9400804)
  ---[ end trace dfeaed4c373a32da ]--

Change the dev_set_proximity() hook prototype so that it returns a value and make it return failure if the PXM->NUMA-node mapping corresponds to an offline node, fixing the crash.

---
 drivers/acpi/arm64/iort.c | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)
drivers/acpi/arm64/iort.c +14 -5
index e48894e002ba8..a46c2c162c03e 100644
@@ -1232,18 +1232,24 @@ static bool __init arm_smmu_v3_is_coherent(struct acpi_iort_node *node)
 /*
  * set numa proximity domain for smmuv3 device
  */
-static void  __init arm_smmu_v3_set_proximity(struct device *dev,
+static int  __init arm_smmu_v3_set_proximity(struct device *dev,
 					      struct acpi_iort_node *node)
 {
 	struct acpi_iort_smmu_v3 *smmu;
 
 	smmu = (struct acpi_iort_smmu_v3 *)node->node_data;
 	if (smmu->flags & ACPI_IORT_SMMU_V3_PXM_VALID) {
-		set_dev_node(dev, acpi_map_pxm_to_node(smmu->pxm));
+		int node = acpi_map_pxm_to_node(smmu->pxm);
+
+		if (node != NUMA_NO_NODE && !node_online(node))
+			return -EINVAL;
+
+		set_dev_node(dev, node);
 		pr_info("SMMU-v3[%llx] Mapped to Proximity domain %d\n",
 			smmu->base_address,
 			smmu->pxm);
 	}
+	return 0;
 }
 #else
 #define arm_smmu_v3_set_proximity NULL
@@ -1318,7 +1324,7 @@ struct iort_dev_config {
 	int (*dev_count_resources)(struct acpi_iort_node *node);
 	void (*dev_init_resources)(struct resource *res,
 				     struct acpi_iort_node *node);
-	void (*dev_set_proximity)(struct device *dev,
+	int (*dev_set_proximity)(struct device *dev,
 				    struct acpi_iort_node *node);
 };
 
@@ -1369,8 +1375,11 @@ static int __init iort_add_platform_device(struct acpi_iort_node *node,
 	if (!pdev)
 		return -ENOMEM;
 
-	if (ops->dev_set_proximity)
-		ops->dev_set_proximity(&pdev->dev, node);
+	if (ops->dev_set_proximity) {
+		ret = ops->dev_set_proximity(&pdev->dev, node);
+		if (ret)
+			goto dev_put;
+	}
 
 	count = ops->dev_count_resources(node);
 
Signature · 1 line
2.20.1
author maintainer
#190 lore raw

PATCH 190/405 arm64: vdso: Fix clock_getres() for CLOCK_REALTIME

In reply to Greg Kroah-Hartman

[ Upstream commit 81fb8736dd81da3fe94f28968dac60f392ec6746 ]

clock_getres() in the vDSO library has to preserve the same behaviour of posix_get_hrtimer_res().

In particular, posix_get_hrtimer_res() does:

    sec = 0;
    ns = hrtimer_resolution;

where 'hrtimer_resolution' depends on whether or not high resolution timers are enabled, which is a runtime decision.

The vDSO incorrectly returns the constant CLOCK_REALTIME_RES. Fix this by exposing 'hrtimer_resolution' in the vDSO datapage and returning that instead.

[will: Use WRITE_ONCE(), move adr off COARSE path, renumber labels, use 'w' reg]
---
 arch/arm64/include/asm/vdso_datapage.h | 1 +
 arch/arm64/kernel/asm-offsets.c        | 2 +-
 arch/arm64/kernel/vdso.c               | 3 +++
 arch/arm64/kernel/vdso/gettimeofday.S  | 7 +++----
 4 files changed, 8 insertions(+), 5 deletions(-)
arch/arm64/include/asm/vdso_datapage.h +1 -0
index 2b9a63771eda8..f89263c8e11af 100644
@@ -38,6 +38,7 @@ struct vdso_data {
 	__u32 tz_minuteswest;	/* Whacky timezone stuff */
 	__u32 tz_dsttime;
 	__u32 use_syscall;
+	__u32 hrtimer_res;
 };
 
 #endif /* !__ASSEMBLY__ */
arch/arm64/kernel/asm-offsets.c +1 -1
index 7f40dcbdd51d0..e10e2a5d9ddcf 100644
@@ -94,7 +94,7 @@ int main(void)
   DEFINE(CLOCK_REALTIME,	CLOCK_REALTIME);
   DEFINE(CLOCK_MONOTONIC,	CLOCK_MONOTONIC);
   DEFINE(CLOCK_MONOTONIC_RAW,	CLOCK_MONOTONIC_RAW);
-  DEFINE(CLOCK_REALTIME_RES,	MONOTONIC_RES_NSEC);
+  DEFINE(CLOCK_REALTIME_RES,	offsetof(struct vdso_data, hrtimer_res));
   DEFINE(CLOCK_REALTIME_COARSE,	CLOCK_REALTIME_COARSE);
   DEFINE(CLOCK_MONOTONIC_COARSE,CLOCK_MONOTONIC_COARSE);
   DEFINE(CLOCK_COARSE_RES,	LOW_RES_NSEC);
arch/arm64/kernel/vdso.c +3 -0
index 2d419006ad433..ec0bb588d7553 100644
@@ -232,6 +232,9 @@ void update_vsyscall(struct timekeeper *tk)
 	vdso_data->wtm_clock_sec		= tk->wall_to_monotonic.tv_sec;
 	vdso_data->wtm_clock_nsec		= tk->wall_to_monotonic.tv_nsec;
 
+	/* Read without the seqlock held by clock_getres() */
+	WRITE_ONCE(vdso_data->hrtimer_res, hrtimer_resolution);
+
 	if (!use_syscall) {
 		/* tkr_mono.cycle_last == tkr_raw.cycle_last */
 		vdso_data->cs_cycle_last	= tk->tkr_mono.cycle_last;
arch/arm64/kernel/vdso/gettimeofday.S +3 -4
index e8f60112818fc..856fee6d35129 100644
@@ -308,13 +308,14 @@ ENTRY(__kernel_clock_getres)
 	ccmp	w0, #CLOCK_MONOTONIC_RAW, #0x4, ne
 	b.ne	1f
 
-	ldr	x2, 5f
+	adr	vdso_data, _vdso_data
+	ldr	w2, [vdso_data, #CLOCK_REALTIME_RES]
 	b	2f
 1:
 	cmp	w0, #CLOCK_REALTIME_COARSE
 	ccmp	w0, #CLOCK_MONOTONIC_COARSE, #0x4, ne
 	b.ne	4f
-	ldr	x2, 6f
+	ldr	x2, 5f
 2:
 	cbz	x1, 3f
 	stp	xzr, x2, [x1]
@@ -328,8 +329,6 @@ ENTRY(__kernel_clock_getres)
 	svc	#0
 	ret
 5:
-	.quad	CLOCK_REALTIME_RES
-6:
 	.quad	CLOCK_COARSE_RES
 	.cfi_endproc
 ENDPROC(__kernel_clock_getres)
Signature · 1 line
2.20.1
author maintainer
#191 lore raw

PATCH 191/405 RDMA/cxgb4: Fix null pointer dereference on alloc_skb failure

In reply to Greg Kroah-Hartman

[ Upstream commit a6d2a5a92e67d151c98886babdc86d530d27111c ]

Currently if alloc_skb fails to allocate the skb a null skb is passed to t4_set_arp_err_handler and this ends up dereferencing the null skb. Avoid the NULL pointer dereference by checking for a NULL skb and returning early.

Addresses-Coverity: ("Dereference null return")

Fixes b38a0ad8ec11 ("RDMA/cxgb4: Set arp error handler for PASS_ACCEPT_RPL messages")
---
 drivers/infiniband/hw/cxgb4/cm.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/infiniband/hw/cxgb4/cm.c +2 -0
index 4d232bdf9e976..689ba6bc2ca9c 100644
@@ -457,6 +457,8 @@ static struct sk_buff *get_skb(struct sk_buff *skb, int len, gfp_t gfp)
 		skb_reset_transport_header(skb);
 	} else {
 		skb = alloc_skb(len, gfp);
+		if (!skb)
+			return NULL;
 	}
 	t4_set_arp_err_handler(skb, NULL, NULL);
 	return skb;
Signature · 1 line
2.20.1
author maintainer
#192 lore raw

PATCH 192/405 fscrypt: use READ_ONCE() to access ->i_crypt_info

In reply to Greg Kroah-Hartman

[ Upstream commit e37a784d8b6a1e726de5ddc7b4809c086a08db09 ]

->i_crypt_info starts out NULL and may later be locklessly set to a non-NULL value by the cmpxchg() in fscrypt_get_encryption_info().

But ->i_crypt_info is used directly, which technically is incorrect. It's a data race, and it doesn't include the data dependency barrier needed to safely dereference the pointer on at least one architecture.

Fix this by using READ_ONCE() instead. Note: we don't need to use smp_load_acquire(), since dereferencing the pointer only requires a data dependency barrier, which is already included in READ_ONCE(). We also don't need READ_ONCE() in places where ->i_crypt_info is unconditionally dereferenced, since it must have already been checked.

Also downgrade the cmpxchg() to cmpxchg_release(), since RELEASE semantics are sufficient on the write side.

---
 fs/crypto/crypto.c      | 2 +-
 fs/crypto/fname.c       | 4 ++--
 fs/crypto/keyinfo.c     | 4 ++--
 fs/crypto/policy.c      | 6 +++---
 include/linux/fscrypt.h | 3 ++-
 5 files changed, 10 insertions(+), 9 deletions(-)
fs/crypto/crypto.c +1 -1
index 4dc788e3bc96b..fe38b53060454 100644
@@ -334,7 +334,7 @@ static int fscrypt_d_revalidate(struct dentry *dentry, unsigned int flags)
 	spin_lock(&dentry->d_lock);
 	cached_with_key = dentry->d_flags & DCACHE_ENCRYPTED_WITH_KEY;
 	spin_unlock(&dentry->d_lock);
-	dir_has_key = (d_inode(dir)->i_crypt_info != NULL);
+	dir_has_key = fscrypt_has_encryption_key(d_inode(dir));
 	dput(dir);
 
 	/*
fs/crypto/fname.c +2 -2
index 7ff40a73dbece..050384c79f40e 100644
@@ -269,7 +269,7 @@ int fscrypt_fname_disk_to_usr(struct inode *inode,
 	if (iname->len < FS_CRYPTO_BLOCK_SIZE)
 		return -EUCLEAN;
 
-	if (inode->i_crypt_info)
+	if (fscrypt_has_encryption_key(inode))
 		return fname_decrypt(inode, iname, oname);
 
 	if (iname->len <= FSCRYPT_FNAME_MAX_UNDIGESTED_SIZE) {
@@ -336,7 +336,7 @@ int fscrypt_setup_filename(struct inode *dir, const struct qstr *iname,
 	if (ret)
 		return ret;
 
-	if (dir->i_crypt_info) {
+	if (fscrypt_has_encryption_key(dir)) {
 		if (!fscrypt_fname_encrypted_size(dir, iname->len,
 						  dir->i_sb->s_cop->max_namelen,
 						  &fname->crypto_buf.len))
fs/crypto/keyinfo.c +2 -2
index 322ce9686bdba..bf291c10c682f 100644
@@ -509,7 +509,7 @@ int fscrypt_get_encryption_info(struct inode *inode)
 	u8 *raw_key = NULL;
 	int res;
 
-	if (inode->i_crypt_info)
+	if (fscrypt_has_encryption_key(inode))
 		return 0;
 
 	res = fscrypt_initialize(inode->i_sb->s_cop->flags);
@@ -573,7 +573,7 @@ int fscrypt_get_encryption_info(struct inode *inode)
 	if (res)
 		goto out;
 
-	if (cmpxchg(&inode->i_crypt_info, NULL, crypt_info) == NULL)
+	if (cmpxchg_release(&inode->i_crypt_info, NULL, crypt_info) == NULL)
 		crypt_info = NULL;
 out:
 	if (res == -ENOKEY)
fs/crypto/policy.c +3 -3
index bd7eaf9b3f003..d536889ac31bf 100644
@@ -194,8 +194,8 @@ int fscrypt_has_permitted_context(struct inode *parent, struct inode *child)
 	res = fscrypt_get_encryption_info(child);
 	if (res)
 		return 0;
-	parent_ci = parent->i_crypt_info;
-	child_ci = child->i_crypt_info;
+	parent_ci = READ_ONCE(parent->i_crypt_info);
+	child_ci = READ_ONCE(child->i_crypt_info);
 
 	if (parent_ci && child_ci) {
 		return memcmp(parent_ci->ci_master_key_descriptor,
@@ -246,7 +246,7 @@ int fscrypt_inherit_context(struct inode *parent, struct inode *child,
 	if (res < 0)
 		return res;
 
-	ci = parent->i_crypt_info;
+	ci = READ_ONCE(parent->i_crypt_info);
 	if (ci == NULL)
 		return -ENOKEY;
 
include/linux/fscrypt.h +2 -1
index e5194fc3983e9..08246f068fd89 100644
@@ -79,7 +79,8 @@ struct fscrypt_ctx {
 
 static inline bool fscrypt_has_encryption_key(const struct inode *inode)
 {
-	return (inode->i_crypt_info != NULL);
+	/* pairs with cmpxchg_release() in fscrypt_get_encryption_info() */
+	return READ_ONCE(inode->i_crypt_info) != NULL;
 }
 
 static inline bool fscrypt_dummy_context_enabled(struct inode *inode)
Signature · 1 line
2.20.1
author maintainer
#193 lore raw

PATCH 193/405 perf/x86/msr: Add Icelake support

In reply to Greg Kroah-Hartman

[ Upstream commit cf50d79a8cfe5adae37fec026220b009559bbeed ]

Icelake is the same as the existing Skylake parts.

---
 arch/x86/events/msr.c | 1 +
 1 file changed, 1 insertion(+)
arch/x86/events/msr.c +1 -0
index a878e6286e4af..f3f4c2263501d 100644
@@ -89,6 +89,7 @@ static bool test_intel(int idx)
 	case INTEL_FAM6_SKYLAKE_X:
 	case INTEL_FAM6_KABYLAKE_MOBILE:
 	case INTEL_FAM6_KABYLAKE_DESKTOP:
+	case INTEL_FAM6_ICELAKE_MOBILE:
 		if (idx == PERF_MSR_SMI || idx == PERF_MSR_PPERF)
 			return true;
 		break;
Signature · 1 line
2.20.1
author maintainer
#194 lore raw

PATCH 194/405 perf/x86/intel/rapl: Add Icelake support

In reply to Greg Kroah-Hartman

[ Upstream commit b3377c3acb9e54cf86efcfe25f2e792bca599ed4 ]

Icelake support the same RAPL counters as Skylake.

---
 arch/x86/events/intel/rapl.c | 2 ++
 1 file changed, 2 insertions(+)
arch/x86/events/intel/rapl.c +2 -0
index 94dc564146ca8..37ebf6fc5415b 100644
@@ -775,6 +775,8 @@ static const struct x86_cpu_id rapl_cpu_match[] __initconst = {
 	X86_RAPL_MODEL_MATCH(INTEL_FAM6_ATOM_GOLDMONT_X, hsw_rapl_init),
 
 	X86_RAPL_MODEL_MATCH(INTEL_FAM6_ATOM_GOLDMONT_PLUS, hsw_rapl_init),
+
+	X86_RAPL_MODEL_MATCH(INTEL_FAM6_ICELAKE_MOBILE,  skl_rapl_init),
 	{},
 };
 
Signature · 1 line
2.20.1
author maintainer
#195 lore raw

PATCH 195/405 perf/x86/intel/cstate: Add Icelake support

In reply to Greg Kroah-Hartman

[ Upstream commit f08c47d1f86c6dc666c7e659d94bf6d4492aa9d7 ]

Icelake uses the same C-state residency events as Sandy Bridge.

---
 arch/x86/events/intel/cstate.c | 2 ++
 1 file changed, 2 insertions(+)
arch/x86/events/intel/cstate.c +2 -0
index d41de9af7a39b..6072f92cb8eaf 100644
@@ -578,6 +578,8 @@ static const struct x86_cpu_id intel_cstates_match[] __initconst = {
 	X86_CSTATES_MODEL(INTEL_FAM6_ATOM_GOLDMONT_X, glm_cstates),
 
 	X86_CSTATES_MODEL(INTEL_FAM6_ATOM_GOLDMONT_PLUS, glm_cstates),
+
+	X86_CSTATES_MODEL(INTEL_FAM6_ICELAKE_MOBILE, snb_cstates),
 	{ },
 };
 MODULE_DEVICE_TABLE(x86cpu, intel_cstates_match);
Signature · 1 line
2.20.1
author maintainer
#196 lore raw

PATCH 196/405 PM / devfreq: Fix static checker warning in try_then_request_governor

In reply to Greg Kroah-Hartman

[ Upstream commit b53b0128052ffd687797d5f4deeb76327e7b5711 ]

The patch 23c7b54ca1cd: "PM / devfreq: Fix devfreq_add_device() when drivers are built as modules." leads to the following static checker warning:

drivers/devfreq/devfreq.c:1043 governor_store()
warn: 'governor' can also be NULL

The reason is that the try_then_request_governor() function returns both error pointers and NULL. It should just return error pointers, so fix this by returning a ERR_PTR to the error intead of returning NULL.

Fixes 23c7b54ca1cd ("PM / devfreq: Fix devfreq_add_device() when drivers are built as modules.")
---
 drivers/devfreq/devfreq.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/devfreq/devfreq.c +2 -2
index 0ae3de76833b7..839621b044f49 100644
@@ -228,7 +228,7 @@ static struct devfreq_governor *find_devfreq_governor(const char *name)
  * if is not found. This can happen when both drivers (the governor driver
  * and the driver that call devfreq_add_device) are built as modules.
  * devfreq_list_lock should be held by the caller. Returns the matched
- * governor's pointer.
+ * governor's pointer or an error pointer.
  */
 static struct devfreq_governor *try_then_request_governor(const char *name)
 {
@@ -254,7 +254,7 @@ static struct devfreq_governor *try_then_request_governor(const char *name)
 		/* Restore previous state before return */
 		mutex_lock(&devfreq_list_lock);
 		if (err)
-			return NULL;
+			return ERR_PTR(err);
 
 		governor = find_devfreq_governor(name);
 	}
Signature · 1 line
2.20.1
author maintainer
#197 lore raw

PATCH 197/405 hwmon: (vt1211) Use request_muxed_region for Super-IO accesses

In reply to Greg Kroah-Hartman

[ Upstream commit 14b97ba5c20056102b3dd22696bf17b057e60976 ]

Super-IO accesses may fail on a system with no or unmapped LPC bus.

Also, other drivers may attempt to access the LPC bus at the same time, resulting in undefined behavior.

Use request_muxed_region() to ensure that IO access on the requested address space is supported, and to ensure that access by multiple drivers is synchronized.

Fixes 2219cd81a6cd ("hwmon/vt1211: Add probing of alternate config index port")
---
 drivers/hwmon/vt1211.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
drivers/hwmon/vt1211.c +12 -3
index 3a6bfa51cb94f..95d5e8ec8b7fc 100644
@@ -226,15 +226,21 @@ static inline void superio_select(int sio_cip, int ldn)
 	outb(ldn, sio_cip + 1);
 }
 
-static inline void superio_enter(int sio_cip)
+static inline int superio_enter(int sio_cip)
 {
+	if (!request_muxed_region(sio_cip, 2, DRVNAME))
+		return -EBUSY;
+
 	outb(0x87, sio_cip);
 	outb(0x87, sio_cip);
+
+	return 0;
 }
 
 static inline void superio_exit(int sio_cip)
 {
 	outb(0xaa, sio_cip);
+	release_region(sio_cip, 2);
 }
 
 /* ---------------------------------------------------------------------
@@ -1282,11 +1288,14 @@ static int __init vt1211_device_add(unsigned short address)
 
 static int __init vt1211_find(int sio_cip, unsigned short *address)
 {
-	int err = -ENODEV;
+	int err;
 	int devid;
 
-	superio_enter(sio_cip);
+	err = superio_enter(sio_cip);
+	if (err)
+		return err;
 
+	err = -ENODEV;
 	devid = force_id ? force_id : superio_inb(sio_cip, SIO_VT1211_DEVID);
 	if (devid != SIO_VT1211_ID)
 		goto EXIT;
Signature · 1 line
2.20.1
author maintainer
#198 lore raw

PATCH 198/405 hwmon: (smsc47m1) Use request_muxed_region for Super-IO accesses

In reply to Greg Kroah-Hartman

[ Upstream commit d6410408ad2a798c4cc685252c1baa713be0ad69 ]

Super-IO accesses may fail on a system with no or unmapped LPC bus.

Also, other drivers may attempt to access the LPC bus at the same time, resulting in undefined behavior.

Use request_muxed_region() to ensure that IO access on the requested address space is supported, and to ensure that access by multiple drivers is synchronized.

Fixes 8d5d45fb1468 ("I2C: Move hwmon drivers (2/3)")
Cc John Garry <[email protected]>
---
 drivers/hwmon/smsc47m1.c | 28 +++++++++++++++++++---------
 1 file changed, 19 insertions(+), 9 deletions(-)
drivers/hwmon/smsc47m1.c +19 -9
index c7b6a425e2c02..5eeac9853d0ae 100644
@@ -73,16 +73,21 @@ superio_inb(int reg)
 /* logical device for fans is 0x0A */
 #define superio_select() superio_outb(0x07, 0x0A)
 
-static inline void
+static inline int
 superio_enter(void)
 {
+	if (!request_muxed_region(REG, 2, DRVNAME))
+		return -EBUSY;
+
 	outb(0x55, REG);
+	return 0;
 }
 
 static inline void
 superio_exit(void)
 {
 	outb(0xAA, REG);
+	release_region(REG, 2);
 }
 
 #define SUPERIO_REG_ACT		0x30
@@ -531,8 +536,12 @@ static int __init smsc47m1_find(struct smsc47m1_sio_data *sio_data)
 {
 	u8 val;
 	unsigned short addr;
+	int err;
+
+	err = superio_enter();
+	if (err)
+		return err;
 
-	superio_enter();
 	val = force_id ? force_id : superio_inb(SUPERIO_REG_DEVID);
 
 	/*
@@ -608,13 +617,14 @@ static int __init smsc47m1_find(struct smsc47m1_sio_data *sio_data)
 static void smsc47m1_restore(const struct smsc47m1_sio_data *sio_data)
 {
 	if ((sio_data->activate & 0x01) == 0) {
-		superio_enter();
-		superio_select();
-
-		pr_info("Disabling device\n");
-		superio_outb(SUPERIO_REG_ACT, sio_data->activate);
-
-		superio_exit();
+		if (!superio_enter()) {
+			superio_select();
+			pr_info("Disabling device\n");
+			superio_outb(SUPERIO_REG_ACT, sio_data->activate);
+			superio_exit();
+		} else {
+			pr_warn("Failed to disable device\n");
+		}
 	}
 }
 
Signature · 1 line
2.20.1
author maintainer
#199 lore raw

PATCH 199/405 hwmon: (smsc47b397) Use request_muxed_region for Super-IO accesses

In reply to Greg Kroah-Hartman

[ Upstream commit 8c0826756744c0ac1df600a5e4cca1a341b13101 ]

Super-IO accesses may fail on a system with no or unmapped LPC bus.

Also, other drivers may attempt to access the LPC bus at the same time, resulting in undefined behavior.

Use request_muxed_region() to ensure that IO access on the requested address space is supported, and to ensure that access by multiple drivers is synchronized.

Fixes 8d5d45fb1468 ("I2C: Move hwmon drivers (2/3)")
Cc John Garry <[email protected]>
---
 drivers/hwmon/smsc47b397.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)
drivers/hwmon/smsc47b397.c +11 -2
index c0775084dde00..60e193f2e9707 100644
@@ -72,14 +72,19 @@ static inline void superio_select(int ld)
 	superio_outb(0x07, ld);
 }
 
-static inline void superio_enter(void)
+static inline int superio_enter(void)
 {
+	if (!request_muxed_region(REG, 2, DRVNAME))
+		return -EBUSY;
+
 	outb(0x55, REG);
+	return 0;
 }
 
 static inline void superio_exit(void)
 {
 	outb(0xAA, REG);
+	release_region(REG, 2);
 }
 
 #define SUPERIO_REG_DEVID	0x20
@@ -300,8 +305,12 @@ static int __init smsc47b397_find(void)
 	u8 id, rev;
 	char *name;
 	unsigned short addr;
+	int err;
+
+	err = superio_enter();
+	if (err)
+		return err;
 
-	superio_enter();
 	id = force_id ? force_id : superio_inb(SUPERIO_REG_DEVID);
 
 	switch (id) {
Signature · 1 line
2.20.1
author maintainer
#200 lore raw

PATCH 200/405 hwmon: (pc87427) Use request_muxed_region for Super-IO accesses

In reply to Greg Kroah-Hartman

[ Upstream commit 755a9b0f8aaa5639ba5671ca50080852babb89ce ]

Super-IO accesses may fail on a system with no or unmapped LPC bus.

Also, other drivers may attempt to access the LPC bus at the same time, resulting in undefined behavior.

Use request_muxed_region() to ensure that IO access on the requested address space is supported, and to ensure that access by multiple drivers is synchronized.

Fixes ba224e2c4f0a7 ("hwmon: New PC87427 hardware monitoring driver")
Cc John Garry <[email protected]>
---
 drivers/hwmon/pc87427.c | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)
drivers/hwmon/pc87427.c +13 -1
index d1a3f2040c007..58eee8fa3e6d6 100644
@@ -106,6 +106,13 @@ static const char *logdev_str[2] = { DRVNAME " FMC", DRVNAME " HMC" };
 #define LD_IN		1
 #define LD_TEMP		1
 
+static inline int superio_enter(int sioaddr)
+{
+	if (!request_muxed_region(sioaddr, 2, DRVNAME))
+		return -EBUSY;
+	return 0;
+}
+
 static inline void superio_outb(int sioaddr, int reg, int val)
 {
 	outb(reg, sioaddr);
@@ -122,6 +129,7 @@ static inline void superio_exit(int sioaddr)
 {
 	outb(0x02, sioaddr);
 	outb(0x02, sioaddr + 1);
+	release_region(sioaddr, 2);
 }
 
 /*
@@ -1195,7 +1203,11 @@ static int __init pc87427_find(int sioaddr, struct pc87427_sio_data *sio_data)
 {
 	u16 val;
 	u8 cfg, cfg_b;
-	int i, err = 0;
+	int i, err;
+
+	err = superio_enter(sioaddr);
+	if (err)
+		return err;
 
 	/* Identify device */
 	val = force_id ? force_id : superio_inb(sioaddr, SIOREG_DEVID);
Signature · 1 line
2.20.1
author maintainer
#201 lore raw

PATCH 201/405 hwmon: (f71805f) Use request_muxed_region for Super-IO accesses

In reply to Greg Kroah-Hartman

[ Upstream commit 73e6ff71a7ea924fb7121d576a2d41e3be3fc6b5 ]

Super-IO accesses may fail on a system with no or unmapped LPC bus.

Unable to handle kernel paging request at virtual address ffffffbffee0002e
pgd = ffffffc1d68d4000
[ffffffbffee0002e] *pgd=0000000000000000, *pud=0000000000000000
Internal error: Oops: 94000046 [#1] PREEMPT SMP
Modules linked in: f71805f(+) hwmon
CPU: 3 PID: 1659 Comm: insmod Not tainted 4.5.0+ #88
Hardware name: linux,dummy-virt (DT)
task: ffffffc1f6665400 ti: ffffffc1d6418000 task.ti: ffffffc1d6418000
PC is at f71805f_find+0x6c/0x358 [f71805f]

Also, other drivers may attempt to access the LPC bus at the same time, resulting in undefined behavior.

Use request_muxed_region() to ensure that IO access on the requested address space is supported, and to ensure that access by multiple drivers is synchronized.

Fixes e53004e20a58e ("hwmon: New f71805f driver")
Cc John Garry <[email protected]>
---
 drivers/hwmon/f71805f.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
drivers/hwmon/f71805f.c +12 -3
index 73c681162653b..623736d2a7c1d 100644
@@ -96,17 +96,23 @@ superio_select(int base, int ld)
 	outb(ld, base + 1);
 }
 
-static inline void
+static inline int
 superio_enter(int base)
 {
+	if (!request_muxed_region(base, 2, DRVNAME))
+		return -EBUSY;
+
 	outb(0x87, base);
 	outb(0x87, base);
+
+	return 0;
 }
 
 static inline void
 superio_exit(int base)
 {
 	outb(0xaa, base);
+	release_region(base, 2);
 }
 
 /*
@@ -1561,7 +1567,7 @@ static int __init f71805f_device_add(unsigned short address,
 static int __init f71805f_find(int sioaddr, unsigned short *address,
 			       struct f71805f_sio_data *sio_data)
 {
-	int err = -ENODEV;
+	int err;
 	u16 devid;
 
 	static const char * const names[] = {
@@ -1569,8 +1575,11 @@ static int __init f71805f_find(int sioaddr, unsigned short *address,
 		"F71872F/FG or F71806F/FG",
 	};
 
-	superio_enter(sioaddr);
+	err = superio_enter(sioaddr);
+	if (err)
+		return err;
 
+	err = -ENODEV;
 	devid = superio_inw(sioaddr, SIO_REG_MANID);
 	if (devid != SIO_FINTEK_ID)
 		goto exit;
Signature · 1 line
2.20.1
author maintainer
#202 lore raw

PATCH 202/405 scsi: libsas: Do discovery on empty PHY to update PHY info

In reply to Greg Kroah-Hartman

[ Upstream commit d8649fc1c5e40e691d589ed825998c36a947491c ]

When we discover the PHY is empty in sas_rediscover_dev(), the PHY information (like negotiated linkrate) is not updated.

As such, for a user examining sysfs for that PHY, they would see incorrect values:

root@(none)$ cd /sys/class/sas_phy/phy-0:0:20
root@(none)$ more negotiated_linkrate
3.0 Gbit
root@(none)$ echo 0 > enable
root@(none)$ more negotiated_linkrate
3.0 Gbit

So fix this, simply discover the PHY again, even though we know it's empty; in the above example, this gives us:

root@(none)$ more negotiated_linkrate
Phy disabled

We must do this after unregistering the device associated with the PHY (in sas_unregister_devs_sas_addr()).

---
 drivers/scsi/libsas/sas_expander.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/scsi/libsas/sas_expander.c +5 -0
index 17b45a0c7bc38..3611a4ef0d150 100644
@@ -2052,6 +2052,11 @@ static int sas_rediscover_dev(struct domain_device *dev, int phy_id, bool last)
 	if ((SAS_ADDR(sas_addr) == 0) || (res == -ECOMM)) {
 		phy->phy_state = PHY_EMPTY;
 		sas_unregister_devs_sas_addr(dev, phy_id, last);
+		/*
+		 * Even though the PHY is empty, for convenience we discover
+		 * the PHY to update the PHY info, like negotiated linkrate.
+		 */
+		sas_ex_phy_discover(dev, phy_id);
 		return res;
 	} else if (SAS_ADDR(sas_addr) == SAS_ADDR(phy->attached_sas_addr) &&
 		   dev_type_flutter(type, phy->attached_dev_type)) {
Signature · 1 line
2.20.1
author maintainer
#203 lore raw

PATCH 203/405 mmc: core: make pwrseq_emmc (partially) support sleepy GPIO controllers

In reply to Greg Kroah-Hartman

[ Upstream commit 002ee28e8b322d4d4b7b83234b5d0f4ebd428eda ]

pwrseq_emmc.c implements a HW reset procedure for eMMC chip by driving a GPIO line.

It registers the .reset() cb on mmc_pwrseq_ops and it registers a system restart notification handler; both of them perform reset by unconditionally calling gpiod_set_value().

If the eMMC reset line is tied to a GPIO controller whose driver can sleep (i.e. I2C GPIO controller), then the kernel would spit warnings when trying to reset the eMMC chip by means of .reset() mmc_pwrseq_ops cb (that is exactly what I'm seeing during boot).

Furthermore, on system reset we would gets to the system restart notification handler with disabled interrupts - local_irq_disable() is called in machine_restart() at least on ARM/ARM64 - and we would be in trouble when the GPIO driver tries to sleep (which indeed doesn't happen here, likely because in my case the machine specific code doesn't call do_kernel_restart(), I guess..).

This patch fixes the .reset() cb to make use of gpiod_set_value_cansleep(), so that the eMMC gets reset on boot without complaints, while, since there isn't that much we can do, we avoid register the restart handler if the GPIO controller has a sleepy driver (and we spit a dev_notice() message to let people know)..

This had been tested on a downstream 4.9 kernel with backported commit 83f37ee7ba33 ("mmc: pwrseq: Add reset callback to the struct mmc_pwrseq_ops") and commit ae60fb031cf2 ("mmc: core: Don't do eMMC HW reset when resuming the eMMC card"), because I couldn't boot my board otherwise. Maybe worth to RFT.

---
 drivers/mmc/core/pwrseq_emmc.c | 38 ++++++++++++++++++----------------
 1 file changed, 20 insertions(+), 18 deletions(-)
drivers/mmc/core/pwrseq_emmc.c +20 -18
index efb8a7965dd4a..154f4204d58cb 100644
@@ -30,19 +30,14 @@ struct mmc_pwrseq_emmc {
 
 #define to_pwrseq_emmc(p) container_of(p, struct mmc_pwrseq_emmc, pwrseq)
 
-static void __mmc_pwrseq_emmc_reset(struct mmc_pwrseq_emmc *pwrseq)
-{
-	gpiod_set_value(pwrseq->reset_gpio, 1);
-	udelay(1);
-	gpiod_set_value(pwrseq->reset_gpio, 0);
-	udelay(200);
-}
-
 static void mmc_pwrseq_emmc_reset(struct mmc_host *host)
 {
 	struct mmc_pwrseq_emmc *pwrseq =  to_pwrseq_emmc(host->pwrseq);
 
-	__mmc_pwrseq_emmc_reset(pwrseq);
+	gpiod_set_value_cansleep(pwrseq->reset_gpio, 1);
+	udelay(1);
+	gpiod_set_value_cansleep(pwrseq->reset_gpio, 0);
+	udelay(200);
 }
 
 static int mmc_pwrseq_emmc_reset_nb(struct notifier_block *this,
@@ -50,8 +45,11 @@ static int mmc_pwrseq_emmc_reset_nb(struct notifier_block *this,
 {
 	struct mmc_pwrseq_emmc *pwrseq = container_of(this,
 					struct mmc_pwrseq_emmc, reset_nb);
+	gpiod_set_value(pwrseq->reset_gpio, 1);
+	udelay(1);
+	gpiod_set_value(pwrseq->reset_gpio, 0);
+	udelay(200);
 
-	__mmc_pwrseq_emmc_reset(pwrseq);
 	return NOTIFY_DONE;
 }
 
@@ -72,14 +70,18 @@ static int mmc_pwrseq_emmc_probe(struct platform_device *pdev)
 	if (IS_ERR(pwrseq->reset_gpio))
 		return PTR_ERR(pwrseq->reset_gpio);
 
-	/*
-	 * register reset handler to ensure emmc reset also from
-	 * emergency_reboot(), priority 255 is the highest priority
-	 * so it will be executed before any system reboot handler.
-	 */
-	pwrseq->reset_nb.notifier_call = mmc_pwrseq_emmc_reset_nb;
-	pwrseq->reset_nb.priority = 255;
-	register_restart_handler(&pwrseq->reset_nb);
+	if (!gpiod_cansleep(pwrseq->reset_gpio)) {
+		/*
+		 * register reset handler to ensure emmc reset also from
+		 * emergency_reboot(), priority 255 is the highest priority
+		 * so it will be executed before any system reboot handler.
+		 */
+		pwrseq->reset_nb.notifier_call = mmc_pwrseq_emmc_reset_nb;
+		pwrseq->reset_nb.priority = 255;
+		register_restart_handler(&pwrseq->reset_nb);
+	} else {
+		dev_notice(dev, "EMMC reset pin tied to a sleepy GPIO driver; reset on emergency-reboot disabled\n");
+	}
 
 	pwrseq->pwrseq.ops = &mmc_pwrseq_emmc_ops;
 	pwrseq->pwrseq.dev = dev;
Signature · 1 line
2.20.1
author maintainer
#204 lore raw

PATCH 204/405 mmc_spi: add a status check for spi_sync_locked

In reply to Greg Kroah-Hartman

[ Upstream commit 611025983b7976df0183390a63a2166411d177f1 ]

In case spi_sync_locked fails, the fix reports the error and returns the error code upstream.

---
 drivers/mmc/host/mmc_spi.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/mmc/host/mmc_spi.c +4 -0
index 1b1498805972c..a3533935e282b 100644
@@ -819,6 +819,10 @@ mmc_spi_readblock(struct mmc_spi_host *host, struct spi_transfer *t,
 	}
 
 	status = spi_sync_locked(spi, &host->m);
+	if (status < 0) {
+		dev_dbg(&spi->dev, "read error %d\n", status);
+		return status;
+	}
 
 	if (host->dma_dev) {
 		dma_sync_single_for_cpu(host->dma_dev,
Signature · 1 line
2.20.1
author maintainer
#205 lore raw

PATCH 205/405 mmc: sdhci-of-esdhc: add erratum eSDHC5 support

In reply to Greg Kroah-Hartman

[ Upstream commit a46e42712596b51874f04c73f1cdf1017f88df52 ]

Software writing to the Transfer Type configuration register (system clock domain) can cause a setup/hold violation in the CRC flops (card clock domain), which can cause write accesses to be sent with corrupt CRC values. This issue occurs only for write preceded by read. this erratum is to fix this issue.

---
 drivers/mmc/host/sdhci-of-esdhc.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/mmc/host/sdhci-of-esdhc.c +3 -0
index 4e669b4edfc11..9da53e548691b 100644
@@ -1074,6 +1074,9 @@ static int sdhci_esdhc_probe(struct platform_device *pdev)
 	if (esdhc->vendor_ver > VENDOR_V_22)
 		host->quirks &= ~SDHCI_QUIRK_NO_BUSY_IRQ;
 
+	if (of_find_compatible_node(NULL, NULL, "fsl,p2020-esdhc"))
+		host->quirks2 |= SDHCI_QUIRK_RESET_AFTER_REQUEST;
+
 	if (of_device_is_compatible(np, "fsl,p5040-esdhc") ||
 	    of_device_is_compatible(np, "fsl,p5020-esdhc") ||
 	    of_device_is_compatible(np, "fsl,p4080-esdhc") ||
Signature · 1 line
2.20.1
author maintainer
#206 lore raw

PATCH 206/405 mmc: sdhci-of-esdhc: add erratum A-009204 support

In reply to Greg Kroah-Hartman

[ Upstream commit 5dd195522562542bc6ebe6e7bd47890d8b7ca93c ]

In the event of that any data error (like, IRQSTAT[DCE]) occurs during an eSDHC data transaction where DMA is used for data transfer to/from the system memory, setting the SYSCTL[RSTD] register may cause a system hang. If software sets the register SYSCTL[RSTD] to 1 for error recovery while DMA transferring is not complete, eSDHC may hang the system bus. This happens because the software register SYSCTL[RSTD] resets the DMA engine without waiting for the completion of pending system transactions. This erratum is to fix this issue.

---
 drivers/mmc/host/sdhci-of-esdhc.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/mmc/host/sdhci-of-esdhc.c +3 -0
index 9da53e548691b..4fc4d2c7643c5 100644
@@ -694,6 +694,9 @@ static void esdhc_reset(struct sdhci_host *host, u8 mask)
 	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
 	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
 
+	if (of_find_compatible_node(NULL, NULL, "fsl,p2020-esdhc"))
+		mdelay(5);
+
 	if (mask & SDHCI_RESET_ALL) {
 		val = sdhci_readl(host, ESDHC_TBCTL);
 		val &= ~ESDHC_TB_EN;
Signature · 1 line
2.20.1
author maintainer
#207 lore raw

PATCH 207/405 mmc: sdhci-of-esdhc: add erratum eSDHC-A001 and A-008358 support

In reply to Greg Kroah-Hartman

[ Upstream commit 05cb6b2a66fa7837211a060878e91be5eb10cb07 ]

eSDHC-A001: The data timeout counter (SYSCTL[DTOCV]) is not reliable for DTOCV values 0x4(2^17 SD clock), 0x8(2^21 SD clock), and 0xC(2^25 SD clock). The data timeout counter can count from 2^13–2^27, but for values 2^17, 2^21, and 2^25, the timeout counter counts for only 2^13 SD clocks.
A-008358: The data timeout counter value loaded into the timeout counter is less than expected and can result into early timeout error in case of eSDHC data transactions. The table below shows the expected vs actual timeout period for different values of SYSCTL[DTOCV]:
these two erratum has the same quirk to control it, and set SDHCI_QUIRK_RESET_AFTER_REQUEST to fix above issue.

---
 drivers/mmc/host/sdhci-of-esdhc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/mmc/host/sdhci-of-esdhc.c +3 -1
index 4fc4d2c7643c5..7e0eae8dafae0 100644
@@ -1077,8 +1077,10 @@ static int sdhci_esdhc_probe(struct platform_device *pdev)
 	if (esdhc->vendor_ver > VENDOR_V_22)
 		host->quirks &= ~SDHCI_QUIRK_NO_BUSY_IRQ;
 
-	if (of_find_compatible_node(NULL, NULL, "fsl,p2020-esdhc"))
+	if (of_find_compatible_node(NULL, NULL, "fsl,p2020-esdhc")) {
 		host->quirks2 |= SDHCI_QUIRK_RESET_AFTER_REQUEST;
+		host->quirks2 |= SDHCI_QUIRK_BROKEN_TIMEOUT_VAL;
+	}
 
 	if (of_device_is_compatible(np, "fsl,p5040-esdhc") ||
 	    of_device_is_compatible(np, "fsl,p5020-esdhc") ||
Signature · 1 line
2.20.1
author maintainer
#208 lore raw

PATCH 208/405 net: hns3: free the pending skb when clean RX ring

In reply to Greg Kroah-Hartman

[ Upstream commit cc5ff6e90f808f9a4c8229bf2f1de0dfe5d7931c ]

If there is pending skb in RX flow when close the port, and the pending buffer is not cleaned, the new packet will be added to the pending skb when the port opens again, and the first new packet has error data.

This patch cleans the pending skb when clean RX ring.

---
 drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c +7 -0
index 0208efe282775..d6b488c2de332 100644
@@ -3851,6 +3851,13 @@ static int hns3_clear_rx_ring(struct hns3_enet_ring *ring)
 		ring_ptr_move_fw(ring, next_to_use);
 	}
 
+	/* Free the pending skb in rx ring */
+	if (ring->skb) {
+		dev_kfree_skb_any(ring->skb);
+		ring->skb = NULL;
+		ring->pending_buf = 0;
+	}
+
 	return 0;
 }
 
Signature · 1 line
2.20.1
author maintainer
#209 lore raw

PATCH 209/405 drm/amdgpu: fix old fence check in amdgpu_fence_emit

In reply to Greg Kroah-Hartman

[ Upstream commit 3d2aca8c8620346abdba96c6300d2c0b90a1d0cc ]

We don't hold a reference to the old fence, so it can go away any time we are waiting for it to signal.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c | 24 ++++++++++++++++-------
 1 file changed, 17 insertions(+), 7 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +17 -7
index ee47c11e92ce7..4dee2326b29c3 100644
@@ -136,8 +136,9 @@ int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **f,
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_fence *fence;
-	struct dma_fence *old, **ptr;
+	struct dma_fence __rcu **ptr;
 	uint32_t seq;
+	int r;
 
 	fence = kmem_cache_alloc(amdgpu_fence_slab, GFP_KERNEL);
 	if (fence == NULL)
@@ -153,15 +154,24 @@ int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **f,
 			       seq, flags | AMDGPU_FENCE_FLAG_INT);
 
 	ptr = &ring->fence_drv.fences[seq & ring->fence_drv.num_fences_mask];
+	if (unlikely(rcu_dereference_protected(*ptr, 1))) {
+		struct dma_fence *old;
+
+		rcu_read_lock();
+		old = dma_fence_get_rcu_safe(ptr);
+		rcu_read_unlock();
+
+		if (old) {
+			r = dma_fence_wait(old, false);
+			dma_fence_put(old);
+			if (r)
+				return r;
+		}
+	}
+
 	/* This function can't be called concurrently anyway, otherwise
 	 * emitting the fence would mess up the hardware ring buffer.
 	 */
-	old = rcu_dereference_protected(*ptr, 1);
-	if (old && !dma_fence_is_signaled(old)) {
-		DRM_INFO("rcu slot is busy\n");
-		dma_fence_wait(old, false);
-	}
-
 	rcu_assign_pointer(*ptr, dma_fence_get(&fence->base));
 
 	*f = &fence->base;
Signature · 1 line
2.20.1
author maintainer
#210 lore raw

PATCH 210/405 PM / core: Propagate dev->power.wakeup_path when no callbacks

In reply to Greg Kroah-Hartman

[ Upstream commit dc351d4c5f4fe4d0f274d6d660227be0c3a03317 ]

The dev->power.direct_complete flag may become set in device_prepare() in case the device don't have any PM callbacks (dev->power.no_pm_callbacks is set). This leads to a broken behaviour, when there is child having wakeup enabled and relies on its parent to be used in the wakeup path.

More precisely, when the direct complete path becomes selected for the child in __device_suspend(), the propagation of the dev->power.wakeup_path becomes skipped as well.

Let's address this problem, by checking if the device is a part the wakeup path or has wakeup enabled, then prevent the direct complete path from being used.

---
 drivers/base/power/main.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/base/power/main.c +4 -0
index f80d298de3fa4..8ad20ed0cb7c3 100644
@@ -1747,6 +1747,10 @@ static int __device_suspend(struct device *dev, pm_message_t state, bool async)
 	if (dev->power.syscore)
 		goto Complete;
 
+	/* Avoid direct_complete to let wakeup_path propagate. */
+	if (device_may_wakeup(dev) || dev->power.wakeup_path)
+		dev->power.direct_complete = false;
+
 	if (dev->power.direct_complete) {
 		if (pm_runtime_status_suspended(dev)) {
 			pm_runtime_disable(dev);
Signature · 1 line
2.20.1
author maintainer
#211 lore raw

PATCH 211/405 clk: rockchip: Fix video codec clocks on rk3288

In reply to Greg Kroah-Hartman

[ Upstream commit 00c0cd9e59d265b393553e9afa54fee8b10e8158 ]

It appears that there is a typo in the rk3288 TRM. For GRF_SOC_CON0[7] it says that 0 means "vepu" and 1 means "vdpu". It's the other way around.

How do I know? Here's my evidence:

1. Prior to commit 4d3e84f99628 ("clk: rockchip: describe aclk_vcodec using the new muxgrf type on rk3288") we always pretended that we were using "aclk_vdpu" and the comment in the code said that this matched the default setting in the system. In fact the default setting is 0 according to the TRM and according to reading memory at bootup. In addition rk3288-based Chromebooks ran like this and the video codecs worked.
2. With the existing clock code if you boot up and try to enable the new VIDEO_ROCKCHIP_VPU as a module (and without "clk_ignore_unused" on the command line), you get errors like "failed to get ack on domain 'pd_video', val=0x80208". After flipping vepu/vdpu things init OK.
3. If I export and add both the vepu and vdpu to the list of clocks for RK3288_PD_VIDEO I can get past the power domain errors, but now I freeze when the vpu_mmu gets initted.
4. If I just mark the "vdpu" as IGNORE_UNUSED then everything boots up and probes OK showing that somehow the "vdpu" was important to keep enabled. This is because we were actually using it as a parent.
5. After this change I can hack "aclk_vcodec_pre" to parent from "aclk_vepu" using assigned-clocks and the video codec still probes OK.
6. Rockchip has said so on the mailing list [1].

...so let's fix it.

Let's also add CLK_SET_RATE_PARENT to "aclk_vcodec_pre" as suggested by Jonas Karlman. Prior to the same commit you could do clk_set_rate() on "aclk_vcodec" and it would change "aclk_vdpu". That's because "aclk_vcodec" was a simple gate clock (always gets CLK_SET_RATE_PARENT) and its direct parent was "aclk_vdpu". After that commit "aclk_vcodec_pre" gets in the way so we need to add CLK_SET_RATE_PARENT to it too.

[1] https://lkml.kernel.org/r/[email protected]

Fixes 4d3e84f99628 ("clk: rockchip: describe aclk_vcodec using the new muxgrf type on rk3288")
---
 drivers/clk/rockchip/clk-rk3288.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/clk/rockchip/clk-rk3288.c +2 -2
index f3bbcdfa88ead..623c5f684987c 100644
@@ -219,7 +219,7 @@ PNAME(mux_hsadcout_p)	= { "hsadc_src", "ext_hsadc" };
 PNAME(mux_edp_24m_p)	= { "ext_edp_24m", "xin24m" };
 PNAME(mux_tspout_p)	= { "cpll", "gpll", "npll", "xin27m" };
 
-PNAME(mux_aclk_vcodec_pre_p)	= { "aclk_vepu", "aclk_vdpu" };
+PNAME(mux_aclk_vcodec_pre_p)	= { "aclk_vdpu", "aclk_vepu" };
 PNAME(mux_usbphy480m_p)		= { "sclk_otgphy1_480m", "sclk_otgphy2_480m",
 				    "sclk_otgphy0_480m" };
 PNAME(mux_hsicphy480m_p)	= { "cpll", "gpll", "usbphy480m_src" };
@@ -420,7 +420,7 @@ static struct rockchip_clk_branch rk3288_clk_branches[] __initdata = {
 	COMPOSITE(0, "aclk_vdpu", mux_pll_src_cpll_gpll_usb480m_p, 0,
 			RK3288_CLKSEL_CON(32), 14, 2, MFLAGS, 8, 5, DFLAGS,
 			RK3288_CLKGATE_CON(3), 11, GFLAGS),
-	MUXGRF(0, "aclk_vcodec_pre", mux_aclk_vcodec_pre_p, 0,
+	MUXGRF(0, "aclk_vcodec_pre", mux_aclk_vcodec_pre_p, CLK_SET_RATE_PARENT,
 			RK3288_GRF_SOC_CON(0), 7, 1, MFLAGS),
 	GATE(ACLK_VCODEC, "aclk_vcodec", "aclk_vcodec_pre", 0,
 		RK3288_CLKGATE_CON(9), 0, GFLAGS),
Signature · 1 line
2.20.1
author maintainer
#212 lore raw

PATCH 212/405 extcon: arizona: Disable mic detect if running when driver is removed

In reply to Greg Kroah-Hartman

[ Upstream commit 00053de52231117ddc154042549f2256183ffb86 ]

Microphone detection provides the button detection features on the Arizona CODECs as such it will be running if the jack is currently inserted. If the driver is unbound whilst the jack is still inserted this will cause warnings from the regulator framework as the MICVDD regulator is put but was never disabled.

Correct this by disabling microphone detection on driver removal and if the microphone detection was running disable the regulator and put the runtime reference that was currently held.

---
 drivers/extcon/extcon-arizona.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
drivers/extcon/extcon-arizona.c +10 -0
index da0e9bc4262fa..9327479c719c2 100644
@@ -1726,6 +1726,16 @@ static int arizona_extcon_remove(struct platform_device *pdev)
 	struct arizona_extcon_info *info = platform_get_drvdata(pdev);
 	struct arizona *arizona = info->arizona;
 	int jack_irq_rise, jack_irq_fall;
+	bool change;
+
+	regmap_update_bits_check(arizona->regmap, ARIZONA_MIC_DETECT_1,
+				 ARIZONA_MICD_ENA, 0,
+				 &change);
+
+	if (change) {
+		regulator_disable(info->micvdd);
+		pm_runtime_put(info->dev);
+	}
 
 	gpiod_put(info->micd_pol_gpio);
 
Signature · 1 line
2.20.1
author maintainer
#213 lore raw

PATCH 213/405 clk: rockchip: Make rkpwm a critical clock on rk3288

In reply to Greg Kroah-Hartman

[ Upstream commit dfe7fb21cd9e730230d55a79bc72cf2ece67cdd5 ]

Most rk3288-based boards are derived from the EVB and thus use a PWM regulator for the logic rail. However, most rk3288-based boards don't specify the PWM regulator in their device tree. We'll deal with that by making it critical.

NOTE: it's important to make it critical and not just IGNORE_UNUSED because all PWMs in the system share the same clock. We don't want another PWM user to turn the clock on and off and kill the logic rail.

This change is in preparation for actually having the PWMs in the rk3288 device tree actually point to the proper PWM clock. Up until now they've all pointed to the clock for the old IP block and they've all worked due to the fact that rkpwm was IGNORE_UNUSED and that the clock rates for both clocks were the same.

---
 drivers/clk/rockchip/clk-rk3288.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/clk/rockchip/clk-rk3288.c +3 -1
index 623c5f684987c..355d6a3611dbf 100644
@@ -697,7 +697,7 @@ static struct rockchip_clk_branch rk3288_clk_branches[] __initdata = {
 	GATE(PCLK_TZPC, "pclk_tzpc", "pclk_cpu", 0, RK3288_CLKGATE_CON(11), 3, GFLAGS),
 	GATE(PCLK_UART2, "pclk_uart2", "pclk_cpu", 0, RK3288_CLKGATE_CON(11), 9, GFLAGS),
 	GATE(PCLK_EFUSE256, "pclk_efuse_256", "pclk_cpu", 0, RK3288_CLKGATE_CON(11), 10, GFLAGS),
-	GATE(PCLK_RKPWM, "pclk_rkpwm", "pclk_cpu", CLK_IGNORE_UNUSED, RK3288_CLKGATE_CON(11), 11, GFLAGS),
+	GATE(PCLK_RKPWM, "pclk_rkpwm", "pclk_cpu", 0, RK3288_CLKGATE_CON(11), 11, GFLAGS),
 
 	/* ddrctrl [DDR Controller PHY clock] gates */
 	GATE(0, "nclk_ddrupctl0", "ddrphy", CLK_IGNORE_UNUSED, RK3288_CLKGATE_CON(11), 4, GFLAGS),
@@ -838,6 +838,8 @@ static const char *const rk3288_critical_clocks[] __initconst = {
 	"pclk_pd_pmu",
 	"pclk_pmu_niu",
 	"pmu_hclk_otg0",
+	/* pwm-regulators on some boards, so handoff-critical later */
+	"pclk_rkpwm",
 };
 
 static void __iomem *rk3288_cru_base;
Signature · 1 line
2.20.1
author maintainer
#214 lore raw

PATCH 214/405 clk: zynqmp: fix check for fractional clock

In reply to Greg Kroah-Hartman

[ Upstream commit c06e64407e031e71c67f45f07981510ca4c880a1 ]

The firmware sets BIT(13) in clkflag to mark a divider as fractional divider. The clock driver copies the clkflag straight to the flags of the common clock framework. In the common clk framework flags, BIT(13) is defined as CLK_DUTY_CYCLE_PARENT.

Add a new field to the zynqmp_clk_divider to specify if a divider is a fractional devider. Set this field based on the clkflag when registering a divider.

At the same time, unset BIT(13) from clkflag when copying the flags to the common clk framework flags.

---
 drivers/clk/zynqmp/divider.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/clk/zynqmp/divider.c +6 -3
index a371c66e72ef6..bd9b5fbc443b3 100644
@@ -31,12 +31,14 @@
  * struct zynqmp_clk_divider - adjustable divider clock
  * @hw:		handle between common and hardware-specific interfaces
  * @flags:	Hardware specific flags
+ * @is_frac:	The divider is a fractional divider
  * @clk_id:	Id of clock
  * @div_type:	divisor type (TYPE_DIV1 or TYPE_DIV2)
  */
 struct zynqmp_clk_divider {
 	struct clk_hw hw;
 	u8 flags;
+	bool is_frac;
 	u32 clk_id;
 	u32 div_type;
 };
@@ -116,8 +118,7 @@ static long zynqmp_clk_divider_round_rate(struct clk_hw *hw,
 
 	bestdiv = zynqmp_divider_get_val(*prate, rate);
 
-	if ((clk_hw_get_flags(hw) & CLK_SET_RATE_PARENT) &&
-	    (divider->flags & CLK_FRAC))
+	if ((clk_hw_get_flags(hw) & CLK_SET_RATE_PARENT) && divider->is_frac)
 		bestdiv = rate % *prate ? 1 : bestdiv;
 	*prate = rate * bestdiv;
 
@@ -195,11 +196,13 @@ struct clk_hw *zynqmp_clk_register_divider(const char *name,
 
 	init.name = name;
 	init.ops = &zynqmp_clk_divider_ops;
-	init.flags = nodes->flag;
+	/* CLK_FRAC is not defined in the common clk framework */
+	init.flags = nodes->flag & ~CLK_FRAC;
 	init.parent_names = parents;
 	init.num_parents = 1;
 
 	/* struct clk_divider assignments */
+	div->is_frac = !!(nodes->flag & CLK_FRAC);
 	div->flags = nodes->type_flag;
 	div->hw.init = &init;
 	div->clk_id = clk_id;
Signature · 1 line
2.20.1
author maintainer
#215 lore raw

PATCH 215/405 s390: zcrypt: initialize variables before_use

In reply to Greg Kroah-Hartman

[ Upstream commit 913140e221567b3ecd21b4242257a7e3fa279026 ]

The 'func_code' variable gets printed in debug statements without a prior initialization in multiple functions, as reported when building with clang:

drivers/s390/crypto/zcrypt_api.c:659:6: warning: variable 'func_code' is used uninitialized whenever 'if' condition is true

      [-Wsometimes-uninitialized]
        if (mex->outputdatalength < mex->inputdatalength) {
            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/s390/crypto/zcrypt_api.c:725:29: note: uninitialized use occurs here
        trace_s390_zcrypt_rep(mex, func_code, rc,
                                   ^~~~~~~~~
drivers/s390/crypto/zcrypt_api.c:659:2: note: remove the 'if' if its condition is always false
        if (mex->outputdatalength < mex->inputdatalength) {
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/s390/crypto/zcrypt_api.c:654:24: note: initialize the variable 'func_code' to silence this warning
        unsigned int func_code;
                              ^

Add initializations to all affected code paths to shut up the warning and make the warning output consistent.

---
 drivers/s390/crypto/zcrypt_api.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/s390/crypto/zcrypt_api.c +4 -0
index 689c2af7026a3..c31b2d31cd832 100644
@@ -659,6 +659,7 @@ static long zcrypt_rsa_modexpo(struct ap_perms *perms,
 	trace_s390_zcrypt_req(mex, TP_ICARSAMODEXPO);
 
 	if (mex->outputdatalength < mex->inputdatalength) {
+		func_code = 0;
 		rc = -EINVAL;
 		goto out;
 	}
@@ -742,6 +743,7 @@ static long zcrypt_rsa_crt(struct ap_perms *perms,
 	trace_s390_zcrypt_req(crt, TP_ICARSACRT);
 
 	if (crt->outputdatalength < crt->inputdatalength) {
+		func_code = 0;
 		rc = -EINVAL;
 		goto out;
 	}
@@ -951,6 +953,7 @@ static long zcrypt_send_ep11_cprb(struct ap_perms *perms,
 
 		targets = kcalloc(target_num, sizeof(*targets), GFP_KERNEL);
 		if (!targets) {
+			func_code = 0;
 			rc = -ENOMEM;
 			goto out;
 		}
@@ -958,6 +961,7 @@ static long zcrypt_send_ep11_cprb(struct ap_perms *perms,
 		uptr = (struct ep11_target_dev __force __user *) xcrb->targets;
 		if (copy_from_user(targets, uptr,
 				   target_num * sizeof(*targets))) {
+			func_code = 0;
 			rc = -EFAULT;
 			goto out_free;
 		}
Signature · 1 line
2.20.1
author maintainer
#216 lore raw

PATCH 216/405 x86/microcode: Fix the ancient deprecated microcode loading method

In reply to Greg Kroah-Hartman

[ Upstream commit 24613a04ad1c0588c10f4b5403ca60a73d164051 ]

Commit

2613f36ed965 ("x86/microcode: Attempt late loading only when new microcode is present")

added the new define UCODE_NEW to denote that an update should happen only when newer microcode (than installed on the system) has been found.

But it missed adjusting that for the old /dev/cpu/microcode loading interface. Fix it.

Fixes 2613f36ed965 ("x86/microcode: Attempt late loading only when new microcode is present")
Cc Jann Horn <[email protected]>
---
 arch/x86/kernel/cpu/microcode/core.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/x86/kernel/cpu/microcode/core.c +2 -1
index 5260185cbf7ba..8a4a7823451ac 100644
@@ -418,8 +418,9 @@ static int do_microcode_update(const void __user *buf, size_t size)
 		if (ustate == UCODE_ERROR) {
 			error = -1;
 			break;
-		} else if (ustate == UCODE_OK)
+		} else if (ustate == UCODE_NEW) {
 			apply_microcode_on_target(cpu);
+		}
 	}
 
 	return error;
Signature · 1 line
2.20.1
author maintainer
#217 lore raw

PATCH 217/405 drm/amd/display: Initialize stream_update with memset

In reply to Greg Kroah-Hartman

[ Upstream commit 2aa632c5ffbedb2ee0e68857683466ea788f17eb ]

The brace initialization used here generates warnings on some compilers. For example, on GCC 4.9:

[...] In function ‘dm_determine_update_type_for_commit’:
[...] error: missing braces around initializer [-Werror=missing-braces]

   struct dc_stream_update stream_update = { 0 };
          ^

Use memset to make this more portable.

v2: Specify the compiler / diagnostic in the commit message (Paul)

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +3 -1
index 3082b55b1e774..66f19d1864b17 100644
@@ -5858,7 +5858,9 @@ dm_determine_update_type_for_commit(struct dc *dc,
 	}
 
 	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
-		struct dc_stream_update stream_update = { 0 };
+		struct dc_stream_update stream_update;
+
+		memset(&stream_update, 0, sizeof(stream_update));
 
 		new_dm_crtc_state = to_dm_crtc_state(new_crtc_state);
 		old_dm_crtc_state = to_dm_crtc_state(old_crtc_state);
Signature · 1 line
2.20.1
author maintainer
#218 lore raw

PATCH 218/405 s390/mm: silence compiler warning when compiling without CONFIG_PGSTE

In reply to Greg Kroah-Hartman

[ Upstream commit 81a8f2beb32a5951ecf04385301f50879abc092b ]

If CONFIG_PGSTE is not set (e.g. when compiling without KVM), GCC complains:

CC arch/s390/mm/pgtable.o
arch/s390/mm/pgtable.c:413:15: warning: ‘pmd_alloc_map’ defined but not
used [-Wunused-function]
static pmd_t *pmd_alloc_map(struct mm_struct *mm, unsigned long addr)

               ^~~~~~~~~~~~~

Wrap the function with "#ifdef CONFIG_PGSTE" to silence the warning.

---
 arch/s390/mm/pgtable.c | 2 ++
 1 file changed, 2 insertions(+)
arch/s390/mm/pgtable.c +2 -0
index 8485d6dc27549..9ebd01219812c 100644
@@ -410,6 +410,7 @@ static inline pmd_t pmdp_flush_lazy(struct mm_struct *mm,
 	return old;
 }
 
+#ifdef CONFIG_PGSTE
 static pmd_t *pmd_alloc_map(struct mm_struct *mm, unsigned long addr)
 {
 	pgd_t *pgd;
@@ -427,6 +428,7 @@ static pmd_t *pmd_alloc_map(struct mm_struct *mm, unsigned long addr)
 	pmd = pmd_alloc(mm, pud, addr);
 	return pmd;
 }
+#endif
 
 pmd_t pmdp_xchg_direct(struct mm_struct *mm, unsigned long addr,
 		       pmd_t *pmdp, pmd_t new)
Signature · 1 line
2.20.1
author maintainer
#219 lore raw

PATCH 219/405 s390: cio: fix cio_irb declaration

In reply to Greg Kroah-Hartman

[ Upstream commit e91012ee855ad9f5ef2ab106a3de51db93fe4d0c ]

clang points out that the declaration of cio_irb does not match the definition exactly, it is missing the alignment attribute:

../drivers/s390/cio/cio.c:50:1: warning: section does not match previous declaration [-Wsection] DEFINE_PER_CPU_ALIGNED(struct irb, cio_irb);
^
../include/linux/percpu-defs.h:150:2: note: expanded from macro 'DEFINE_PER_CPU_ALIGNED'

        DEFINE_PER_CPU_SECTION(type, name, PER_CPU_ALIGNED_SECTION)     \
        ^
../include/linux/percpu-defs.h:93:9: note: expanded from macro 'DEFINE_PER_CPU_SECTION'
        extern __PCPU_ATTRS(sec) __typeof__(type) name;                 \
               ^
../include/linux/percpu-defs.h:49:26: note: expanded from macro '__PCPU_ATTRS'
        __percpu __attribute__((section(PER_CPU_BASE_SECTION sec)))     \
                                ^

../drivers/s390/cio/cio.h:118:1: note: previous attribute is here DECLARE_PER_CPU(struct irb, cio_irb);
^
../include/linux/percpu-defs.h:111:2: note: expanded from macro 'DECLARE_PER_CPU'
DECLARE_PER_CPU_SECTION(type, name, "")
^
../include/linux/percpu-defs.h:87:9: note: expanded from macro 'DECLARE_PER_CPU_SECTION'
extern __PCPU_ATTRS(sec) __typeof__(type) name
^
../include/linux/percpu-defs.h:49:26: note: expanded from macro '__PCPU_ATTRS'

        __percpu __attribute__((section(PER_CPU_BASE_SECTION sec)))     \
                                ^

Use DECLARE_PER_CPU_ALIGNED() here, to make the two match.

---
 drivers/s390/cio/cio.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/s390/cio/cio.h +1 -1
index 9811fd8a0c731..92eabbb5f18d4 100644
@@ -115,7 +115,7 @@ struct subchannel {
 	struct schib_config config;
 } __attribute__ ((aligned(8)));
 
-DECLARE_PER_CPU(struct irb, cio_irb);
+DECLARE_PER_CPU_ALIGNED(struct irb, cio_irb);
 
 #define to_subchannel(n) container_of(n, struct subchannel, dev)
 
Signature · 1 line
2.20.1
author maintainer
#220 lore raw

PATCH 220/405 drm/amd/display: use proper formula to calculate bandwidth from timing

In reply to Greg Kroah-Hartman

[ Upstream commit e49f69363adf8920883fff7e8ffecb802d897c6b ]

[why]
The existing calculation uses a wrong formula to
calculate bandwidth from timing.

[how]
Expose the existing proper function that calculates the bandwidth, so dc_link can use it to calculate timing bandwidth correctly.

---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c | 48 ++++++++++++++++-
 .../gpu/drm/amd/display/dc/core/dc_link_dp.c  | 51 +------------------
 drivers/gpu/drm/amd/display/dc/dc_link.h      |  2 +
 3 files changed, 51 insertions(+), 50 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +47 -1
index ea18e9c2d8cea..ea2f271e234bd 100644
@@ -2316,7 +2316,7 @@ static struct fixed31_32 get_pbn_from_timing(struct pipe_ctx *pipe_ctx)
 	uint32_t denominator;
 
 	bpc = get_color_depth(pipe_ctx->stream_res.pix_clk_params.color_depth);
-	kbps = pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz / 10 * bpc * 3;
+	kbps = dc_bandwidth_in_kbps_from_timing(&pipe_ctx->stream->timing);
 
 	/*
 	 * margin 5300ppm + 300ppm ~ 0.6% as per spec, factor is 1.006
@@ -2736,3 +2736,49 @@ void dc_link_enable_hpd_filter(struct dc_link *link, bool enable)
 	}
 }
 
+uint32_t dc_bandwidth_in_kbps_from_timing(
+	const struct dc_crtc_timing *timing)
+{
+	uint32_t bits_per_channel = 0;
+	uint32_t kbps;
+
+	switch (timing->display_color_depth) {
+	case COLOR_DEPTH_666:
+		bits_per_channel = 6;
+		break;
+	case COLOR_DEPTH_888:
+		bits_per_channel = 8;
+		break;
+	case COLOR_DEPTH_101010:
+		bits_per_channel = 10;
+		break;
+	case COLOR_DEPTH_121212:
+		bits_per_channel = 12;
+		break;
+	case COLOR_DEPTH_141414:
+		bits_per_channel = 14;
+		break;
+	case COLOR_DEPTH_161616:
+		bits_per_channel = 16;
+		break;
+	default:
+		break;
+	}
+
+	ASSERT(bits_per_channel != 0);
+
+	kbps = timing->pix_clk_100hz / 10;
+	kbps *= bits_per_channel;
+
+	if (timing->flags.Y_ONLY != 1) {
+		/*Only YOnly make reduce bandwidth by 1/3 compares to RGB*/
+		kbps *= 3;
+		if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420)
+			kbps /= 2;
+		else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR422)
+			kbps = kbps * 2 / 3;
+	}
+
+	return kbps;
+
+}
drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +2 -49

69 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dc_link.h +2 -0
index 8fc223defed4a..a83e1c60f9db2 100644
@@ -252,4 +252,6 @@ bool dc_submit_i2c(
 		uint32_t link_index,
 		struct i2c_command *cmd);
 
+uint32_t dc_bandwidth_in_kbps_from_timing(
+	const struct dc_crtc_timing *timing);
 #endif /* DC_LINK_H_ */
Signature · 1 line
2.20.1
author maintainer
#221 lore raw

PATCH 221/405 selftests: cgroup: fix cleanup path in test_memcg_subtree_control()

In reply to Greg Kroah-Hartman

[ Upstream commit e14d314c7a489f060d6d691866fef5f131281718 ]

Dan reported, that cleanup path in test_memcg_subtree_control() triggers a static checker warning:
./tools/testing/selftests/cgroup/test_memcontrol.c:76 \
test_memcg_subtree_control()
error: uninitialized symbol 'child2'.

Fix this by initializing child2 and parent2 variables and split the cleanup path into few stages.

Fixes 84092dbcf901 ("selftests: cgroup: add memory controller self-tests")
Cc Dan Carpenter <[email protected]>
Cc Shuah Khan (Samsung OSG) <[email protected]>
Cc Mike Rapoport <[email protected]>
---
 .../selftests/cgroup/test_memcontrol.c        | 38 ++++++++++---------
 1 file changed, 21 insertions(+), 17 deletions(-)
tools/testing/selftests/cgroup/test_memcontrol.c +21 -17
index 28d321ba311b4..6f339882a6ca1 100644
@@ -26,7 +26,7 @@
  */
 static int test_memcg_subtree_control(const char *root)
 {
-	char *parent, *child, *parent2, *child2;
+	char *parent, *child, *parent2 = NULL, *child2 = NULL;
 	int ret = KSFT_FAIL;
 	char buf[PAGE_SIZE];
 
@@ -34,50 +34,54 @@ static int test_memcg_subtree_control(const char *root)
 	parent = cg_name(root, "memcg_test_0");
 	child = cg_name(root, "memcg_test_0/memcg_test_1");
 	if (!parent || !child)
-		goto cleanup;
+		goto cleanup_free;
 
 	if (cg_create(parent))
-		goto cleanup;
+		goto cleanup_free;
 
 	if (cg_write(parent, "cgroup.subtree_control", "+memory"))
-		goto cleanup;
+		goto cleanup_parent;
 
 	if (cg_create(child))
-		goto cleanup;
+		goto cleanup_parent;
 
 	if (cg_read_strstr(child, "cgroup.controllers", "memory"))
-		goto cleanup;
+		goto cleanup_child;
 
 	/* Create two nested cgroups without enabling memory controller */
 	parent2 = cg_name(root, "memcg_test_1");
 	child2 = cg_name(root, "memcg_test_1/memcg_test_1");
 	if (!parent2 || !child2)
-		goto cleanup;
+		goto cleanup_free2;
 
 	if (cg_create(parent2))
-		goto cleanup;
+		goto cleanup_free2;
 
 	if (cg_create(child2))
-		goto cleanup;
+		goto cleanup_parent2;
 
 	if (cg_read(child2, "cgroup.controllers", buf, sizeof(buf)))
-		goto cleanup;
+		goto cleanup_all;
 
 	if (!cg_read_strstr(child2, "cgroup.controllers", "memory"))
-		goto cleanup;
+		goto cleanup_all;
 
 	ret = KSFT_PASS;
 
-cleanup:
-	cg_destroy(child);
-	cg_destroy(parent);
-	free(parent);
-	free(child);
-
+cleanup_all:
 	cg_destroy(child2);
+cleanup_parent2:
 	cg_destroy(parent2);
+cleanup_free2:
 	free(parent2);
 	free(child2);
+cleanup_child:
+	cg_destroy(child);
+cleanup_parent:
+	cg_destroy(parent);
+cleanup_free:
+	free(parent);
+	free(child);
 
 	return ret;
 }
Signature · 1 line
2.20.1
author maintainer
#222 lore raw

PATCH 222/405 net: hns3: fix keep_alive_timer not stop problem

In reply to Greg Kroah-Hartman

[ Upstream commit e233516e6a92baeec20aa40fa5b63be6b94f1627 ]

When hclgevf_client_start() fails or VF driver unloaded, there is nobody to disable keep_alive_timer.

So this patch fixes them.

Fixes a6d818e31d08 ("net: hns3: Add vport alive state checking support")
---
 .../ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c    | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c +11 -1
index 8bc28e6f465f1..8dd7fef863f68 100644
@@ -2007,9 +2007,15 @@ static int hclgevf_set_alive(struct hnae3_handle *handle, bool alive)
 static int hclgevf_client_start(struct hnae3_handle *handle)
 {
 	struct hclgevf_dev *hdev = hclgevf_ae_get_hdev(handle);
+	int ret;
+
+	ret = hclgevf_set_alive(handle, true);
+	if (ret)
+		return ret;
 
 	mod_timer(&hdev->keep_alive_timer, jiffies + 2 * HZ);
-	return hclgevf_set_alive(handle, true);
+
+	return 0;
 }
 
 static void hclgevf_client_stop(struct hnae3_handle *handle)
@@ -2051,6 +2057,10 @@ static void hclgevf_state_uninit(struct hclgevf_dev *hdev)
 {
 	set_bit(HCLGEVF_STATE_DOWN, &hdev->state);
 
+	if (hdev->keep_alive_timer.function)
+		del_timer_sync(&hdev->keep_alive_timer);
+	if (hdev->keep_alive_task.func)
+		cancel_work_sync(&hdev->keep_alive_task);
 	if (hdev->service_timer.function)
 		del_timer_sync(&hdev->service_timer);
 	if (hdev->service_task.func)
Signature · 1 line
2.20.1
author maintainer
#223 lore raw

PATCH 223/405 qmi_wwan: Add quirk for Quectel dynamic config

In reply to Greg Kroah-Hartman

[ Upstream commit e4bf63482c309287ca84d91770ffa7dcc18e37eb ]

Most, if not all, Quectel devices use dynamic interface numbers, and users are able to change the USB configuration at will. Matching on for example interface number is therefore not possible.

Instead, the QMI device can be identified by looking at the interface class, subclass and protocol (all 0xff), as well as the number of endpoints. The reason we need to look at the number of endpoints, is that the diagnostic port interface has the same class, subclass and protocol as QMI. However, the diagnostic port only has two endpoints, while QMI has three.

Until now, we have identified the QMI device by combining a match on class, subclass and protocol, with a call to the function quectel_diag_detect(). In quectel_diag_detect(), we check if the number of endpoints matches for known Quectel vendor/product ids.

Adding new vendor/product ids to quectel_diag_detect() is not a good long-term solution. This commit replaces the function with a quirk, and applies the quirk to affected Quectel devices that I have been able to test the change with (EP06, EM12 and EC25). If the quirk is set and the number of endpoints equal two, we return from qmi_wwan_probe() with -ENODEV.

---
 drivers/net/usb/qmi_wwan.c | 65 ++++++++++++++++++--------------------
 1 file changed, 31 insertions(+), 34 deletions(-)
drivers/net/usb/qmi_wwan.c +31 -34
index 366217263d704..d9a6699abe592 100644
@@ -63,6 +63,7 @@ enum qmi_wwan_flags {
 
 enum qmi_wwan_quirks {
 	QMI_WWAN_QUIRK_DTR = 1 << 0,	/* needs "set DTR" request */
+	QMI_WWAN_QUIRK_QUECTEL_DYNCFG = 1 << 1,	/* check num. endpoints */
 };
 
 struct qmimux_hdr {
@@ -845,6 +846,16 @@ static const struct driver_info	qmi_wwan_info_quirk_dtr = {
 	.data           = QMI_WWAN_QUIRK_DTR,
 };
 
+static const struct driver_info	qmi_wwan_info_quirk_quectel_dyncfg = {
+	.description	= "WWAN/QMI device",
+	.flags		= FLAG_WWAN | FLAG_SEND_ZLP,
+	.bind		= qmi_wwan_bind,
+	.unbind		= qmi_wwan_unbind,
+	.manage_power	= qmi_wwan_manage_power,
+	.rx_fixup       = qmi_wwan_rx_fixup,
+	.data           = QMI_WWAN_QUIRK_DTR | QMI_WWAN_QUIRK_QUECTEL_DYNCFG,
+};
+
 #define HUAWEI_VENDOR_ID	0x12D1
 
 /* map QMI/wwan function by a fixed interface number */
@@ -865,6 +876,15 @@ static const struct driver_info	qmi_wwan_info_quirk_dtr = {
 #define QMI_GOBI_DEVICE(vend, prod) \
 	QMI_FIXED_INTF(vend, prod, 0)
 
+/* Quectel does not use fixed interface numbers on at least some of their
+ * devices. We need to check the number of endpoints to ensure that we bind to
+ * the correct interface.
+ */
+#define QMI_QUIRK_QUECTEL_DYNCFG(vend, prod) \
+	USB_DEVICE_AND_INTERFACE_INFO(vend, prod, USB_CLASS_VENDOR_SPEC, \
+				      USB_SUBCLASS_VENDOR_SPEC, 0xff), \
+	.driver_info = (unsigned long)&qmi_wwan_info_quirk_quectel_dyncfg
+
 static const struct usb_device_id products[] = {
 	/* 1. CDC ECM like devices match on the control interface */
 	{	/* Huawei E392, E398 and possibly others sharing both device id and more... */
@@ -969,20 +989,9 @@ static const struct usb_device_id products[] = {
 		USB_DEVICE_AND_INTERFACE_INFO(0x03f0, 0x581d, USB_CLASS_VENDOR_SPEC, 1, 7),
 		.driver_info = (unsigned long)&qmi_wwan_info,
 	},
-	{	/* Quectel EP06/EG06/EM06 */
-		USB_DEVICE_AND_INTERFACE_INFO(0x2c7c, 0x0306,
-					      USB_CLASS_VENDOR_SPEC,
-					      USB_SUBCLASS_VENDOR_SPEC,
-					      0xff),
-		.driver_info	    = (unsigned long)&qmi_wwan_info_quirk_dtr,
-	},
-	{	/* Quectel EG12/EM12 */
-		USB_DEVICE_AND_INTERFACE_INFO(0x2c7c, 0x0512,
-					      USB_CLASS_VENDOR_SPEC,
-					      USB_SUBCLASS_VENDOR_SPEC,
-					      0xff),
-		.driver_info	    = (unsigned long)&qmi_wwan_info_quirk_dtr,
-	},
+	{QMI_QUIRK_QUECTEL_DYNCFG(0x2c7c, 0x0125)},	/* Quectel EC25, EC20 R2.0  Mini PCIe */
+	{QMI_QUIRK_QUECTEL_DYNCFG(0x2c7c, 0x0306)},	/* Quectel EP06/EG06/EM06 */
+	{QMI_QUIRK_QUECTEL_DYNCFG(0x2c7c, 0x0512)},	/* Quectel EG12/EM12 */
 
 	/* 3. Combined interface devices matching on interface number */
 	{QMI_FIXED_INTF(0x0408, 0xea42, 4)},	/* Yota / Megafon M100-1 */
@@ -1283,7 +1292,6 @@ static const struct usb_device_id products[] = {
 	{QMI_FIXED_INTF(0x03f0, 0x9d1d, 1)},	/* HP lt4120 Snapdragon X5 LTE */
 	{QMI_FIXED_INTF(0x22de, 0x9061, 3)},	/* WeTelecom WPD-600N */
 	{QMI_QUIRK_SET_DTR(0x1e0e, 0x9001, 5)},	/* SIMCom 7100E, 7230E, 7600E ++ */
-	{QMI_QUIRK_SET_DTR(0x2c7c, 0x0125, 4)},	/* Quectel EC25, EC20 R2.0  Mini PCIe */
 	{QMI_QUIRK_SET_DTR(0x2c7c, 0x0121, 4)},	/* Quectel EC21 Mini PCIe */
 	{QMI_QUIRK_SET_DTR(0x2c7c, 0x0191, 4)},	/* Quectel EG91 */
 	{QMI_FIXED_INTF(0x2c7c, 0x0296, 4)},	/* Quectel BG96 */
@@ -1363,27 +1371,12 @@ static bool quectel_ec20_detected(struct usb_interface *intf)
 	return false;
 }
 
-static bool quectel_diag_detected(struct usb_interface *intf)
-{
-	struct usb_device *dev = interface_to_usbdev(intf);
-	struct usb_interface_descriptor intf_desc = intf->cur_altsetting->desc;
-	u16 id_vendor = le16_to_cpu(dev->descriptor.idVendor);
-	u16 id_product = le16_to_cpu(dev->descriptor.idProduct);
-
-	if (id_vendor != 0x2c7c || intf_desc.bNumEndpoints != 2)
-		return false;
-
-	if (id_product == 0x0306 || id_product == 0x0512)
-		return true;
-	else
-		return false;
-}
-
 static int qmi_wwan_probe(struct usb_interface *intf,
 			  const struct usb_device_id *prod)
 {
 	struct usb_device_id *id = (struct usb_device_id *)prod;
 	struct usb_interface_descriptor *desc = &intf->cur_altsetting->desc;
+	const struct driver_info *info;
 
 	/* Workaround to enable dynamic IDs.  This disables usbnet
 	 * blacklisting functionality.  Which, if required, can be
@@ -1417,10 +1410,14 @@ static int qmi_wwan_probe(struct usb_interface *intf,
 	 * we need to match on class/subclass/protocol. These values are
 	 * identical for the diagnostic- and QMI-interface, but bNumEndpoints is
 	 * different. Ignore the current interface if the number of endpoints
-	 * the number for the diag interface (two).
+	 * equals the number for the diag interface (two).
 	 */
-	if (quectel_diag_detected(intf))
-		return -ENODEV;
+	info = (void *)&id->driver_info;
+
+	if (info->data & QMI_WWAN_QUIRK_QUECTEL_DYNCFG) {
+		if (desc->bNumEndpoints == 2)
+			return -ENODEV;
+	}
 
 	return usbnet_probe(intf, id);
 }
Signature · 1 line
2.20.1
author maintainer
#224 lore raw

PATCH 224/405 net: hns3: add error handler for initializing command queue

In reply to Greg Kroah-Hartman

[ Upstream commit 4339ef396ab65a61f7f22f36d7ba94b6e9e0939b ]

This patch adds error handler for the failure of command queue initialization both PF and VF.

---
 .../net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c    | 11 ++++++++---
 .../net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c  | 11 ++++++++---
 2 files changed, 16 insertions(+), 6 deletions(-)
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.c +8 -3
index 3a093a92eac51..d92e4af11b1fe 100644
@@ -373,21 +373,26 @@ int hclge_cmd_init(struct hclge_dev *hdev)
 	 * reset may happen when lower level reset is being processed.
 	 */
 	if ((hclge_is_reset_pending(hdev))) {
-		set_bit(HCLGE_STATE_CMD_DISABLE, &hdev->state);
-		return -EBUSY;
+		ret = -EBUSY;
+		goto err_cmd_init;
 	}
 
 	ret = hclge_cmd_query_firmware_version(&hdev->hw, &version);
 	if (ret) {
 		dev_err(&hdev->pdev->dev,
 			"firmware version query failed %d\n", ret);
-		return ret;
+		goto err_cmd_init;
 	}
 	hdev->fw_version = version;
 
 	dev_info(&hdev->pdev->dev, "The firmware version is %08x\n", version);
 
 	return 0;
+
+err_cmd_init:
+	set_bit(HCLGE_STATE_CMD_DISABLE, &hdev->state);
+
+	return ret;
 }
 
 static void hclge_cmd_uninit_regs(struct hclge_hw *hw)
drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_cmd.c +8 -3
index 9a0a501908aec..382ecb15e7435 100644
@@ -344,8 +344,8 @@ int hclgevf_cmd_init(struct hclgevf_dev *hdev)
 	 * reset may happen when lower level reset is being processed.
 	 */
 	if (hclgevf_is_reset_pending(hdev)) {
-		set_bit(HCLGEVF_STATE_CMD_DISABLE, &hdev->state);
-		return -EBUSY;
+		ret = -EBUSY;
+		goto err_cmd_init;
 	}
 
 	/* get firmware version */
@@ -353,13 +353,18 @@ int hclgevf_cmd_init(struct hclgevf_dev *hdev)
 	if (ret) {
 		dev_err(&hdev->pdev->dev,
 			"failed(%d) to query firmware version\n", ret);
-		return ret;
+		goto err_cmd_init;
 	}
 	hdev->fw_version = version;
 
 	dev_info(&hdev->pdev->dev, "The firmware version is %08x\n", version);
 
 	return 0;
+
+err_cmd_init:
+	set_bit(HCLGEVF_STATE_CMD_DISABLE, &hdev->state);
+
+	return ret;
 }
 
 static void hclgevf_cmd_uninit_regs(struct hclgevf_hw *hw)
Signature · 1 line
2.20.1
author maintainer
#225 lore raw

PATCH 225/405 cpufreq: ppc_cbe: fix possible object reference leak

In reply to Greg Kroah-Hartman

[ Upstream commit 233298032803f2802fe99892d0de4ab653bfece4 ]

The call to of_get_cpu_node returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./drivers/cpufreq/ppc_cbe_cpufreq.c:89:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 76, but without a corresponding object release within this function. ./drivers/cpufreq/ppc_cbe_cpufreq.c:89:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 76, but without a corresponding object release within this function.

---
 drivers/cpufreq/ppc_cbe_cpufreq.c | 1 +
 1 file changed, 1 insertion(+)
drivers/cpufreq/ppc_cbe_cpufreq.c +1 -0
index 41a0f0be3f9ff..8414c3a4ea08c 100644
@@ -86,6 +86,7 @@ static int cbe_cpufreq_cpu_init(struct cpufreq_policy *policy)
 	if (!cbe_get_cpu_pmd_regs(policy->cpu) ||
 	    !cbe_get_cpu_mic_tm_regs(policy->cpu)) {
 		pr_info("invalid CBE regs pointers for cpufreq\n");
+		of_node_put(cpu);
 		return -EINVAL;
 	}
 
Signature · 1 line
2.20.1
author maintainer
#226 lore raw

PATCH 226/405 cpufreq/pasemi: fix possible object reference leak

In reply to Greg Kroah-Hartman

[ Upstream commit a9acc26b75f652f697e02a9febe2ab0da648a571 ]

The call to of_get_cpu_node returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./drivers/cpufreq/pasemi-cpufreq.c:212:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 147, but without a corresponding object release within this function. ./drivers/cpufreq/pasemi-cpufreq.c:220:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 147, but without a corresponding object release within this function.

---
 drivers/cpufreq/pasemi-cpufreq.c | 1 +
 1 file changed, 1 insertion(+)
drivers/cpufreq/pasemi-cpufreq.c +1 -0
index 75dfbd2a58ea6..c7710c149de85 100644
@@ -146,6 +146,7 @@ static int pas_cpufreq_cpu_init(struct cpufreq_policy *policy)
 
 	cpu = of_get_cpu_node(policy->cpu, NULL);
 
+	of_node_put(cpu);
 	if (!cpu)
 		goto out;
 
Signature · 1 line
2.20.1
author maintainer
#227 lore raw

PATCH 227/405 cpufreq: pmac32: fix possible object reference leak

In reply to Greg Kroah-Hartman

[ Upstream commit 8d10dc28a9ea6e8c02e825dab28699f3c72b02d9 ]

The call to of_find_node_by_name returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./drivers/cpufreq/pmac32-cpufreq.c:557:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 552, but without a corresponding object release within this function. ./drivers/cpufreq/pmac32-cpufreq.c:569:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 552, but without a corresponding object release within this function. ./drivers/cpufreq/pmac32-cpufreq.c:598:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 587, but without a corresponding object release within this function.

Cc "Rafael J. Wysocki" <[email protected]>
Cc Viresh Kumar <[email protected]>
Cc Benjamin Herrenschmidt <[email protected]>
Cc Paul Mackerras <[email protected]>
Cc Michael Ellerman <[email protected]>
---
 drivers/cpufreq/pmac32-cpufreq.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/cpufreq/pmac32-cpufreq.c +2 -0
index 52f0d91d30c17..9b4ce2eb8222c 100644
@@ -552,6 +552,7 @@ static int pmac_cpufreq_init_7447A(struct device_node *cpunode)
 	volt_gpio_np = of_find_node_by_name(NULL, "cpu-vcore-select");
 	if (volt_gpio_np)
 		voltage_gpio = read_gpio(volt_gpio_np);
+	of_node_put(volt_gpio_np);
 	if (!voltage_gpio){
 		pr_err("missing cpu-vcore-select gpio\n");
 		return 1;
@@ -588,6 +589,7 @@ static int pmac_cpufreq_init_750FX(struct device_node *cpunode)
 	if (volt_gpio_np)
 		voltage_gpio = read_gpio(volt_gpio_np);
 
+	of_node_put(volt_gpio_np);
 	pvr = mfspr(SPRN_PVR);
 	has_cpu_l2lve = !((pvr & 0xf00) == 0x100);
 
Signature · 1 line
2.20.1
author maintainer
#228 lore raw

PATCH 228/405 cpufreq: kirkwood: fix possible object reference leak

In reply to Greg Kroah-Hartman

[ Upstream commit 7c468966f05ac9c17bb5948275283d34e6fe0660 ]

The call to of_get_child_by_name returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./drivers/cpufreq/kirkwood-cpufreq.c:127:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 118, but without a corresponding object release within this function. ./drivers/cpufreq/kirkwood-cpufreq.c:133:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 118, but without a corresponding object release within this function.

and also do some cleanup:
- of_node_put(np);
- np = NULL;
...
of_node_put(np);

---
 drivers/cpufreq/kirkwood-cpufreq.c | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)
drivers/cpufreq/kirkwood-cpufreq.c +11 -8
index c2dd43f3f5d8a..8d63a6dc8383c 100644
@@ -124,13 +124,14 @@ static int kirkwood_cpufreq_probe(struct platform_device *pdev)
 	priv.cpu_clk = of_clk_get_by_name(np, "cpu_clk");
 	if (IS_ERR(priv.cpu_clk)) {
 		dev_err(priv.dev, "Unable to get cpuclk\n");
-		return PTR_ERR(priv.cpu_clk);
+		err = PTR_ERR(priv.cpu_clk);
+		goto out_node;
 	}
 
 	err = clk_prepare_enable(priv.cpu_clk);
 	if (err) {
 		dev_err(priv.dev, "Unable to prepare cpuclk\n");
-		return err;
+		goto out_node;
 	}
 
 	kirkwood_freq_table[0].frequency = clk_get_rate(priv.cpu_clk) / 1000;
@@ -161,20 +162,22 @@ static int kirkwood_cpufreq_probe(struct platform_device *pdev)
 		goto out_ddr;
 	}
 
-	of_node_put(np);
-	np = NULL;
-
 	err = cpufreq_register_driver(&kirkwood_cpufreq_driver);
-	if (!err)
-		return 0;
+	if (err) {
+		dev_err(priv.dev, "Failed to register cpufreq driver\n");
+		goto out_powersave;
+	}
 
-	dev_err(priv.dev, "Failed to register cpufreq driver\n");
+	of_node_put(np);
+	return 0;
 
+out_powersave:
 	clk_disable_unprepare(priv.powersave_clk);
 out_ddr:
 	clk_disable_unprepare(priv.ddr_clk);
 out_cpu:
 	clk_disable_unprepare(priv.cpu_clk);
+out_node:
 	of_node_put(np);
 
 	return err;
Signature · 1 line
2.20.1
author maintainer
#229 lore raw

PATCH 229/405 cpufreq: imx6q: fix possible object reference leak

In reply to Greg Kroah-Hartman

[ Upstream commit ddb64c5db3cc8fb9c1242214d5798b2c2865681c ]

The call to of_node_get returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./drivers/cpufreq/imx6q-cpufreq.c:391:4-10: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 348, but without a corresponding object release within this function. ./drivers/cpufreq/imx6q-cpufreq.c:395:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 348, but without a corresponding object release within this function.

Cc "Rafael J. Wysocki" <[email protected]>
Cc Viresh Kumar <[email protected]>
Cc Shawn Guo <[email protected]>
Cc Sascha Hauer <[email protected]>
Cc Pengutronix Kernel Team <[email protected]>
Cc Fabio Estevam <[email protected]>
Cc NXP Linux Team <[email protected]>
---
 drivers/cpufreq/imx6q-cpufreq.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/cpufreq/imx6q-cpufreq.c +2 -2
index a4ff09f91c8f8..3e17560b1efe3 100644
@@ -388,11 +388,11 @@ static int imx6q_cpufreq_probe(struct platform_device *pdev)
 		ret = imx6ul_opp_check_speed_grading(cpu_dev);
 		if (ret) {
 			if (ret == -EPROBE_DEFER)
-				return ret;
+				goto put_node;
 
 			dev_err(cpu_dev, "failed to read ocotp: %d\n",
 				ret);
-			return ret;
+			goto put_node;
 		}
 	} else {
 		imx6q_opp_check_speed_grading(cpu_dev);
Signature · 9 lines
2.20.1




_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
author maintainer
#230 lore raw

PATCH 230/405 cpufreq: ap806: fix possible object reference leak

In reply to Greg Kroah-Hartman

[ Upstream commit b623fa320f8360f049a6f3c3ccc487cb85af4c5b ]

The call to of_find_compatible_node returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./drivers/cpufreq/armada-8k-cpufreq.c:187:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 130, but without a corresponding object release within this function. ./drivers/cpufreq/armada-8k-cpufreq.c:191:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 130, but without a corresponding object release within this function.

Cc Jason Cooper <[email protected]>
Cc Andrew Lunn <[email protected]>
Cc Gregory Clement <[email protected]>
Cc Sebastian Hesselbarth <[email protected]>
Cc "Rafael J. Wysocki" <[email protected]>
Cc Viresh Kumar <[email protected]>
---
 drivers/cpufreq/armada-8k-cpufreq.c | 1 +
 1 file changed, 1 insertion(+)
drivers/cpufreq/armada-8k-cpufreq.c +1 -0
index b3f4bd647e9b3..988ebc326bdbb 100644
@@ -132,6 +132,7 @@ static int __init armada_8k_cpufreq_init(void)
 		of_node_put(node);
 		return -ENODEV;
 	}
+	of_node_put(node);
 
 	nb_cpus = num_possible_cpus();
 	freq_tables = kcalloc(nb_cpus, sizeof(*freq_tables), GFP_KERNEL);
Signature · 9 lines
2.20.1




_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
author maintainer
#231 lore raw

PATCH 231/405 block: sed-opal: fix IOC_OPAL_ENABLE_DISABLE_MBR

In reply to Greg Kroah-Hartman

[ Upstream commit 78bf47353b0041865564deeed257a54f047c2fdc ]

The implementation of IOC_OPAL_ENABLE_DISABLE_MBR handled the value opal_mbr_data.enable_disable incorrectly: enable_disable is expected to be one of OPAL_MBR_ENABLE(0) or OPAL_MBR_DISABLE(1). enable_disable was passed directly to set_mbr_done and set_mbr_enable_disable where is was interpreted as either OPAL_TRUE(1) or OPAL_FALSE(0). The end result was that calling IOC_OPAL_ENABLE_DISABLE_MBR with OPAL_MBR_ENABLE actually disabled the shadow MBR and vice versa.

This patch adds correct conversion from OPAL_MBR_DISABLE/ENABLE to OPAL_FALSE/TRUE. The change affects existing programs using IOC_OPAL_ENABLE_DISABLE_MBR but this is typically used only once when setting up an Opal drive.

---
 block/sed-opal.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
block/sed-opal.c +6 -3
index e0de4dd448b3c..1196408972937 100644
@@ -2095,13 +2095,16 @@ static int opal_erase_locking_range(struct opal_dev *dev,
 static int opal_enable_disable_shadow_mbr(struct opal_dev *dev,
 					  struct opal_mbr_data *opal_mbr)
 {
+	u8 enable_disable = opal_mbr->enable_disable == OPAL_MBR_ENABLE ?
+		OPAL_TRUE : OPAL_FALSE;
+
 	const struct opal_step mbr_steps[] = {
 		{ opal_discovery0, },
 		{ start_admin1LSP_opal_session, &opal_mbr->key },
-		{ set_mbr_done, &opal_mbr->enable_disable },
+		{ set_mbr_done, &enable_disable },
 		{ end_opal_session, },
 		{ start_admin1LSP_opal_session, &opal_mbr->key },
-		{ set_mbr_enable_disable, &opal_mbr->enable_disable },
+		{ set_mbr_enable_disable, &enable_disable },
 		{ end_opal_session, },
 		{ NULL, }
 	};
@@ -2221,7 +2224,7 @@ static int __opal_lock_unlock(struct opal_dev *dev,
 
 static int __opal_set_mbr_done(struct opal_dev *dev, struct opal_key *key)
 {
-	u8 mbr_done_tf = 1;
+	u8 mbr_done_tf = OPAL_TRUE;
 	const struct opal_step mbrdone_step [] = {
 		{ opal_discovery0, },
 		{ start_admin1LSP_opal_session, key },
Signature · 1 line
2.20.1
author maintainer
#232 lore raw

PATCH 232/405 habanalabs: prevent device PTE read/write during hard-reset

In reply to Greg Kroah-Hartman

[ Upstream commit 9f201aba56b92c3daa4b76efae056ddbb80d91e6 ]

During hard-reset, contexts are closed as part of the tear-down process. After a context is closed, the driver cleans up the page tables of that context in the device's DRAM. This action is both dangerous and unnecessary.

It is unnecessary, because the device is going through a hard-reset, which means the device's DRAM contents are no longer valid and the device's MMU is being reset.

It is dangerous, because if the hard-reset came as a result of a PCI freeze, this action may cause the entire host machine to hang.

Therefore, prevent all device PTE updates when a hard-reset operation is pending.

---
 drivers/misc/habanalabs/goya/goya.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/misc/habanalabs/goya/goya.c +6 -0
index 3c509e19d69dc..1533cb3205400 100644
@@ -4407,6 +4407,9 @@ static u64 goya_read_pte(struct hl_device *hdev, u64 addr)
 {
 	struct goya_device *goya = hdev->asic_specific;
 
+	if (hdev->hard_reset_pending)
+		return U64_MAX;
+
 	return readq(hdev->pcie_bar[DDR_BAR_ID] +
 			(addr - goya->ddr_bar_cur_addr));
 }
@@ -4415,6 +4418,9 @@ static void goya_write_pte(struct hl_device *hdev, u64 addr, u64 val)
 {
 	struct goya_device *goya = hdev->asic_specific;
 
+	if (hdev->hard_reset_pending)
+		return;
+
 	writeq(val, hdev->pcie_bar[DDR_BAR_ID] +
 			(addr - goya->ddr_bar_cur_addr));
 }
Signature · 1 line
2.20.1
author maintainer
#233 lore raw

PATCH 233/405 habanalabs: all FD must be closed before removing device

In reply to Greg Kroah-Hartman

[ Upstream commit caa3c8e52582fc4d2ed82afd5e7ea164c18ef4fe ]

This patch fixes a bug in the implementation of the function that removes the device.

The bug can happen when the device is removed but not the driver itself (e.g. remove by the OS due to PCI freeze in Power architecture).

In that case, there maybe open users that are calling IOCTLs while the device is removed. This is a possible race condition that the driver must handle. Otherwise, a kernel panic may occur.

This race is prevented in the hard-reset flow, because the driver makes sure the users are closed before continuing with the hard-reset. This race can not occur when the driver itself is removed because the OS makes sure all the file descriptors are closed.

The fix is to make sure the open users close their file descriptors and if they don't (after a certain amount of time), the driver sends them a SIGKILL, because the remove of the device can't be stopped.

The patch re-uses the same code that is called from the hard-reset flow.

---
 drivers/misc/habanalabs/device.c | 32 +++++++++++++++++++++++++++-----
 1 file changed, 27 insertions(+), 5 deletions(-)
drivers/misc/habanalabs/device.c +27 -5
index 77d51be66c7e8..652c8edb2164c 100644
@@ -498,11 +498,8 @@ int hl_device_resume(struct hl_device *hdev)
 	return rc;
 }
 
-static void hl_device_hard_reset_pending(struct work_struct *work)
+static void device_kill_open_processes(struct hl_device *hdev)
 {
-	struct hl_device_reset_work *device_reset_work =
-		container_of(work, struct hl_device_reset_work, reset_work);
-	struct hl_device *hdev = device_reset_work->hdev;
 	u16 pending_total, pending_cnt;
 	struct task_struct *task = NULL;
 
@@ -537,6 +534,12 @@ static void hl_device_hard_reset_pending(struct work_struct *work)
 		}
 	}
 
+	/* We killed the open users, but because the driver cleans up after the
+	 * user contexts are closed (e.g. mmu mappings), we need to wait again
+	 * to make sure the cleaning phase is finished before continuing with
+	 * the reset
+	 */
+
 	pending_cnt = pending_total;
 
 	while ((atomic_read(&hdev->fd_open_cnt)) && (pending_cnt)) {
@@ -552,6 +555,16 @@ static void hl_device_hard_reset_pending(struct work_struct *work)
 
 	mutex_unlock(&hdev->fd_open_cnt_lock);
 
+}
+
+static void device_hard_reset_pending(struct work_struct *work)
+{
+	struct hl_device_reset_work *device_reset_work =
+		container_of(work, struct hl_device_reset_work, reset_work);
+	struct hl_device *hdev = device_reset_work->hdev;
+
+	device_kill_open_processes(hdev);
+
 	hl_device_reset(hdev, true, true);
 
 	kfree(device_reset_work);
@@ -635,7 +648,7 @@ int hl_device_reset(struct hl_device *hdev, bool hard_reset,
 		 * from a dedicated work
 		 */
 		INIT_WORK(&device_reset_work->reset_work,
-				hl_device_hard_reset_pending);
+				device_hard_reset_pending);
 		device_reset_work->hdev = hdev;
 		schedule_work(&device_reset_work->reset_work);
 
@@ -1035,6 +1048,15 @@ void hl_device_fini(struct hl_device *hdev)
 	/* Mark device as disabled */
 	hdev->disabled = true;
 
+	/*
+	 * Flush anyone that is inside the critical section of enqueue
+	 * jobs to the H/W
+	 */
+	hdev->asic_funcs->hw_queues_lock(hdev);
+	hdev->asic_funcs->hw_queues_unlock(hdev);
+
+	device_kill_open_processes(hdev);
+
 	hl_hwmon_fini(hdev);
 
 	device_late_fini(hdev);
Signature · 1 line
2.20.1
author maintainer
#234 lore raw

PATCH 234/405 samples/bpf: fix build with new clang

In reply to Greg Kroah-Hartman

[ Upstream commit 636e78b1cdb40b77a79b143dbd9d94847b360efa ]

clang started to error on invalid asm clobber usage in x86 headers and many bpf program samples failed to build with the message:

CLANG-bpf /data/users/ast/bpf-next/samples/bpf/xdp_redirect_kern.o
In file included from /data/users/ast/bpf-next/samples/bpf/xdp_redirect_kern.c:14:
In file included from ../include/linux/in.h:23:
In file included from ../include/uapi/linux/in.h:24:
In file included from ../include/linux/socket.h:8:
In file included from ../include/linux/uio.h:14:
In file included from ../include/crypto/hash.h:16:
In file included from ../include/linux/crypto.h:26:
In file included from ../include/linux/uaccess.h:5:
In file included from ../include/linux/sched.h:15:
In file included from ../include/linux/sem.h:5:
In file included from ../include/uapi/linux/sem.h:5:
In file included from ../include/linux/ipc.h:9:
In file included from ../include/linux/refcount.h:72:
../arch/x86/include/asm/refcount.h:72:36: error: asm-specifier for input or output variable conflicts with asm clobber list

                                         r->refs.counter, e, "er", i, "cx");
                                                                      ^
../arch/x86/include/asm/refcount.h:86:27: error: asm-specifier for input or output variable conflicts with asm clobber list
                                         r->refs.counter, e, "cx");
                                                             ^

2 errors generated.

Override volatile() to workaround the problem.

---
 samples/bpf/asm_goto_workaround.h | 1 +
 1 file changed, 1 insertion(+)
samples/bpf/asm_goto_workaround.h +1 -0
index 5cd7c1d1a5d56..7409722727ca1 100644
@@ -13,4 +13,5 @@
 #define asm_volatile_goto(x...) asm volatile("invalid use of asm_volatile_goto")
 #endif
 
+#define volatile(x...) volatile("")
 #endif
Signature · 1 line
2.20.1
author maintainer
#235 lore raw

PATCH 235/405 x86/build: Keep local relocations with ld.lld

In reply to Greg Kroah-Hartman

[ Upstream commit 7c21383f3429dd70da39c0c7f1efa12377a47ab6 ]

The LLVM linker (ld.lld) defaults to removing local relocations, which causes KASLR boot failures. ld.bfd and ld.gold already handle this correctly. This adds the explicit instruction "--discard-none" during the link phase. There is no change in output for ld.bfd and ld.gold, but ld.lld now produces an image with all the needed relocations.

---
 arch/x86/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/Makefile +1 -1
index a587805c6687f..56e748a7679f4 100644
@@ -47,7 +47,7 @@ export REALMODE_CFLAGS
 export BITS
 
 ifdef CONFIG_X86_NEED_RELOCS
-        LDFLAGS_vmlinux := --emit-relocs
+        LDFLAGS_vmlinux := --emit-relocs --discard-none
 endif
 
 #
Signature · 1 line
2.20.1
author maintainer
#236 lore raw

PATCH 236/405 spi: Dont call spi_get_gpio_descs() before device name is set

In reply to Greg Kroah-Hartman

[ Upstream commit 0a919ae49223d32ac0e8be3494547fcd1e4aa0aa ]

Move code calling spi_get_gpio_descs() to happen after ctlr->dev's name is set in order to have proper GPIO consumer names.

Before:

cat /sys/kernel/debug/gpio
gpiochip0: GPIOs 0-31, parent: platform/40049000.gpio, vf610-gpio:
gpio-6 ( |regulator-usb0-vbus ) out lo

gpiochip1: GPIOs 32-63, parent: platform/4004a000.gpio, vf610-gpio:

 gpio-36  (                    |scl                 ) in  hi
 gpio-37  (                    |sda                 ) in  hi
 gpio-40  (                    |(null) CS1          ) out lo
 gpio-41  (                    |(null) CS0          ) out lo ACTIVE LOW
 gpio-42  (                    |miso                ) in  hi
 gpio-43  (                    |mosi                ) in  lo
 gpio-44  (                    |sck                 ) out lo

After:

cat /sys/kernel/debug/gpio
gpiochip0: GPIOs 0-31, parent: platform/40049000.gpio, vf610-gpio:
gpio-6 ( |regulator-usb0-vbus ) out lo

gpiochip1: GPIOs 32-63, parent: platform/4004a000.gpio, vf610-gpio:

 gpio-36  (                    |scl                 ) in  hi
 gpio-37  (                    |sda                 ) in  hi
 gpio-40  (                    |spi0 CS1            ) out lo
 gpio-41  (                    |spi0 CS0            ) out lo ACTIVE LOW
 gpio-42  (                    |miso                ) in  hi
 gpio-43  (                    |mosi                ) in  lo
 gpio-44  (                    |sck                 ) out lo
---
 drivers/spi/spi.c | 37 +++++++++++++++++++------------------
 1 file changed, 19 insertions(+), 18 deletions(-)
drivers/spi/spi.c +19 -18
index 93986f879b09e..d17f68775a4bb 100644
@@ -2275,24 +2275,6 @@ int spi_register_controller(struct spi_controller *ctlr)
 	if (status)
 		return status;
 
-	if (!spi_controller_is_slave(ctlr)) {
-		if (ctlr->use_gpio_descriptors) {
-			status = spi_get_gpio_descs(ctlr);
-			if (status)
-				return status;
-			/*
-			 * A controller using GPIO descriptors always
-			 * supports SPI_CS_HIGH if need be.
-			 */
-			ctlr->mode_bits |= SPI_CS_HIGH;
-		} else {
-			/* Legacy code path for GPIOs from DT */
-			status = of_spi_register_master(ctlr);
-			if (status)
-				return status;
-		}
-	}
-
 	/* even if it's just one always-selected device, there must
 	 * be at least one chipselect
 	 */
@@ -2349,6 +2331,25 @@ int spi_register_controller(struct spi_controller *ctlr)
 	 * registration fails if the bus ID is in use.
 	 */
 	dev_set_name(&ctlr->dev, "spi%u", ctlr->bus_num);
+
+	if (!spi_controller_is_slave(ctlr)) {
+		if (ctlr->use_gpio_descriptors) {
+			status = spi_get_gpio_descs(ctlr);
+			if (status)
+				return status;
+			/*
+			 * A controller using GPIO descriptors always
+			 * supports SPI_CS_HIGH if need be.
+			 */
+			ctlr->mode_bits |= SPI_CS_HIGH;
+		} else {
+			/* Legacy code path for GPIOs from DT */
+			status = of_spi_register_master(ctlr);
+			if (status)
+				return status;
+		}
+	}
+
 	status = device_add(&ctlr->dev);
 	if (status < 0) {
 		/* free bus id */
Signature · 1 line
2.20.1
author maintainer
#237 lore raw

PATCH 237/405 regulator: core: Avoid potential deadlock on regulator_unregister

In reply to Greg Kroah-Hartman

[ Upstream commit 063773011d33bb36588a90385aa9eb75d13c6d80 ]

Lockdep reports the following issue on my setup:

Possible unsafe locking scenario:

CPU0 CPU1
---- ----

lock((work_completion)(&(&rdev->disable_work)->work));
                        lock(regulator_list_mutex);
                        lock((work_completion)(&(&rdev->disable_work)->work));

lock(regulator_list_mutex);

The problem is that regulator_unregister takes the regulator_list_mutex and then calls flush_work on disable_work. But regulator_disable_work calls regulator_lock_dependent which will also take the regulator_list_mutex. Resulting in a deadlock if the flush_work call actually needs to flush the work.

Fix this issue by moving the flush_work outside of the regulator_list_mutex. The list mutex is not used to guard the point at which the delayed work is queued, so its use adds no additional safety.

Fixes f8702f9e4aa7 ("regulator: core: Use ww_mutex for regulators locking")
---
 drivers/regulator/core.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/regulator/core.c +2 -1
index 6da41207e479a..35a7d020afecd 100644
@@ -5062,10 +5062,11 @@ void regulator_unregister(struct regulator_dev *rdev)
 		regulator_put(rdev->supply);
 	}
 
+	flush_work(&rdev->disable_work.work);
+
 	mutex_lock(&regulator_list_mutex);
 
 	debugfs_remove_recursive(rdev->debugfs);
-	flush_work(&rdev->disable_work.work);
 	WARN_ON(rdev->open_count);
 	regulator_remove_coupling(rdev);
 	unset_regulator_supplies(rdev);
Signature · 1 line
2.20.1
author maintainer
#238 lore raw

PATCH 238/405 ASoC: core: remove link components before cleaning up card resources

In reply to Greg Kroah-Hartman

[ Upstream commit f96fb7d198ca624fe33c4145a004eb5a3d0eddec ]

When the card is registered by the machine driver, dai link components are probed after the snd_card is created. This is done in snd_soc_bind_card() which calls snd_soc_instantiate_card() to first create the snd_card and then probes the link components by calling soc_probe_link_components(). The snd_card is used by the component driver to add the kcontrols associated with dapm widgets to the card.

When the machine driver is unregistered, the snd_card is freed when the card resources are cleaned up. But the snd_card needs to be valid while unloading the topology dapm widgets in order to remove the kcontrols from the card.

Since, unloading topology is done when the component driver is removed, the link components should be removed in snd_soc_unbind_card(). This will ensure that the kcontrols are removed before the card resources are cleaned up and the snd_card itself is freed.

---
 sound/soc/soc-core.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
sound/soc/soc-core.c +11 -0
index 46e3ab0fced47..fe99b02bbf171 100644
@@ -2828,10 +2828,21 @@ EXPORT_SYMBOL_GPL(snd_soc_register_card);
 
 static void snd_soc_unbind_card(struct snd_soc_card *card, bool unregister)
 {
+	struct snd_soc_pcm_runtime *rtd;
+	int order;
+
 	if (card->instantiated) {
 		card->instantiated = false;
 		snd_soc_dapm_shutdown(card);
 		snd_soc_flush_all_delayed_work(card);
+
+		/* remove all components used by DAI links on this card */
+		for_each_comp_order(order) {
+			for_each_card_rtds(card, rtd) {
+				soc_remove_link_components(card, rtd, order);
+			}
+		}
+
 		soc_cleanup_card_resources(card);
 		if (!unregister)
 			list_add(&card->list, &unbind_card_list);
Signature · 1 line
2.20.1
author maintainer
#239 lore raw

PATCH 239/405 drm/pl111: fix possible object reference leak

In reply to Greg Kroah-Hartman

[ Upstream commit bc29d3a69d4c1bd1a103e8b3c1ed81b807c1870b ]

The call to of_find_matching_node_and_match returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
drivers/gpu/drm/pl111/pl111_versatile.c:333:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 317, but without a corresponding object release within this function. drivers/gpu/drm/pl111/pl111_versatile.c:340:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 317, but without a corresponding object release within this function. drivers/gpu/drm/pl111/pl111_versatile.c:346:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 317, but without a corresponding object release within this function. drivers/gpu/drm/pl111/pl111_versatile.c:354:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 317, but without a corresponding object release within this function. drivers/gpu/drm/pl111/pl111_versatile.c:395:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 317, but without a corresponding object release within this function. drivers/gpu/drm/pl111/pl111_versatile.c:402:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 317, but without a corresponding object release within this function.

Cc Eric Anholt <[email protected]> (supporter:DRM DRIVER FOR ARM PL111 CLCD)
Cc David Airlie <[email protected]> (maintainer:DRM DRIVERS)
Cc Daniel Vetter <[email protected]> (maintainer:DRM DRIVERS)
Cc [email protected] (open list:DRM DRIVERS)
Cc [email protected] (open list)
---
 drivers/gpu/drm/pl111/pl111_versatile.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/pl111/pl111_versatile.c +4 -0
index b9baefdba38a1..1c318ad32a8cd 100644
@@ -330,6 +330,7 @@ int pl111_versatile_init(struct device *dev, struct pl111_drm_dev_private *priv)
 		ret = vexpress_muxfpga_init();
 		if (ret) {
 			dev_err(dev, "unable to initialize muxfpga driver\n");
+			of_node_put(np);
 			return ret;
 		}
 
@@ -337,17 +338,20 @@ int pl111_versatile_init(struct device *dev, struct pl111_drm_dev_private *priv)
 		pdev = of_find_device_by_node(np);
 		if (!pdev) {
 			dev_err(dev, "can't find the sysreg device, deferring\n");
+			of_node_put(np);
 			return -EPROBE_DEFER;
 		}
 		map = dev_get_drvdata(&pdev->dev);
 		if (!map) {
 			dev_err(dev, "sysreg has not yet probed\n");
 			platform_device_put(pdev);
+			of_node_put(np);
 			return -EPROBE_DEFER;
 		}
 	} else {
 		map = syscon_node_to_regmap(np);
 	}
+	of_node_put(np);
 
 	if (IS_ERR(map)) {
 		dev_err(dev, "no Versatile syscon regmap\n");
Signature · 8 lines
2.20.1



_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
author maintainer
#240 lore raw

PATCH 240/405 iio: ad_sigma_delta: Properly handle SPI bus locking vs CS assertion

In reply to Greg Kroah-Hartman

[ Upstream commit df1d80aee963480c5c2938c64ec0ac3e4a0df2e0 ]

For devices from the SigmaDelta family we need to keep CS low when doing a conversion, since the device will use the MISO line as a interrupt to indicate that the conversion is complete.

This is why the driver locks the SPI bus and when the SPI bus is locked keeps as long as a conversion is going on. The current implementation gets one small detail wrong though. CS is only de-asserted after the SPI bus is unlocked. This means it is possible for a different SPI device on the same bus to send a message which would be wrongfully be addressed to the SigmaDelta device as well. Make sure that the last SPI transfer that is done while holding the SPI bus lock de-asserts the CS signal.

---
 drivers/iio/adc/ad_sigma_delta.c       | 16 +++++++++++-----
 include/linux/iio/adc/ad_sigma_delta.h |  1 +
 2 files changed, 12 insertions(+), 5 deletions(-)
drivers/iio/adc/ad_sigma_delta.c +11 -5
index 54d9978b27405..a4310600a8536 100644
@@ -62,7 +62,7 @@ int ad_sd_write_reg(struct ad_sigma_delta *sigma_delta, unsigned int reg,
 	struct spi_transfer t = {
 		.tx_buf		= data,
 		.len		= size + 1,
-		.cs_change	= sigma_delta->bus_locked,
+		.cs_change	= sigma_delta->keep_cs_asserted,
 	};
 	struct spi_message m;
 	int ret;
@@ -218,6 +218,7 @@ static int ad_sd_calibrate(struct ad_sigma_delta *sigma_delta,
 
 	spi_bus_lock(sigma_delta->spi->master);
 	sigma_delta->bus_locked = true;
+	sigma_delta->keep_cs_asserted = true;
 	reinit_completion(&sigma_delta->completion);
 
 	ret = ad_sigma_delta_set_mode(sigma_delta, mode);
@@ -235,9 +236,10 @@ static int ad_sd_calibrate(struct ad_sigma_delta *sigma_delta,
 		ret = 0;
 	}
 out:
+	sigma_delta->keep_cs_asserted = false;
+	ad_sigma_delta_set_mode(sigma_delta, AD_SD_MODE_IDLE);
 	sigma_delta->bus_locked = false;
 	spi_bus_unlock(sigma_delta->spi->master);
-	ad_sigma_delta_set_mode(sigma_delta, AD_SD_MODE_IDLE);
 
 	return ret;
 }
@@ -290,6 +292,7 @@ int ad_sigma_delta_single_conversion(struct iio_dev *indio_dev,
 
 	spi_bus_lock(sigma_delta->spi->master);
 	sigma_delta->bus_locked = true;
+	sigma_delta->keep_cs_asserted = true;
 	reinit_completion(&sigma_delta->completion);
 
 	ad_sigma_delta_set_mode(sigma_delta, AD_SD_MODE_SINGLE);
@@ -299,9 +302,6 @@ int ad_sigma_delta_single_conversion(struct iio_dev *indio_dev,
 	ret = wait_for_completion_interruptible_timeout(
 			&sigma_delta->completion, HZ);
 
-	sigma_delta->bus_locked = false;
-	spi_bus_unlock(sigma_delta->spi->master);
-
 	if (ret == 0)
 		ret = -EIO;
 	if (ret < 0)
@@ -322,7 +322,10 @@ int ad_sigma_delta_single_conversion(struct iio_dev *indio_dev,
 		sigma_delta->irq_dis = true;
 	}
 
+	sigma_delta->keep_cs_asserted = false;
 	ad_sigma_delta_set_mode(sigma_delta, AD_SD_MODE_IDLE);
+	sigma_delta->bus_locked = false;
+	spi_bus_unlock(sigma_delta->spi->master);
 	mutex_unlock(&indio_dev->mlock);
 
 	if (ret)
@@ -359,6 +362,8 @@ static int ad_sd_buffer_postenable(struct iio_dev *indio_dev)
 
 	spi_bus_lock(sigma_delta->spi->master);
 	sigma_delta->bus_locked = true;
+	sigma_delta->keep_cs_asserted = true;
+
 	ret = ad_sigma_delta_set_mode(sigma_delta, AD_SD_MODE_CONTINUOUS);
 	if (ret)
 		goto err_unlock;
@@ -387,6 +392,7 @@ static int ad_sd_buffer_postdisable(struct iio_dev *indio_dev)
 		sigma_delta->irq_dis = true;
 	}
 
+	sigma_delta->keep_cs_asserted = false;
 	ad_sigma_delta_set_mode(sigma_delta, AD_SD_MODE_IDLE);
 
 	sigma_delta->bus_locked = false;
include/linux/iio/adc/ad_sigma_delta.h +1 -0
index 7e84351fa2c05..6e9fb1932dde9 100644
@@ -69,6 +69,7 @@ struct ad_sigma_delta {
 	bool			irq_dis;
 
 	bool			bus_locked;
+	bool			keep_cs_asserted;
 
 	uint8_t			comm;
 
Signature · 1 line
2.20.1
author maintainer
#241 lore raw

PATCH 241/405 iio: hmc5843: fix potential NULL pointer dereferences

In reply to Greg Kroah-Hartman

[ Upstream commit 536cc27deade8f1ec3c1beefa60d5fbe0f6fcb28 ]

devm_regmap_init_i2c may fail and return NULL. The fix returns the error when it fails.

---
 drivers/iio/magnetometer/hmc5843_i2c.c | 7 ++++++-
 drivers/iio/magnetometer/hmc5843_spi.c | 7 ++++++-
 2 files changed, 12 insertions(+), 2 deletions(-)
drivers/iio/magnetometer/hmc5843_i2c.c +6 -1
index 3de7f4426ac40..86abba5827a25 100644
@@ -58,8 +58,13 @@ static const struct regmap_config hmc5843_i2c_regmap_config = {
 static int hmc5843_i2c_probe(struct i2c_client *cli,
 			     const struct i2c_device_id *id)
 {
+	struct regmap *regmap = devm_regmap_init_i2c(cli,
+			&hmc5843_i2c_regmap_config);
+	if (IS_ERR(regmap))
+		return PTR_ERR(regmap);
+
 	return hmc5843_common_probe(&cli->dev,
-			devm_regmap_init_i2c(cli, &hmc5843_i2c_regmap_config),
+			regmap,
 			id->driver_data, id->name);
 }
 
drivers/iio/magnetometer/hmc5843_spi.c +6 -1
index 535f03a70d630..79b2b707f90e7 100644
@@ -58,6 +58,7 @@ static const struct regmap_config hmc5843_spi_regmap_config = {
 static int hmc5843_spi_probe(struct spi_device *spi)
 {
 	int ret;
+	struct regmap *regmap;
 	const struct spi_device_id *id = spi_get_device_id(spi);
 
 	spi->mode = SPI_MODE_3;
@@ -67,8 +68,12 @@ static int hmc5843_spi_probe(struct spi_device *spi)
 	if (ret)
 		return ret;
 
+	regmap = devm_regmap_init_spi(spi, &hmc5843_spi_regmap_config);
+	if (IS_ERR(regmap))
+		return PTR_ERR(regmap);
+
 	return hmc5843_common_probe(&spi->dev,
-			devm_regmap_init_spi(spi, &hmc5843_spi_regmap_config),
+			regmap,
 			id->driver_data, id->name);
 }
 
Signature · 1 line
2.20.1
author maintainer
#242 lore raw

PATCH 242/405 iio: common: ssp_sensors: Initialize calculated_time in ssp_common_process_data

In reply to Greg Kroah-Hartman

[ Upstream commit 6f9ca1d3eb74b81f811a87002de2d51640d135b1 ]

When building with -Wsometimes-uninitialized, Clang warns:

drivers/iio/common/ssp_sensors/ssp_iio.c:95:6: warning: variable 'calculated_time' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized]

While it isn't wrong, this will never be a problem because iio_push_to_buffers_with_timestamp only uses calculated_time on the same condition that it is assigned (when scan_timestamp is not zero). While iio_push_to_buffers_with_timestamp is marked as inline, Clang does inlining in the optimization stage, which happens after the semantic analysis phase (plus inline is merely a hint to the compiler).

Fix this by just zero initializing calculated_time.

---
 drivers/iio/common/ssp_sensors/ssp_iio.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/iio/common/ssp_sensors/ssp_iio.c +1 -1
index 645f2e3975db4..e38f704d88b7e 100644
@@ -81,7 +81,7 @@ int ssp_common_process_data(struct iio_dev *indio_dev, void *buf,
 			    unsigned int len, int64_t timestamp)
 {
 	__le32 time;
-	int64_t calculated_time;
+	int64_t calculated_time = 0;
 	struct ssp_sensor_data *spd = iio_priv(indio_dev);
 
 	if (indio_dev->scan_bytes == 0)
Signature · 1 line
2.20.1
author maintainer
#243 lore raw

PATCH 243/405 iio: adc: ti-ads7950: Fix improper use of mlock

In reply to Greg Kroah-Hartman

[ Upstream commit abbde2792999c9ad3514dd25d7f8d9a96034fe16 ]

Indio->mlock is used for protecting the different iio device modes. It is currently not being used in this way. Replace the lock with an internal lock specifically used for protecting the SPI transfer buffer.

---
 drivers/iio/adc/ti-ads7950.c | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)
drivers/iio/adc/ti-ads7950.c +15 -4
index 0ad63592cc3c9..1e47bef72bb79 100644
@@ -56,6 +56,9 @@ struct ti_ads7950_state {
 	struct spi_message	ring_msg;
 	struct spi_message	scan_single_msg;
 
+	/* Lock to protect the spi xfer buffers */
+	struct mutex		slock;
+
 	struct regulator	*reg;
 	unsigned int		vref_mv;
 
@@ -268,6 +271,7 @@ static irqreturn_t ti_ads7950_trigger_handler(int irq, void *p)
 	struct ti_ads7950_state *st = iio_priv(indio_dev);
 	int ret;
 
+	mutex_lock(&st->slock);
 	ret = spi_sync(st->spi, &st->ring_msg);
 	if (ret < 0)
 		goto out;
@@ -276,6 +280,7 @@ static irqreturn_t ti_ads7950_trigger_handler(int irq, void *p)
 					   iio_get_time_ns(indio_dev));
 
 out:
+	mutex_unlock(&st->slock);
 	iio_trigger_notify_done(indio_dev->trig);
 
 	return IRQ_HANDLED;
@@ -286,7 +291,7 @@ static int ti_ads7950_scan_direct(struct iio_dev *indio_dev, unsigned int ch)
 	struct ti_ads7950_state *st = iio_priv(indio_dev);
 	int ret, cmd;
 
-	mutex_lock(&indio_dev->mlock);
+	mutex_lock(&st->slock);
 
 	cmd = TI_ADS7950_CR_WRITE | TI_ADS7950_CR_CHAN(ch) | st->settings;
 	st->single_tx = cmd;
@@ -298,7 +303,7 @@ static int ti_ads7950_scan_direct(struct iio_dev *indio_dev, unsigned int ch)
 	ret = st->single_rx;
 
 out:
-	mutex_unlock(&indio_dev->mlock);
+	mutex_unlock(&st->slock);
 
 	return ret;
 }
@@ -432,16 +437,19 @@ static int ti_ads7950_probe(struct spi_device *spi)
 	if (ACPI_COMPANION(&spi->dev))
 		st->vref_mv = TI_ADS7950_VA_MV_ACPI_DEFAULT;
 
+	mutex_init(&st->slock);
+
 	st->reg = devm_regulator_get(&spi->dev, "vref");
 	if (IS_ERR(st->reg)) {
 		dev_err(&spi->dev, "Failed get get regulator \"vref\"\n");
-		return PTR_ERR(st->reg);
+		ret = PTR_ERR(st->reg);
+		goto error_destroy_mutex;
 	}
 
 	ret = regulator_enable(st->reg);
 	if (ret) {
 		dev_err(&spi->dev, "Failed to enable regulator \"vref\"\n");
-		return ret;
+		goto error_destroy_mutex;
 	}
 
 	ret = iio_triggered_buffer_setup(indio_dev, NULL,
@@ -463,6 +471,8 @@ static int ti_ads7950_probe(struct spi_device *spi)
 	iio_triggered_buffer_cleanup(indio_dev);
 error_disable_reg:
 	regulator_disable(st->reg);
+error_destroy_mutex:
+	mutex_destroy(&st->slock);
 
 	return ret;
 }
@@ -475,6 +485,7 @@ static int ti_ads7950_remove(struct spi_device *spi)
 	iio_device_unregister(indio_dev);
 	iio_triggered_buffer_cleanup(indio_dev);
 	regulator_disable(st->reg);
+	mutex_destroy(&st->slock);
 
 	return 0;
 }
Signature · 1 line
2.20.1
author maintainer
#244 lore raw

PATCH 244/405 net: hns3: check resetting status in hns3_get_stats()

In reply to Greg Kroah-Hartman

[ Upstream commit c4e401e5a934bb0798ebbba98e08dab129695eff ]

hns3_get_stats() should check the resetting status firstly, since the device will be reinitialized when resetting. If the reset has not completed, the hns3_get_stats() may access invalid memory.

---
 drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c +5 -0
index 359d4731fb2db..ea94b5152963f 100644
@@ -483,6 +483,11 @@ static void hns3_get_stats(struct net_device *netdev,
 	struct hnae3_handle *h = hns3_get_handle(netdev);
 	u64 *p = data;
 
+	if (hns3_nic_resetting(netdev)) {
+		netdev_err(netdev, "dev resetting, could not get stats\n");
+		return;
+	}
+
 	if (!h->ae_algo->ops->get_stats || !h->ae_algo->ops->update_stats) {
 		netdev_err(netdev, "could not get any statistics\n");
 		return;
Signature · 1 line
2.20.1
author maintainer
#245 lore raw

PATCH 245/405 net: hns3: add protect when handling mac addr list

In reply to Greg Kroah-Hartman

[ Upstream commit 389775a6605e040dddea21a778a88eaaa57c068d ]

It used netdev->uc and netdev->mc list in function hns3_recover_hw_addr() and hns3_remove_hw_addr(). We should add protect for them.

Fixes f05e21097121 ("net: hns3: Clear mac vlan table entries when unload driver or function reset")
---
 drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c +8 -2
index d6b488c2de332..c7d310903319f 100644
@@ -3774,12 +3774,13 @@ static int hns3_recover_hw_addr(struct net_device *ndev)
 	struct netdev_hw_addr *ha, *tmp;
 	int ret = 0;
 
+	netif_addr_lock_bh(ndev);
 	/* go through and sync uc_addr entries to the device */
 	list = &ndev->uc;
 	list_for_each_entry_safe(ha, tmp, &list->list, list) {
 		ret = hns3_nic_uc_sync(ndev, ha->addr);
 		if (ret)
-			return ret;
+			goto out;
 	}
 
 	/* go through and sync mc_addr entries to the device */
@@ -3787,9 +3788,11 @@ static int hns3_recover_hw_addr(struct net_device *ndev)
 	list_for_each_entry_safe(ha, tmp, &list->list, list) {
 		ret = hns3_nic_mc_sync(ndev, ha->addr);
 		if (ret)
-			return ret;
+			goto out;
 	}
 
+out:
+	netif_addr_unlock_bh(ndev);
 	return ret;
 }
 
@@ -3800,6 +3803,7 @@ static void hns3_remove_hw_addr(struct net_device *netdev)
 
 	hns3_nic_uc_unsync(netdev, netdev->dev_addr);
 
+	netif_addr_lock_bh(netdev);
 	/* go through and unsync uc_addr entries to the device */
 	list = &netdev->uc;
 	list_for_each_entry_safe(ha, tmp, &list->list, list)
@@ -3810,6 +3814,8 @@ static void hns3_remove_hw_addr(struct net_device *netdev)
 	list_for_each_entry_safe(ha, tmp, &list->list, list)
 		if (ha->refcount > 1)
 			hns3_nic_mc_unsync(netdev, ha->addr);
+
+	netif_addr_unlock_bh(netdev);
 }
 
 static void hns3_clear_tx_ring(struct hns3_enet_ring *ring)
Signature · 1 line
2.20.1
author maintainer
#246 lore raw

PATCH 246/405 selftests/bpf: ksym_search wont check symbols exists

In reply to Greg Kroah-Hartman

[ Upstream commit 0979ff7992fb6f4eb837995b12f4071dcafebd2d ]

Currently, ksym_search located at trace_helpers won't check symbols are existing or not.

In ksym_search, when symbol is not found, it will return &syms[0](_stext). But when the kernel symbols are not loaded, it will return NULL, which is not a desired action.

This commit will add verification logic whether symbols are loaded prior to the symbol search.

---
 tools/testing/selftests/bpf/trace_helpers.c | 4 ++++
 1 file changed, 4 insertions(+)
tools/testing/selftests/bpf/trace_helpers.c +4 -0
index 4cdb63bf0521d..9a9fc6c9b70b5 100644
@@ -52,6 +52,10 @@ struct ksym *ksym_search(long key)
 	int start = 0, end = sym_cnt;
 	int result;
 
+	/* kallsyms not loaded. return NULL */
+	if (sym_cnt <= 0)
+		return NULL;
+
 	while (start < end) {
 		size_t mid = start + (end - start) / 2;
 
Signature · 1 line
2.20.1
author maintainer
#247 lore raw

PATCH 247/405 rtlwifi: fix a potential NULL pointer dereference

In reply to Greg Kroah-Hartman

[ Upstream commit 765976285a8c8db3f0eb7f033829a899d0c2786e ]

In case alloc_workqueue fails, the fix reports the error and returns to avoid NULL pointer dereference.

---
 drivers/net/wireless/realtek/rtlwifi/base.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/net/wireless/realtek/rtlwifi/base.c +5 -0
index 217d2a7a43c74..ac746c322554b 100644
@@ -448,6 +448,11 @@ static void _rtl_init_deferred_work(struct ieee80211_hw *hw)
 	/* <2> work queue */
 	rtlpriv->works.hw = hw;
 	rtlpriv->works.rtl_wq = alloc_workqueue("%s", 0, 0, rtlpriv->cfg->name);
+	if (unlikely(!rtlpriv->works.rtl_wq)) {
+		pr_err("Failed to allocate work queue\n");
+		return;
+	}
+
 	INIT_DELAYED_WORK(&rtlpriv->works.watchdog_wq,
 			  (void *)rtl_watchdog_wq_callback);
 	INIT_DELAYED_WORK(&rtlpriv->works.ips_nic_off_wq,
Signature · 1 line
2.20.1
author maintainer
#248 lore raw

PATCH 248/405 mwifiex: Fix mem leak in mwifiex_tm_cmd

In reply to Greg Kroah-Hartman

[ Upstream commit 003b686ace820ce2d635a83f10f2d7f9c147dabc ]

'hostcmd' is alloced by kzalloc, should be freed before leaving from the error handling cases, otherwise it will cause mem leak.

Fixes 3935ccc14d2c ("mwifiex: add cfg80211 testmode support")
---
 drivers/net/wireless/marvell/mwifiex/cfg80211.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/net/wireless/marvell/mwifiex/cfg80211.c +5 -1
index c46f0a54a0c76..e582d9b3e50c2 100644
@@ -4082,16 +4082,20 @@ static int mwifiex_tm_cmd(struct wiphy *wiphy, struct wireless_dev *wdev,
 
 		if (mwifiex_send_cmd(priv, 0, 0, 0, hostcmd, true)) {
 			dev_err(priv->adapter->dev, "Failed to process hostcmd\n");
+			kfree(hostcmd);
 			return -EFAULT;
 		}
 
 		/* process hostcmd response*/
 		skb = cfg80211_testmode_alloc_reply_skb(wiphy, hostcmd->len);
-		if (!skb)
+		if (!skb) {
+			kfree(hostcmd);
 			return -ENOMEM;
+		}
 		err = nla_put(skb, MWIFIEX_TM_ATTR_DATA,
 			      hostcmd->len, hostcmd->cmd);
 		if (err) {
+			kfree(hostcmd);
 			kfree_skb(skb);
 			return -EMSGSIZE;
 		}
Signature · 1 line
2.20.1
author maintainer
#249 lore raw

PATCH 249/405 brcmfmac: fix missing checks for kmemdup

In reply to Greg Kroah-Hartman

[ Upstream commit 46953f97224d56a12ccbe9c6acaa84ca0dab2780 ]

In case kmemdup fails, the fix sets conn_info->req_ie_len and conn_info->resp_ie_len to zero to avoid buffer overflows.

---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c +4 -0
index e92f6351bd224..8ee8af4e7ec4f 100644
@@ -5464,6 +5464,8 @@ static s32 brcmf_get_assoc_ies(struct brcmf_cfg80211_info *cfg,
 		conn_info->req_ie =
 		    kmemdup(cfg->extra_buf, conn_info->req_ie_len,
 			    GFP_KERNEL);
+		if (!conn_info->req_ie)
+			conn_info->req_ie_len = 0;
 	} else {
 		conn_info->req_ie_len = 0;
 		conn_info->req_ie = NULL;
@@ -5480,6 +5482,8 @@ static s32 brcmf_get_assoc_ies(struct brcmf_cfg80211_info *cfg,
 		conn_info->resp_ie =
 		    kmemdup(cfg->extra_buf, conn_info->resp_ie_len,
 			    GFP_KERNEL);
+		if (!conn_info->resp_ie)
+			conn_info->resp_ie_len = 0;
 	} else {
 		conn_info->resp_ie_len = 0;
 		conn_info->resp_ie = NULL;
Signature · 1 line
2.20.1
author maintainer
#250 lore raw

PATCH 250/405 b43: shut up clang -Wuninitialized variable warning

In reply to Greg Kroah-Hartman

[ Upstream commit d825db346270dbceef83b7b750dbc29f1d7dcc0e ]

Clang warns about what is clearly a case of passing an uninitalized variable into a static function:

drivers/net/wireless/broadcom/b43/phy_lp.c:1852:23: error: variable 'gains' is uninitialized when used here

      [-Werror,-Wuninitialized]
                lpphy_papd_cal(dev, gains, 0, 1, 30);
                                    ^~~~~
drivers/net/wireless/broadcom/b43/phy_lp.c:1838:2: note: variable 'gains' is declared here
        struct lpphy_tx_gains gains, oldgains;
        ^

1 error generated.

However, this function is empty, and its arguments are never evaluated, so gcc in contrast does not warn here. Both compilers behave in a reasonable way as far as I can tell, so we should change the code to avoid the warning everywhere.

We could just eliminate the lpphy_papd_cal() function entirely, given that it has had the TODO comment in it for 10 years now and is rather unlikely to ever get done. I'm doing a simpler change here, and just pass the 'oldgains' variable in that has been initialized, based on the guess that this is what was originally meant.

Fixes 2c0d6100da3e ("b43: LP-PHY: Begin implementing calibration & software RFKILL support")
---
 drivers/net/wireless/broadcom/b43/phy_lp.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/net/wireless/broadcom/b43/phy_lp.c +3 -3
index 46408a560814c..aedee026c5e24 100644
@@ -1835,7 +1835,7 @@ static void lpphy_papd_cal(struct b43_wldev *dev, struct lpphy_tx_gains gains,
 static void lpphy_papd_cal_txpwr(struct b43_wldev *dev)
 {
 	struct b43_phy_lp *lpphy = dev->phy.lp;
-	struct lpphy_tx_gains gains, oldgains;
+	struct lpphy_tx_gains oldgains;
 	int old_txpctl, old_afe_ovr, old_rf, old_bbmult;
 
 	lpphy_read_tx_pctl_mode_from_hardware(dev);
@@ -1849,9 +1849,9 @@ static void lpphy_papd_cal_txpwr(struct b43_wldev *dev)
 	lpphy_set_tx_power_control(dev, B43_LPPHY_TXPCTL_OFF);
 
 	if (dev->dev->chip_id == 0x4325 && dev->dev->chip_rev == 0)
-		lpphy_papd_cal(dev, gains, 0, 1, 30);
+		lpphy_papd_cal(dev, oldgains, 0, 1, 30);
 	else
-		lpphy_papd_cal(dev, gains, 0, 1, 65);
+		lpphy_papd_cal(dev, oldgains, 0, 1, 65);
 
 	if (old_afe_ovr)
 		lpphy_set_tx_gains(dev, oldgains);
Signature · 1 line
2.20.1
author maintainer
#251 lore raw

PATCH 251/405 brcmfmac: convert dev_init_lock mutex to completion

In reply to Greg Kroah-Hartman

[ Upstream commit a9fd0953fa4a62887306be28641b4b0809f3b2fd ]

Leaving dev_init_lock mutex locked in probe causes BUG and a WARNING when kernel is compiled with CONFIG_PROVE_LOCKING. Convert mutex to completion which silences those warnings and improves code readability.

Fix below errors when connecting the USB WiFi dongle:

brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43143 for chip BCM43143/2

BUG: workqueue leaked lock or atomic: kworker/0:2/0x00000000/434
     last function: hub_event
1 lock held by kworker/0:2/434:
 #0: 18d5dcdf (&devinfo->dev_init_lock){+.+.}, at: brcmf_usb_probe+0x78/0x550 [brcmfmac]
CPU: 0 PID: 434 Comm: kworker/0:2 Not tainted 4.19.23-00084-g454a789-dirty #123
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
Workqueue: usb_hub_wq hub_event
[<8011237c>] (unwind_backtrace) from [<8010d74c>] (show_stack+0x10/0x14)
[<8010d74c>] (show_stack) from [<809c4324>] (dump_stack+0xa8/0xd4)
[<809c4324>] (dump_stack) from [<8014195c>] (process_one_work+0x710/0x808)
[<8014195c>] (process_one_work) from [<80141a80>] (worker_thread+0x2c/0x564)
[<80141a80>] (worker_thread) from [<80147bcc>] (kthread+0x13c/0x16c)
[<80147bcc>] (kthread) from [<801010b4>] (ret_from_fork+0x14/0x20)
Exception stack(0xed1d9fb0 to 0xed1d9ff8)

9fa0: 00000000 00000000 00000000 00000000
9fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 9fe0: 00000000 00000000 00000000 00000000 00000013 00000000

======================================================
WARNING: possible circular locking dependency detected 4.19.23-00084-g454a789-dirty #123 Not tainted
------------------------------------------------------
kworker/0:2/434 is trying to acquire lock:
e29cf799 ((wq_completion)"events"){+.+.}, at: process_one_work+0x174/0x808

but task is already holding lock:
18d5dcdf (&devinfo->dev_init_lock){+.+.}, at: brcmf_usb_probe+0x78/0x550 [brcmfmac]

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #2 (&devinfo->dev_init_lock){+.+.}:

       mutex_lock_nested+0x1c/0x24
       brcmf_usb_probe+0x78/0x550 [brcmfmac]
       usb_probe_interface+0xc0/0x1bc
       really_probe+0x228/0x2c0
       __driver_attach+0xe4/0xe8
       bus_for_each_dev+0x68/0xb4
       bus_add_driver+0x19c/0x214
       driver_register+0x78/0x110
       usb_register_driver+0x84/0x148
       process_one_work+0x228/0x808
       worker_thread+0x2c/0x564
       kthread+0x13c/0x16c
       ret_from_fork+0x14/0x20
         (null)

-> #1 (brcmf_driver_work){+.+.}:

       worker_thread+0x2c/0x564
       kthread+0x13c/0x16c
       ret_from_fork+0x14/0x20
         (null)

-> #0 ((wq_completion)"events"){+.+.}:

       process_one_work+0x1b8/0x808
       worker_thread+0x2c/0x564
       kthread+0x13c/0x16c
       ret_from_fork+0x14/0x20
         (null)

other info that might help us debug this:

Chain exists of:
  (wq_completion)"events" --> brcmf_driver_work --> &devinfo->dev_init_lock

 Possible unsafe locking scenario:
       CPU0                    CPU1
       ----                    ----
  lock(&devinfo->dev_init_lock);
                               lock(brcmf_driver_work);
                               lock(&devinfo->dev_init_lock);

lock((wq_completion)"events");

*** DEADLOCK ***

1 lock held by kworker/0:2/434:
#0: 18d5dcdf (&devinfo->dev_init_lock){+.+.}, at: brcmf_usb_probe+0x78/0x550 [brcmfmac]

stack backtrace:
CPU: 0 PID: 434 Comm: kworker/0:2 Not tainted 4.19.23-00084-g454a789-dirty #123
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
Workqueue: events request_firmware_work_func
[<8011237c>] (unwind_backtrace) from [<8010d74c>] (show_stack+0x10/0x14)
[<8010d74c>] (show_stack) from [<809c4324>] (dump_stack+0xa8/0xd4)
[<809c4324>] (dump_stack) from [<80172838>] (print_circular_bug+0x210/0x330)
[<80172838>] (print_circular_bug) from [<80175940>] (__lock_acquire+0x160c/0x1a30)
[<80175940>] (__lock_acquire) from [<8017671c>] (lock_acquire+0xe0/0x268)
[<8017671c>] (lock_acquire) from [<80141404>] (process_one_work+0x1b8/0x808)
[<80141404>] (process_one_work) from [<80141a80>] (worker_thread+0x2c/0x564)
[<80141a80>] (worker_thread) from [<80147bcc>] (kthread+0x13c/0x16c)
[<80147bcc>] (kthread) from [<801010b4>] (ret_from_fork+0x14/0x20)
Exception stack(0xed1d9fb0 to 0xed1d9ff8)

9fa0: 00000000 00000000 00000000 00000000
9fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 9fe0: 00000000 00000000 00000000 00000000 00000013 00000000

---
 .../wireless/broadcom/brcm80211/brcmfmac/usb.c  | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/usb.c +8 -9
index e9cbfd077710a..a513990cd1d6a 100644
@@ -160,7 +160,7 @@ struct brcmf_usbdev_info {
 
 	struct usb_device *usbdev;
 	struct device *dev;
-	struct mutex dev_init_lock;
+	struct completion dev_init_done;
 
 	int ctl_in_pipe, ctl_out_pipe;
 	struct urb *ctl_urb; /* URB for control endpoint */
@@ -1193,11 +1193,11 @@ static void brcmf_usb_probe_phase2(struct device *dev, int ret,
 	if (ret)
 		goto error;
 
-	mutex_unlock(&devinfo->dev_init_lock);
+	complete(&devinfo->dev_init_done);
 	return;
 error:
 	brcmf_dbg(TRACE, "failed: dev=%s, err=%d\n", dev_name(dev), ret);
-	mutex_unlock(&devinfo->dev_init_lock);
+	complete(&devinfo->dev_init_done);
 	device_release_driver(dev);
 }
 
@@ -1265,7 +1265,7 @@ static int brcmf_usb_probe_cb(struct brcmf_usbdev_info *devinfo)
 		if (ret)
 			goto fail;
 		/* we are done */
-		mutex_unlock(&devinfo->dev_init_lock);
+		complete(&devinfo->dev_init_done);
 		return 0;
 	}
 	bus->chip = bus_pub->devid;
@@ -1325,11 +1325,10 @@ brcmf_usb_probe(struct usb_interface *intf, const struct usb_device_id *id)
 
 	devinfo->usbdev = usb;
 	devinfo->dev = &usb->dev;
-	/* Take an init lock, to protect for disconnect while still loading.
+	/* Init completion, to protect for disconnect while still loading.
 	 * Necessary because of the asynchronous firmware load construction
 	 */
-	mutex_init(&devinfo->dev_init_lock);
-	mutex_lock(&devinfo->dev_init_lock);
+	init_completion(&devinfo->dev_init_done);
 
 	usb_set_intfdata(intf, devinfo);
 
@@ -1407,7 +1406,7 @@ brcmf_usb_probe(struct usb_interface *intf, const struct usb_device_id *id)
 	return 0;
 
 fail:
-	mutex_unlock(&devinfo->dev_init_lock);
+	complete(&devinfo->dev_init_done);
 	kfree(devinfo);
 	usb_set_intfdata(intf, NULL);
 	return ret;
@@ -1422,7 +1421,7 @@ brcmf_usb_disconnect(struct usb_interface *intf)
 	devinfo = (struct brcmf_usbdev_info *)usb_get_intfdata(intf);
 
 	if (devinfo) {
-		mutex_lock(&devinfo->dev_init_lock);
+		wait_for_completion(&devinfo->dev_init_done);
 		/* Make sure that devinfo still exists. Firmware probe routines
 		 * may have released the device and cleared the intfdata.
 		 */
Signature · 1 line
2.20.1
author maintainer
#252 lore raw

PATCH 252/405 brcmfmac: fix WARNING during USB disconnect in case of unempty psq

In reply to Greg Kroah-Hartman

[ Upstream commit c80d26e81ef1802f30364b4ad1955c1443a592b9 ]

brcmu_pkt_buf_free_skb emits WARNING when attempting to free a sk_buff which is part of any queue. After USB disconnect this may have happened when brcmf_fws_hanger_cleanup() is called as per-interface psq was never cleaned when removing the interface.
Change brcmf_fws_macdesc_cleanup() in a way that it removes the corresponding packets from hanger table (to avoid double-free when brcmf_fws_hanger_cleanup() is called) and add a call to clean-up the interface specific packet queue.

Below is a WARNING during USB disconnect with Raspberry Pi WiFi dongle running in AP mode. This was reproducible when the interface was transmitting during the disconnect and is fixed with this commit.

------------[ cut here ]------------
WARNING: CPU: 0 PID: 1171 at drivers/net/wireless/broadcom/brcm80211/brcmutil/utils.c:49 brcmu_pkt_buf_free_skb+0x3c/0x40
Modules linked in: nf_log_ipv4 nf_log_common xt_LOG xt_limit iptable_mangle xt_connmark xt_tcpudp xt_conntrack nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 iptable_filter ip_tables x_tables usb_f_mass_storage usb_f_rndis u_ether cdc_acm smsc95xx usbnet ci_hdrc_imx ci_hdrc ulpi usbmisc_imx 8250_exar 8250_pci 8250 8250_base libcomposite configfs udc_core
CPU: 0 PID: 1171 Comm: kworker/0:0 Not tainted 4.19.23-00075-gde33ed8 #99
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
Workqueue: usb_hub_wq hub_event
[<8010ff84>] (unwind_backtrace) from [<8010bb64>] (show_stack+0x10/0x14)
[<8010bb64>] (show_stack) from [<80840278>] (dump_stack+0x88/0x9c)
[<80840278>] (dump_stack) from [<8011f5ec>] (__warn+0xfc/0x114)
[<8011f5ec>] (__warn) from [<8011f71c>] (warn_slowpath_null+0x40/0x48)
[<8011f71c>] (warn_slowpath_null) from [<805a476c>] (brcmu_pkt_buf_free_skb+0x3c/0x40)
[<805a476c>] (brcmu_pkt_buf_free_skb) from [<805bb6c4>] (brcmf_fws_cleanup+0x1e4/0x22c)
[<805bb6c4>] (brcmf_fws_cleanup) from [<805bc854>] (brcmf_fws_del_interface+0x58/0x68)
[<805bc854>] (brcmf_fws_del_interface) from [<805b66ac>] (brcmf_remove_interface+0x40/0x150)
[<805b66ac>] (brcmf_remove_interface) from [<805b6870>] (brcmf_detach+0x6c/0xb0)
[<805b6870>] (brcmf_detach) from [<805bdbb8>] (brcmf_usb_disconnect+0x30/0x4c)
[<805bdbb8>] (brcmf_usb_disconnect) from [<805e5d64>] (usb_unbind_interface+0x5c/0x1e0)
[<805e5d64>] (usb_unbind_interface) from [<804aab10>] (device_release_driver_internal+0x154/0x1ec)
[<804aab10>] (device_release_driver_internal) from [<804a97f4>] (bus_remove_device+0xcc/0xf8)
[<804a97f4>] (bus_remove_device) from [<804a6fc0>] (device_del+0x118/0x308)
[<804a6fc0>] (device_del) from [<805e488c>] (usb_disable_device+0xa0/0x1c8)
[<805e488c>] (usb_disable_device) from [<805dcf98>] (usb_disconnect+0x70/0x1d8)
[<805dcf98>] (usb_disconnect) from [<805ddd84>] (hub_event+0x464/0xf50)
[<805ddd84>] (hub_event) from [<80135a70>] (process_one_work+0x138/0x3f8)
[<80135a70>] (process_one_work) from [<80135d5c>] (worker_thread+0x2c/0x554)
[<80135d5c>] (worker_thread) from [<8013b1a0>] (kthread+0x124/0x154)
[<8013b1a0>] (kthread) from [<801010e8>] (ret_from_fork+0x14/0x2c)
Exception stack(0xecf8dfb0 to 0xecf8dff8)
dfa0:                                     00000000 00000000 00000000 00000000
dfc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
dfe0: 00000000 00000000 00000000 00000000 00000013 00000000
---[ end trace 38d234018e9e2a90 ]---
------------[ cut here ]------------
---
 .../broadcom/brcm80211/brcmfmac/fwsignal.c    | 42 +++++++++++--------
 1 file changed, 24 insertions(+), 18 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/fwsignal.c +24 -18
index abeb305492e01..d48b8b2d946fe 100644
@@ -580,24 +580,6 @@ static bool brcmf_fws_ifidx_match(struct sk_buff *skb, void *arg)
 	return ifidx == *(int *)arg;
 }
 
-static void brcmf_fws_psq_flush(struct brcmf_fws_info *fws, struct pktq *q,
-				int ifidx)
-{
-	bool (*matchfn)(struct sk_buff *, void *) = NULL;
-	struct sk_buff *skb;
-	int prec;
-
-	if (ifidx != -1)
-		matchfn = brcmf_fws_ifidx_match;
-	for (prec = 0; prec < q->num_prec; prec++) {
-		skb = brcmu_pktq_pdeq_match(q, prec, matchfn, &ifidx);
-		while (skb) {
-			brcmu_pkt_buf_free_skb(skb);
-			skb = brcmu_pktq_pdeq_match(q, prec, matchfn, &ifidx);
-		}
-	}
-}
-
 static void brcmf_fws_hanger_init(struct brcmf_fws_hanger *hanger)
 {
 	int i;
@@ -669,6 +651,28 @@ static inline int brcmf_fws_hanger_poppkt(struct brcmf_fws_hanger *h,
 	return 0;
 }
 
+static void brcmf_fws_psq_flush(struct brcmf_fws_info *fws, struct pktq *q,
+				int ifidx)
+{
+	bool (*matchfn)(struct sk_buff *, void *) = NULL;
+	struct sk_buff *skb;
+	int prec;
+	u32 hslot;
+
+	if (ifidx != -1)
+		matchfn = brcmf_fws_ifidx_match;
+	for (prec = 0; prec < q->num_prec; prec++) {
+		skb = brcmu_pktq_pdeq_match(q, prec, matchfn, &ifidx);
+		while (skb) {
+			hslot = brcmf_skb_htod_tag_get_field(skb, HSLOT);
+			brcmf_fws_hanger_poppkt(&fws->hanger, hslot, &skb,
+						true);
+			brcmu_pkt_buf_free_skb(skb);
+			skb = brcmu_pktq_pdeq_match(q, prec, matchfn, &ifidx);
+		}
+	}
+}
+
 static int brcmf_fws_hanger_mark_suppressed(struct brcmf_fws_hanger *h,
 					    u32 slot_id)
 {
@@ -2200,6 +2204,8 @@ void brcmf_fws_del_interface(struct brcmf_if *ifp)
 	brcmf_fws_lock(fws);
 	ifp->fws_desc = NULL;
 	brcmf_dbg(TRACE, "deleting %s\n", entry->name);
+	brcmf_fws_macdesc_cleanup(fws, &fws->desc.iface[ifp->ifidx],
+				  ifp->ifidx);
 	brcmf_fws_macdesc_deinit(entry);
 	brcmf_fws_cleanup(fws, ifp->ifidx);
 	brcmf_fws_unlock(fws);
Signature · 1 line
2.20.1
author maintainer
#253 lore raw

PATCH 253/405 brcmfmac: fix race during disconnect when USB completion is in progress

In reply to Greg Kroah-Hartman

[ Upstream commit db3b9e2e1d58080d0754bdf9293dabf8c6491b67 ]

It was observed that rarely during USB disconnect happening shortly after connect (before full initialization completes) usb_hub_wq would wait forever for the dev_init_lock to be unlocked. dev_init_lock would remain locked though because of infinite wait during usb_kill_urb:

[ 2730.656472] kworker/0:2     D    0   260      2 0x00000000
[ 2730.660700] Workqueue: events request_firmware_work_func
[ 2730.664807] [<809dca20>] (__schedule) from [<809dd164>] (schedule+0x4c/0xac)
[ 2730.670587] [<809dd164>] (schedule) from [<8069af44>] (usb_kill_urb+0xdc/0x114)
[ 2730.676815] [<8069af44>] (usb_kill_urb) from [<7f258b50>] (brcmf_usb_free_q+0x34/0xa8 [brcmfmac])
[ 2730.684833] [<7f258b50>] (brcmf_usb_free_q [brcmfmac]) from [<7f2517d4>] (brcmf_detach+0xa0/0xb8 [brcmfmac])
[ 2730.693557] [<7f2517d4>] (brcmf_detach [brcmfmac]) from [<7f251a34>] (brcmf_attach+0xac/0x3d8 [brcmfmac])
[ 2730.702094] [<7f251a34>] (brcmf_attach [brcmfmac]) from [<7f2587ac>] (brcmf_usb_probe_phase2+0x468/0x4a0 [brcmfmac])
[ 2730.711601] [<7f2587ac>] (brcmf_usb_probe_phase2 [brcmfmac]) from [<7f252888>] (brcmf_fw_request_done+0x194/0x220 [brcmfmac])
[ 2730.721795] [<7f252888>] (brcmf_fw_request_done [brcmfmac]) from [<805748e4>] (request_firmware_work_func+0x4c/0x88)
[ 2730.731125] [<805748e4>] (request_firmware_work_func) from [<80141474>] (process_one_work+0x228/0x808)
[ 2730.739223] [<80141474>] (process_one_work) from [<80141a80>] (worker_thread+0x2c/0x564)
[ 2730.746105] [<80141a80>] (worker_thread) from [<80147bcc>] (kthread+0x13c/0x16c)
[ 2730.752227] [<80147bcc>] (kthread) from [<801010b4>] (ret_from_fork+0x14/0x20)
[ 2733.099695] kworker/0:3     D    0  1065      2 0x00000000
[ 2733.103926] Workqueue: usb_hub_wq hub_event
[ 2733.106914] [<809dca20>] (__schedule) from [<809dd164>] (schedule+0x4c/0xac)
[ 2733.112693] [<809dd164>] (schedule) from [<809e2a8c>] (schedule_timeout+0x214/0x3e4)
[ 2733.119621] [<809e2a8c>] (schedule_timeout) from [<809dde2c>] (wait_for_common+0xc4/0x1c0)
[ 2733.126810] [<809dde2c>] (wait_for_common) from [<7f258d00>] (brcmf_usb_disconnect+0x1c/0x4c [brcmfmac])
[ 2733.135206] [<7f258d00>] (brcmf_usb_disconnect [brcmfmac]) from [<8069e0c8>] (usb_unbind_interface+0x5c/0x1e4)
[ 2733.143943] [<8069e0c8>] (usb_unbind_interface) from [<8056d3e8>] (device_release_driver_internal+0x164/0x1fc)
[ 2733.152769] [<8056d3e8>] (device_release_driver_internal) from [<8056c078>] (bus_remove_device+0xd0/0xfc)
[ 2733.161138] [<8056c078>] (bus_remove_device) from [<8056977c>] (device_del+0x11c/0x310)
[ 2733.167939] [<8056977c>] (device_del) from [<8069cba8>] (usb_disable_device+0xa0/0x1cc)
[ 2733.174743] [<8069cba8>] (usb_disable_device) from [<8069507c>] (usb_disconnect+0x74/0x1dc)
[ 2733.181823] [<8069507c>] (usb_disconnect) from [<80695e88>] (hub_event+0x478/0xf88)
[ 2733.188278] [<80695e88>] (hub_event) from [<80141474>] (process_one_work+0x228/0x808)
[ 2733.194905] [<80141474>] (process_one_work) from [<80141a80>] (worker_thread+0x2c/0x564)
[ 2733.201724] [<80141a80>] (worker_thread) from [<80147bcc>] (kthread+0x13c/0x16c)
[ 2733.207913] [<80147bcc>] (kthread) from [<801010b4>] (ret_from_fork+0x14/0x20)

It was traced down to a case where usb_kill_urb would be called on an URB structure containing more or less random data, including large number in its use_count. During the debugging it appeared that in brcmf_usb_free_q() the traversal over URBs' lists is not synchronized with operations on those lists in brcmf_usb_rx_complete() leading to handling brcmf_usbdev_info structure (holding lists' head) as lists' element and in result causing above problem.

Fix it by walking through all URBs during brcmf_cancel_all_urbs using the arrays of requests instead of linked lists.

---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/usb.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/usb.c +8 -2
index a513990cd1d6a..81e1842f1d8c1 100644
@@ -682,12 +682,18 @@ static int brcmf_usb_up(struct device *dev)
 
 static void brcmf_cancel_all_urbs(struct brcmf_usbdev_info *devinfo)
 {
+	int i;
+
 	if (devinfo->ctl_urb)
 		usb_kill_urb(devinfo->ctl_urb);
 	if (devinfo->bulk_urb)
 		usb_kill_urb(devinfo->bulk_urb);
-	brcmf_usb_free_q(&devinfo->tx_postq, true);
-	brcmf_usb_free_q(&devinfo->rx_postq, true);
+	if (devinfo->tx_reqs)
+		for (i = 0; i < devinfo->bus_pub.ntxq; i++)
+			usb_kill_urb(devinfo->tx_reqs[i].urb);
+	if (devinfo->rx_reqs)
+		for (i = 0; i < devinfo->bus_pub.nrxq; i++)
+			usb_kill_urb(devinfo->rx_reqs[i].urb);
 }
 
 static void brcmf_usb_down(struct device *dev)
Signature · 1 line
2.20.1
author maintainer
#254 lore raw

PATCH 254/405 brcmfmac: fix Oops when bringing up interface during USB disconnect

In reply to Greg Kroah-Hartman

[ Upstream commit 24d413a31afaee9bbbf79226052c386b01780ce2 ]

Fix a race which leads to an Oops with NULL pointer dereference. The dereference is in brcmf_config_dongle() when cfg_to_ndev() attempts to get net_device structure of interface with index 0 via if2bss mapping. This shouldn't fail because of check for bus being ready in brcmf_netdev_open(), but it's not synchronised with USB disconnect and there is a race: after the check the bus can be marked down and the mapping for interface 0 may be gone.

Solve this by modifying disconnect handling so that the removal of mapping of ifidx to brcmf_if structure happens after netdev removal (which is synchronous with brcmf_netdev_open() thanks to rtln being locked in devinet_ioctl()). This assures brcmf_netdev_open() returns before the mapping is removed during disconnect.

Unable to handle kernel NULL pointer dereference at virtual address 00000008
pgd = bcae2612
[00000008] *pgd=8be73831
Internal error: Oops: 17 [#1] PREEMPT SMP ARM
Modules linked in: brcmfmac brcmutil nf_log_ipv4 nf_log_common xt_LOG xt_limit

iptable_mangle xt_connmark xt_tcpudp xt_conntrack nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 iptable_filter ip_tables x_tables usb_f_mass_storage usb_f_rndis u_ether usb_serial_simple usbserial cdc_acm smsc95xx usbnet ci_hdrc_imx ci_hdrc usbmisc_imx ulpi 8250_exar 8250_pci 8250 8250_base libcomposite configfs udc_core [last unloaded: brcmutil]

CPU: 2 PID: 24478 Comm: ifconfig Not tainted 4.19.23-00078-ga62866d-dirty #115
Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
PC is at brcmf_cfg80211_up+0x94/0x29c [brcmfmac]
LR is at brcmf_cfg80211_up+0x8c/0x29c [brcmfmac]
pc : [<7f26a91c>]    lr : [<7f26a914>]    psr: a0070013

sp : eca99d28 ip : 00000000 fp : ee9c6c00
r10: 00000036 r9 : 00000000 r8 : ece4002c
r7 : edb5b800 r6 : 00000000 r5 : 80f08448 r4 : edb5b968 r3 : ffffffff r2 : 00000000 r1 : 00000002 r0 : 00000000
Flags: NzCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none
Control: 10c5387d Table: 7ca0c04a DAC: 00000051 Process ifconfig (pid: 24478, stack limit = 0xd9e85a0e)

Stack: (0xeca99d28 to 0xeca9a000)

9d20: 00000000 80f873b0 0000000d 80f08448 eca99d68 50d45f32
9d40: 7f27de94 ece40000 80f08448 80f08448 7f27de94 ece4002c 00000000 00000036
9d60: ee9c6c00 7f27262c 00001002 50d45f32 ece40000 00000000 80f08448 80772008
9d80: 00000001 00001043 00001002 ece40000 00000000 50d45f32 ece40000 00000001
9da0: 80f08448 00001043 00001002 807723d0 00000000 50d45f32 80f08448 eca99e58
9dc0: 80f87113 50d45f32 80f08448 ece40000 ece40138 00001002 80f08448 00000000
9de0: 00000000 80772434 edbd5380 eca99e58 edbd5380 80f08448 ee9c6c0c 80805f70
9e00: 00000000 ede08e00 00008914 ece40000 00000014 ee9c6c0c 600c0013 00001043
9e20: 0208a8c0 ffffffff 00000000 50d45f32 eca98000 80f08448 7ee9fc38 00008914
9e40: 80f68e40 00000051 eca98000 00000036 00000003 80808b9c 6e616c77 00000030
9e60: 00000000 00000000 00001043 0208a8c0 ffffffff 00000000 80f08448 00000000
9e80: 00000000 816d8b20 600c0013 00000001 ede09320 801763d4 00000000 50d45f32
9ea0: eca98000 80f08448 7ee9fc38 50d45f32 00008914 80f08448 7ee9fc38 80f68e40
9ec0: ed531540 8074721c 00000800 00000001 00000000 6e616c77 00000030 00000000
9ee0: 00000000 00001002 0208a8c0 ffffffff 00000000 50d45f32 80f08448 7ee9fc38
9f00: ed531560 ec8fc900 80285a6c 80285138 edb910c0 00000000 ecd91008 ede08e00
9f20: 80f08448 00000000 00000000 816d8b20 600c0013 00000001 ede09320 801763d4
9f40: 00000000 50d45f32 00021000 edb91118 edb910c0 80f08448 01b29000 edb91118
9f60: eca99f7c 50d45f32 00021000 ec8fc900 00000003 ec8fc900 00008914 7ee9fc38
9f80: eca98000 00000036 00000003 80285a6c 00086364 7ee9fe1c 000000c3 00000036
9fa0: 801011c4 80101000 00086364 7ee9fe1c 00000003 00008914 7ee9fc38 00086364
9fc0: 00086364 7ee9fe1c 000000c3 00000036 0008630c 7ee9fe1c 7ee9fc38 00000003
9fe0: 000a42b8 7ee9fbd4 00019914 76e09acc 600c0010 00000003 00000000 00000000 [<7f26a91c>] (brcmf_cfg80211_up [brcmfmac]) from [<7f27262c>] (brcmf_netdev_open+0x74/0xe8 [brcmfmac]) [<7f27262c>] (brcmf_netdev_open [brcmfmac]) from [<80772008>] (__dev_open+0xcc/0x150) [<80772008>] (__dev_open) from [<807723d0>] (__dev_change_flags+0x168/0x1b4) [<807723d0>] (__dev_change_flags) from [<80772434>] (dev_change_flags+0x18/0x48) [<80772434>] (dev_change_flags) from [<80805f70>] (devinet_ioctl+0x67c/0x79c) [<80805f70>] (devinet_ioctl) from [<80808b9c>] (inet_ioctl+0x210/0x3d4)
[<80808b9c>] (inet_ioctl) from [<8074721c>] (sock_ioctl+0x350/0x524)
[<8074721c>] (sock_ioctl) from [<80285138>] (do_vfs_ioctl+0xb0/0x9b0)
[<80285138>] (do_vfs_ioctl) from [<80285a6c>] (ksys_ioctl+0x34/0x5c)
[<80285a6c>] (ksys_ioctl) from [<80101000>] (ret_fast_syscall+0x0/0x28)

Exception stack(0xeca99fa8 to 0xeca99ff0)
9fa0:                   00086364 7ee9fe1c 00000003 00008914 7ee9fc38 00086364
9fc0: 00086364 7ee9fe1c 000000c3 00000036 0008630c 7ee9fe1c 7ee9fc38 00000003
9fe0: 000a42b8 7ee9fbd4 00019914 76e09acc
Code: e5970328 eb002021 e1a02006 e3a01002 (e5909008)
---[ end trace 5cbac2333f3ac5df ]---
---
 .../net/wireless/broadcom/brcm80211/brcmfmac/core.c    | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c +7 -3
index 4fbe8791f6749..24ed19ed116ea 100644
@@ -841,17 +841,17 @@ static void brcmf_del_if(struct brcmf_pub *drvr, s32 bsscfgidx,
 			 bool rtnl_locked)
 {
 	struct brcmf_if *ifp;
+	int ifidx;
 
 	ifp = drvr->iflist[bsscfgidx];
-	drvr->iflist[bsscfgidx] = NULL;
 	if (!ifp) {
 		bphy_err(drvr, "Null interface, bsscfgidx=%d\n", bsscfgidx);
 		return;
 	}
 	brcmf_dbg(TRACE, "Enter, bsscfgidx=%d, ifidx=%d\n", bsscfgidx,
 		  ifp->ifidx);
-	if (drvr->if2bss[ifp->ifidx] == bsscfgidx)
-		drvr->if2bss[ifp->ifidx] = BRCMF_BSSIDX_INVALID;
+	ifidx = ifp->ifidx;
+
 	if (ifp->ndev) {
 		if (bsscfgidx == 0) {
 			if (ifp->ndev->netdev_ops == &brcmf_netdev_ops_pri) {
@@ -879,6 +879,10 @@ static void brcmf_del_if(struct brcmf_pub *drvr, s32 bsscfgidx,
 		brcmf_p2p_ifp_removed(ifp, rtnl_locked);
 		kfree(ifp);
 	}
+
+	drvr->iflist[bsscfgidx] = NULL;
+	if (drvr->if2bss[ifidx] == bsscfgidx)
+		drvr->if2bss[ifidx] = BRCMF_BSSIDX_INVALID;
 }
 
 void brcmf_remove_interface(struct brcmf_if *ifp, bool rtnl_locked)
Signature · 1 line
2.20.1
author maintainer
#255 lore raw

PATCH 255/405 rtc: xgene: fix possible race condition

In reply to Greg Kroah-Hartman

[ Upstream commit a652e00ee1233e251a337c28e18a1da59224e5ce ]

The IRQ is requested before the struct rtc is allocated and registered, but this struct is used in the IRQ handler. This may lead to a NULL pointer dereference.

Switch to devm_rtc_allocate_device/rtc_register_device to allocate the rtc struct before requesting the IRQ.

---
 drivers/rtc/rtc-xgene.c | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)
drivers/rtc/rtc-xgene.c +11 -7
index 153820876a820..2f741f455c30a 100644
@@ -168,6 +168,10 @@ static int xgene_rtc_probe(struct platform_device *pdev)
 	if (IS_ERR(pdata->csr_base))
 		return PTR_ERR(pdata->csr_base);
 
+	pdata->rtc = devm_rtc_allocate_device(&pdev->dev);
+	if (IS_ERR(pdata->rtc))
+		return PTR_ERR(pdata->rtc);
+
 	irq = platform_get_irq(pdev, 0);
 	if (irq < 0) {
 		dev_err(&pdev->dev, "No IRQ resource\n");
@@ -198,15 +202,15 @@ static int xgene_rtc_probe(struct platform_device *pdev)
 		return ret;
 	}
 
-	pdata->rtc = devm_rtc_device_register(&pdev->dev, pdev->name,
-					 &xgene_rtc_ops, THIS_MODULE);
-	if (IS_ERR(pdata->rtc)) {
-		clk_disable_unprepare(pdata->clk);
-		return PTR_ERR(pdata->rtc);
-	}
-
 	/* HW does not support update faster than 1 seconds */
 	pdata->rtc->uie_unsupported = 1;
+	pdata->rtc->ops = &xgene_rtc_ops;
+
+	ret = rtc_register_device(pdata->rtc);
+	if (ret) {
+		clk_disable_unprepare(pdata->clk);
+		return ret;
+	}
 
 	return 0;
 }
Signature · 1 line
2.20.1
author maintainer
#256 lore raw

PATCH 256/405 spi: Add missing error handling for CS GPIOs

In reply to Greg Kroah-Hartman

[ Upstream commit 1723fdec5fcbc4de3d26bbb23a9e1704ee258955 ]

While devm_gpiod_get_index_optional() returns NULL if the GPIO is not present (i.e. -ENOENT), it may still return other error codes, like -EPROBE_DEFER. Currently these are not handled, leading to unrecoverable failures later in case of probe deferral:

gpiod_set_consumer_name: invalid GPIO (errorpointer)
gpiod_direction_output: invalid GPIO (errorpointer)
gpiod_set_value_cansleep: invalid GPIO (errorpointer)
gpiod_set_value_cansleep: invalid GPIO (errorpointer)
gpiod_set_value_cansleep: invalid GPIO (errorpointer)

Detect and propagate errors to fix this.

Fixes f3186dd876697e69 ("spi: Optionally use GPIO descriptors for CS GPIOs")
---
 drivers/spi/spi.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/spi/spi.c +2 -0
index d17f68775a4bb..e3f2e15b75ad4 100644
@@ -2195,6 +2195,8 @@ static int spi_get_gpio_descs(struct spi_controller *ctlr)
 		 */
 		cs[i] = devm_gpiod_get_index_optional(dev, "cs", i,
 						      GPIOD_OUT_LOW);
+		if (IS_ERR(cs[i]))
+			return PTR_ERR(cs[i]);
 
 		if (cs[i]) {
 			/*
Signature · 1 line
2.20.1
author maintainer
#257 lore raw

PATCH 257/405 rtlwifi: fix potential NULL pointer dereference

In reply to Greg Kroah-Hartman

[ Upstream commit 60209d482b97743915883d293c8b85226d230c19 ]

In case dev_alloc_skb fails, the fix safely returns to avoid potential NULL pointer dereference.

---
 drivers/net/wireless/realtek/rtlwifi/rtl8188ee/fw.c       | 2 ++
 drivers/net/wireless/realtek/rtlwifi/rtl8192c/fw_common.c | 2 ++
 drivers/net/wireless/realtek/rtlwifi/rtl8192ee/fw.c       | 2 ++
 drivers/net/wireless/realtek/rtlwifi/rtl8723ae/fw.c       | 2 ++
 drivers/net/wireless/realtek/rtlwifi/rtl8723be/fw.c       | 2 ++
 drivers/net/wireless/realtek/rtlwifi/rtl8821ae/fw.c       | 4 ++++
 6 files changed, 14 insertions(+)
drivers/net/wireless/realtek/rtlwifi/rtl8188ee/fw.c +2 -0
index 203e7b574e845..e2e0bfbc24fe2 100644
@@ -600,6 +600,8 @@ void rtl88e_set_fw_rsvdpagepkt(struct ieee80211_hw *hw, bool b_dl_finished)
 		      u1rsvdpageloc, 3);
 
 	skb = dev_alloc_skb(totalpacketlen);
+	if (!skb)
+		return;
 	skb_put_data(skb, &reserved_page_packet, totalpacketlen);
 
 	rtstatus = rtl_cmd_send_packet(hw, skb);
drivers/net/wireless/realtek/rtlwifi/rtl8192c/fw_common.c +2 -0
index 18c76990a0898..86b1b88cc4ed8 100644
@@ -623,6 +623,8 @@ void rtl92c_set_fw_rsvdpagepkt(struct ieee80211_hw *hw,
 		      u1rsvdpageloc, 3);
 
 	skb = dev_alloc_skb(totalpacketlen);
+	if (!skb)
+		return;
 	skb_put_data(skb, &reserved_page_packet, totalpacketlen);
 
 	if (cmd_send_packet)
drivers/net/wireless/realtek/rtlwifi/rtl8192ee/fw.c +2 -0
index 7c5b54b71a92f..67305ce915ec4 100644
@@ -744,6 +744,8 @@ void rtl92ee_set_fw_rsvdpagepkt(struct ieee80211_hw *hw, bool b_dl_finished)
 		      u1rsvdpageloc, 3);
 
 	skb = dev_alloc_skb(totalpacketlen);
+	if (!skb)
+		return;
 	skb_put_data(skb, &reserved_page_packet, totalpacketlen);
 
 	rtstatus = rtl_cmd_send_packet(hw, skb);
drivers/net/wireless/realtek/rtlwifi/rtl8723ae/fw.c +2 -0
index be451a6f7dbe5..33481232fad01 100644
@@ -448,6 +448,8 @@ void rtl8723e_set_fw_rsvdpagepkt(struct ieee80211_hw *hw, bool b_dl_finished)
 		      u1rsvdpageloc, 3);
 
 	skb = dev_alloc_skb(totalpacketlen);
+	if (!skb)
+		return;
 	skb_put_data(skb, &reserved_page_packet, totalpacketlen);
 
 	rtstatus = rtl_cmd_send_packet(hw, skb);
drivers/net/wireless/realtek/rtlwifi/rtl8723be/fw.c +2 -0
index 4d7fa27f55caa..aa56058af56ef 100644
@@ -562,6 +562,8 @@ void rtl8723be_set_fw_rsvdpagepkt(struct ieee80211_hw *hw,
 		      u1rsvdpageloc, sizeof(u1rsvdpageloc));
 
 	skb = dev_alloc_skb(totalpacketlen);
+	if (!skb)
+		return;
 	skb_put_data(skb, &reserved_page_packet, totalpacketlen);
 
 	rtstatus = rtl_cmd_send_packet(hw, skb);
drivers/net/wireless/realtek/rtlwifi/rtl8821ae/fw.c +4 -0
index dc0eb692088f6..fe32d397d2875 100644
@@ -1623,6 +1623,8 @@ void rtl8812ae_set_fw_rsvdpagepkt(struct ieee80211_hw *hw,
 		      &reserved_page_packet_8812[0], totalpacketlen);
 
 	skb = dev_alloc_skb(totalpacketlen);
+	if (!skb)
+		return;
 	skb_put_data(skb, &reserved_page_packet_8812, totalpacketlen);
 
 	rtstatus = rtl_cmd_send_packet(hw, skb);
@@ -1759,6 +1761,8 @@ void rtl8821ae_set_fw_rsvdpagepkt(struct ieee80211_hw *hw,
 		      &reserved_page_packet_8821[0], totalpacketlen);
 
 	skb = dev_alloc_skb(totalpacketlen);
+	if (!skb)
+		return;
 	skb_put_data(skb, &reserved_page_packet_8821, totalpacketlen);
 
 	rtstatus = rtl_cmd_send_packet(hw, skb);
Signature · 1 line
2.20.1
author maintainer
#258 lore raw

PATCH 258/405 scsi: ufs: Fix regulator load and icc-level configuration

In reply to Greg Kroah-Hartman

[ Upstream commit 0487fff76632ec023d394a05b82e87a971db8c03 ]

Currently if a regulator has "<name>-fixed-regulator" property in device tree, it will skip current limit initialization. This lead to a zero "max_uA" value in struct ufs_vreg.

However, "regulator_set_load" operation shall be required on regulators which have valid current limits, otherwise a zero "max_uA" set by "regulator_set_load" may cause unexpected behavior when this regulator is enabled or set as high power mode.

Similarly, in device's icc_level configuration flow, the target icc_level shall be updated if regulator also has valid current limit, otherwise a wrong icc_level will be calculated by zero "max_uA" and thus causes unexpected results after it is written to device.

---
 drivers/scsi/ufs/ufshcd.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
drivers/scsi/ufs/ufshcd.c +12 -3
index e040f9dd9ff32..58e0bd1dac9b4 100644
@@ -6294,19 +6294,19 @@ static u32 ufshcd_find_max_sup_active_icc_level(struct ufs_hba *hba,
 		goto out;
 	}
 
-	if (hba->vreg_info.vcc)
+	if (hba->vreg_info.vcc && hba->vreg_info.vcc->max_uA)
 		icc_level = ufshcd_get_max_icc_level(
 				hba->vreg_info.vcc->max_uA,
 				POWER_DESC_MAX_ACTV_ICC_LVLS - 1,
 				&desc_buf[PWR_DESC_ACTIVE_LVLS_VCC_0]);
 
-	if (hba->vreg_info.vccq)
+	if (hba->vreg_info.vccq && hba->vreg_info.vccq->max_uA)
 		icc_level = ufshcd_get_max_icc_level(
 				hba->vreg_info.vccq->max_uA,
 				icc_level,
 				&desc_buf[PWR_DESC_ACTIVE_LVLS_VCCQ_0]);
 
-	if (hba->vreg_info.vccq2)
+	if (hba->vreg_info.vccq2 && hba->vreg_info.vccq2->max_uA)
 		icc_level = ufshcd_get_max_icc_level(
 				hba->vreg_info.vccq2->max_uA,
 				icc_level,
@@ -7004,6 +7004,15 @@ static int ufshcd_config_vreg_load(struct device *dev, struct ufs_vreg *vreg,
 	if (!vreg)
 		return 0;
 
+	/*
+	 * "set_load" operation shall be required on those regulators
+	 * which specifically configured current limitation. Otherwise
+	 * zero max_uA may cause unexpected behavior when regulator is
+	 * enabled or set as high power mode.
+	 */
+	if (!vreg->max_uA)
+		return 0;
+
 	ret = regulator_set_load(vreg->reg, ua);
 	if (ret < 0) {
 		dev_err(dev, "%s: %s set load (ua=%d) failed, err=%d\n",
Signature · 1 line
2.20.1
author maintainer
#259 lore raw

PATCH 259/405 scsi: ufs: Avoid configuring regulator with undefined voltage range

In reply to Greg Kroah-Hartman

[ Upstream commit 3b141e8cfd54ba3e5c610717295b2a02aab26a05 ]

For regulators used by UFS, vcc, vccq and vccq2 will have voltage range initialized by ufshcd_populate_vreg(), however other regulators may have undefined voltage range if dt-bindings have no such definition.

In above undefined case, both "min_uV" and "max_uV" fields in ufs_vreg struct will be zero values and these values will be configured on regulators in different power modes.

Currently this may have no harm if both "min_uV" and "max_uV" always keep "zero values" because regulator_set_voltage() will always bypass such invalid values and return "good" results.

However improper values shall be fixed to avoid potential bugs. Simply bypass voltage configuration if voltage range is not defined.

---
 drivers/scsi/ufs/ufshcd.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)
drivers/scsi/ufs/ufshcd.c +8 -5
index 58e0bd1dac9b4..5ba49c8cd2a36 100644
@@ -7048,12 +7048,15 @@ static int ufshcd_config_vreg(struct device *dev,
 	name = vreg->name;
 
 	if (regulator_count_voltages(reg) > 0) {
-		min_uV = on ? vreg->min_uV : 0;
-		ret = regulator_set_voltage(reg, min_uV, vreg->max_uV);
-		if (ret) {
-			dev_err(dev, "%s: %s set voltage failed, err=%d\n",
+		if (vreg->min_uV && vreg->max_uV) {
+			min_uV = on ? vreg->min_uV : 0;
+			ret = regulator_set_voltage(reg, min_uV, vreg->max_uV);
+			if (ret) {
+				dev_err(dev,
+					"%s: %s set voltage failed, err=%d\n",
 					__func__, name, ret);
-			goto out;
+				goto out;
+			}
 		}
 
 		uA_load = on ? vreg->max_uA : 0;
Signature · 1 line
2.20.1
author maintainer
#260 lore raw

PATCH 260/405 drm/panel: otm8009a: Add delay at the end of initialization

In reply to Greg Kroah-Hartman

[ Upstream commit 0084c3c71126fc878c6dab8a6ab8ecc484c2be02 ]

At the end of initialization, a delay is required by the panel. Without this delay, the panel could received a frame early & generate a crash of panel (black screen).

---
 drivers/gpu/drm/panel/panel-orisetech-otm8009a.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/panel/panel-orisetech-otm8009a.c +3 -0
index 87fa316e1d7b0..58ccf648b70fb 100644
@@ -248,6 +248,9 @@ static int otm8009a_init_sequence(struct otm8009a *ctx)
 	/* Send Command GRAM memory write (no parameters) */
 	dcs_write_seq(ctx, MIPI_DCS_WRITE_MEMORY_START);
 
+	/* Wait a short while to let the panel be ready before the 1st frame */
+	mdelay(10);
+
 	return 0;
 }
 
Signature · 1 line
2.20.1
author maintainer
#261 lore raw

PATCH 261/405 drm/amd/display: Prevent cursor hotspot overflow for RV overlay planes

In reply to Greg Kroah-Hartman

[ Upstream commit 6752bea8b03e77c98be7d8d25b0a9d86a00b3cf7 ]

[Why]
The actual position for the cursor on the screen is essentially:

x_out = x - x_plane - x_hotspot
y_out = y - y_plane - y_hotspot

The register values for cursor position and cursor hotspot need to be greater than zero when programmed, but we also need to subtract off the plane position to display the cursor at the correct position.

Since we don't want x or y to be less than zero, we add the plane position as a positive value to x_hotspot or y_hotspot. However, what this doesn't take into account is that the hotspot registers are limited by the maximum cursor size.

On DCN10 the cursor hotspot regitsers are masked to 0xFF, so they have a maximum value of 0-255. Values greater this will wrap, causing the cursor to display in the wrong position.

In practice this means that for sufficiently large plane positions, the cursor will be drawn twice on the screen, and can cause screen flashes or p-state WARNS depending on what the wrapped value is.

So we need a way to remove the value from x_plane and y_plane without exceeding the maximum cursor size.

[How]
Subtract as much as x_plane/y_plane as possible from x and y and place the remainder in the cursor hotspot register.

The value for x_hotspot and y_hotspot can still wrap around but it won't happen in a case where the cursor is actually enabled.

The cursor plane needs to intersect at least one pixel of the plane's rectangle to be enabled, so the cursor position + hotspot provided by userspace must always be strictly less than the maximum cursor size for the cursor to actually be enabled.

---
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c    | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +9 -3
index d1a8f1c302a96..401ea9561618e 100644
@@ -2692,9 +2692,15 @@ static void dcn10_set_cursor_position(struct pipe_ctx *pipe_ctx)
 		.rotation = pipe_ctx->plane_state->rotation,
 		.mirror = pipe_ctx->plane_state->horizontal_mirror
 	};
-
-	pos_cpy.x_hotspot += pipe_ctx->plane_state->dst_rect.x;
-	pos_cpy.y_hotspot += pipe_ctx->plane_state->dst_rect.y;
+	uint32_t x_plane = pipe_ctx->plane_state->dst_rect.x;
+	uint32_t y_plane = pipe_ctx->plane_state->dst_rect.y;
+	uint32_t x_offset = min(x_plane, pos_cpy.x);
+	uint32_t y_offset = min(y_plane, pos_cpy.y);
+
+	pos_cpy.x -= x_offset;
+	pos_cpy.y -= y_offset;
+	pos_cpy.x_hotspot += (x_plane - x_offset);
+	pos_cpy.y_hotspot += (y_plane - y_offset);
 
 	if (pipe_ctx->plane_state->address.type
 			== PLN_ADDR_TYPE_VIDEO_PROGRESSIVE)
Signature · 1 line
2.20.1
author maintainer
#262 lore raw

PATCH 262/405 arm64: cpu_ops: fix a leaked reference by adding missing of_node_put

In reply to Greg Kroah-Hartman

[ Upstream commit 92606ec9285fb84cd9b5943df23f07d741384bfc ]

The call to of_get_next_child returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./arch/arm64/kernel/cpu_ops.c:102:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 69, but without a corresponding object release within this function.

---
 arch/arm64/kernel/cpu_ops.c | 1 +
 1 file changed, 1 insertion(+)
arch/arm64/kernel/cpu_ops.c +1 -0
index ea001241bdd47..00f8b8612b69f 100644
@@ -85,6 +85,7 @@ static const char *__init cpu_read_enable_method(int cpu)
 				pr_err("%pOF: missing enable-method property\n",
 					dn);
 		}
+		of_node_put(dn);
 	} else {
 		enable_method = acpi_get_enable_method(cpu);
 		if (!enable_method) {
Signature · 9 lines
2.20.1




_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
author maintainer
#263 lore raw

PATCH 263/405 locking/static_key: Fix false positive warnings on concurrent dec/inc

In reply to Greg Kroah-Hartman

[ Upstream commit a1247d06d01045d7ab2882a9c074fbf21137c690 ]

Even though the atomic_dec_and_mutex_lock() in __static_key_slow_dec_cpuslocked() can never see a negative value in key->enabled the subsequent sanity check is re-reading key->enabled, which may have been set to -1 in the meantime by static_key_slow_inc_cpuslocked().

CPU A CPU B

__static_key_slow_dec_cpuslocked(): static_key_slow_inc_cpuslocked():

                               # enabled = 1
   atomic_dec_and_mutex_lock()
                               # enabled = 0
                                              atomic_read() == 0
                                              atomic_set(-1)
                               # enabled = -1
   val = atomic_read()
   # Oops - val == -1!

The test case is TCP's clean_acked_data_enable() / clean_acked_data_disable() as tickled by KTLS (net/ktls).

---
 kernel/jump_label.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)
kernel/jump_label.c +13 -8
index bad96b476eb6e..a799b1ac6b2fe 100644
@@ -206,6 +206,8 @@ static void __static_key_slow_dec_cpuslocked(struct static_key *key,
 					   unsigned long rate_limit,
 					   struct delayed_work *work)
 {
+	int val;
+
 	lockdep_assert_cpus_held();
 
 	/*
@@ -215,17 +217,20 @@ static void __static_key_slow_dec_cpuslocked(struct static_key *key,
 	 * returns is unbalanced, because all other static_key_slow_inc()
 	 * instances block while the update is in progress.
 	 */
-	if (!atomic_dec_and_mutex_lock(&key->enabled, &jump_label_mutex)) {
-		WARN(atomic_read(&key->enabled) < 0,
-		     "jump label: negative count!\n");
+	val = atomic_fetch_add_unless(&key->enabled, -1, 1);
+	if (val != 1) {
+		WARN(val < 0, "jump label: negative count!\n");
 		return;
 	}
 
-	if (rate_limit) {
-		atomic_inc(&key->enabled);
-		schedule_delayed_work(work, rate_limit);
-	} else {
-		jump_label_update(key);
+	jump_label_lock();
+	if (atomic_dec_and_test(&key->enabled)) {
+		if (rate_limit) {
+			atomic_inc(&key->enabled);
+			schedule_delayed_work(work, rate_limit);
+		} else {
+			jump_label_update(key);
+		}
 	}
 	jump_label_unlock();
 }
Signature · 1 line
2.20.1
author maintainer
#264 lore raw

PATCH 264/405 wil6210: fix return code of wmi_mgmt_tx and wmi_mgmt_tx_ext

In reply to Greg Kroah-Hartman

[ Upstream commit 49122ec42634f73babb1dc96f170023e5228d080 ]

The functions that send management TX frame have 3 possible
results: success and other side acknowledged receive (ACK=1), success and other side did not acknowledge receive(ACK=0) and failure to send the frame. The current implementation incorrectly reports the ACK=0 case as failure.

---
 drivers/net/wireless/ath/wil6210/cfg80211.c |  5 +++++
 drivers/net/wireless/ath/wil6210/wmi.c      | 11 ++++++-----
 2 files changed, 11 insertions(+), 5 deletions(-)
drivers/net/wireless/ath/wil6210/cfg80211.c +5 -0
index a1e226652b4ab..692730415d781 100644
@@ -1274,7 +1274,12 @@ int wil_cfg80211_mgmt_tx(struct wiphy *wiphy, struct wireless_dev *wdev,
 			     params->wait);
 
 out:
+	/* when the sent packet was not acked by receiver(ACK=0), rc will
+	 * be -EAGAIN. In this case this function needs to return success,
+	 * the ACK=0 will be reflected in tx_status.
+	 */
 	tx_status = (rc == 0);
+	rc = (rc == -EAGAIN) ? 0 : rc;
 	cfg80211_mgmt_tx_status(wdev, cookie ? *cookie : 0, buf, len,
 				tx_status, GFP_KERNEL);
 
drivers/net/wireless/ath/wil6210/wmi.c +6 -5
index bda4a9712f91f..63116f4b62c7f 100644
@@ -3502,8 +3502,9 @@ int wmi_mgmt_tx(struct wil6210_vif *vif, const u8 *buf, size_t len)
 	rc = wmi_call(wil, WMI_SW_TX_REQ_CMDID, vif->mid, cmd, total,
 		      WMI_SW_TX_COMPLETE_EVENTID, &evt, sizeof(evt), 2000);
 	if (!rc && evt.evt.status != WMI_FW_STATUS_SUCCESS) {
-		wil_err(wil, "mgmt_tx failed with status %d\n", evt.evt.status);
-		rc = -EINVAL;
+		wil_dbg_wmi(wil, "mgmt_tx failed with status %d\n",
+			    evt.evt.status);
+		rc = -EAGAIN;
 	}
 
 	kfree(cmd);
@@ -3555,9 +3556,9 @@ int wmi_mgmt_tx_ext(struct wil6210_vif *vif, const u8 *buf, size_t len,
 	rc = wmi_call(wil, WMI_SW_TX_REQ_EXT_CMDID, vif->mid, cmd, total,
 		      WMI_SW_TX_COMPLETE_EVENTID, &evt, sizeof(evt), 2000);
 	if (!rc && evt.evt.status != WMI_FW_STATUS_SUCCESS) {
-		wil_err(wil, "mgmt_tx_ext failed with status %d\n",
-			evt.evt.status);
-		rc = -EINVAL;
+		wil_dbg_wmi(wil, "mgmt_tx_ext failed with status %d\n",
+			    evt.evt.status);
+		rc = -EAGAIN;
 	}
 
 	kfree(cmd);
Signature · 1 line
2.20.1
author maintainer
#265 lore raw

PATCH 265/405 x86/uaccess, ftrace: Fix ftrace_likely_update() vs. SMAP

In reply to Greg Kroah-Hartman

[ Upstream commit 4a6c91fbdef846ec7250b82f2eeeb87ac5f18cf9 ]

For CONFIG_TRACE_BRANCH_PROFILING=y the likely/unlikely things get overloaded and generate callouts to this code, and thus also when AC=1.

Make it safe.

Cc Borislav Petkov <[email protected]>
Cc Josh Poimboeuf <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Steven Rostedt <[email protected]>
Cc Thomas Gleixner <[email protected]>
---
 kernel/trace/trace_branch.c | 4 ++++
 1 file changed, 4 insertions(+)
kernel/trace/trace_branch.c +4 -0
index 4ad967453b6fb..3ea65cdff30d5 100644
@@ -205,6 +205,8 @@ void trace_likely_condition(struct ftrace_likely_data *f, int val, int expect)
 void ftrace_likely_update(struct ftrace_likely_data *f, int val,
 			  int expect, int is_constant)
 {
+	unsigned long flags = user_access_save();
+
 	/* A constant is always correct */
 	if (is_constant) {
 		f->constant++;
@@ -223,6 +225,8 @@ void ftrace_likely_update(struct ftrace_likely_data *f, int val,
 		f->data.correct++;
 	else
 		f->data.incorrect++;
+
+	user_access_restore(flags);
 }
 EXPORT_SYMBOL(ftrace_likely_update);
 
Signature · 1 line
2.20.1
author maintainer
#266 lore raw

PATCH 266/405 iwlwifi: mvm: IBSS: use BE FIFO for multicast

In reply to Greg Kroah-Hartman

[ Upstream commit 192a7e1f731fd9a64216cce35287eb23360437f6 ]

Back in commit 4d339989acd7 ("iwlwifi: mvm: support ibss in dqa mode") we changed queue selection for IBSS to be:

    if (ieee80211_is_probe_resp(fc) || ieee80211_is_auth(fc) ||
        ieee80211_is_deauth(fc))
            return IWL_MVM_DQA_AP_PROBE_RESP_QUEUE;
    if (info->hw_queue == info->control.vif->cab_queue)
            return info->hw_queue;
    return IWL_MVM_DQA_AP_PROBE_RESP_QUEUE;

Clearly, the thought at the time must've been that mac80211 will select the hw_queue as the cab_queue, so that we'll return and use that, where we store the multicast queue for IBSS. This, however, isn't true because mac80211 doesn't implement powersave for IBSS and thus selects the normal IBSS interface AC queue (best effort).

This therefore always used the probe response queue, which maps to the BE FIFO.

In commit cfbc6c4c5b91 ("iwlwifi: mvm: support mac80211 TXQs model") we rethought this code, and as a consequence now started mapping the multicast traffic to the multicast hardware queue since we no longer relied on mac80211 selecting the queue, doing it ourselves instead. This queue is mapped to the MCAST FIFO. however, this isn't actually enabled/controlled by the firmware in IBSS mode because we don't implement powersave, and frames from this queue can never go out in this case.

Therefore, we got queue hang reports such as
https://bugzilla.kernel.org/show_bug.cgi?id=201707

Fix this by mapping the multicast queue to the BE FIFO in IBSS so that all the frames can go out.

Fixes cfbc6c4c5b91 ("iwlwifi: mvm: support mac80211 TXQs model")
---
 drivers/net/wireless/intel/iwlwifi/mvm/sta.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/net/wireless/intel/iwlwifi/mvm/sta.c +2 -1
index 98d123dd71778..eb452e9dce057 100644
@@ -2277,7 +2277,8 @@ int iwl_mvm_add_mcast_sta(struct iwl_mvm *mvm, struct ieee80211_vif *vif)
 	static const u8 _maddr[] = {0x03, 0x00, 0x00, 0x00, 0x00, 0x00};
 	const u8 *maddr = _maddr;
 	struct iwl_trans_txq_scd_cfg cfg = {
-		.fifo = IWL_MVM_TX_FIFO_MCAST,
+		.fifo = vif->type == NL80211_IFTYPE_AP ?
+			IWL_MVM_TX_FIFO_MCAST : IWL_MVM_TX_FIFO_BE,
 		.sta_id = msta->sta_id,
 		.tid = 0,
 		.aggregate = false,
Signature · 1 line
2.20.1
author maintainer
#267 lore raw

PATCH 267/405 x86/uaccess, signal: Fix AC=1 bloat

In reply to Greg Kroah-Hartman

[ Upstream commit 88e4718275c1bddca6f61f300688b4553dc8584b ]

Occasionally GCC is less agressive with inlining and the following is observed:

arch/x86/kernel/signal.o: warning: objtool: restore_sigcontext()+0x3cc: call to force_valid_ss.isra.5() with UACCESS enabled arch/x86/kernel/signal.o: warning: objtool: do_signal()+0x384: call to frame_uc_flags.isra.0() with UACCESS enabled

Cure this by moving this code out of the AC=1 region, since it really isn't needed for the user access.

Cc Borislav Petkov <[email protected]>
Cc Josh Poimboeuf <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Thomas Gleixner <[email protected]>
---
 arch/x86/kernel/signal.c | 29 +++++++++++++++++------------
 1 file changed, 17 insertions(+), 12 deletions(-)
arch/x86/kernel/signal.c +17 -12
index 08dfd4c1a4f95..c8aa58a2bab97 100644
@@ -132,16 +132,6 @@ static int restore_sigcontext(struct pt_regs *regs,
 		COPY_SEG_CPL3(cs);
 		COPY_SEG_CPL3(ss);
 
-#ifdef CONFIG_X86_64
-		/*
-		 * Fix up SS if needed for the benefit of old DOSEMU and
-		 * CRIU.
-		 */
-		if (unlikely(!(uc_flags & UC_STRICT_RESTORE_SS) &&
-			     user_64bit_mode(regs)))
-			force_valid_ss(regs);
-#endif
-
 		get_user_ex(tmpflags, &sc->flags);
 		regs->flags = (regs->flags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS);
 		regs->orig_ax = -1;		/* disable syscall checks */
@@ -150,6 +140,15 @@ static int restore_sigcontext(struct pt_regs *regs,
 		buf = (void __user *)buf_val;
 	} get_user_catch(err);
 
+#ifdef CONFIG_X86_64
+	/*
+	 * Fix up SS if needed for the benefit of old DOSEMU and
+	 * CRIU.
+	 */
+	if (unlikely(!(uc_flags & UC_STRICT_RESTORE_SS) && user_64bit_mode(regs)))
+		force_valid_ss(regs);
+#endif
+
 	err |= fpu__restore_sig(buf, IS_ENABLED(CONFIG_X86_32));
 
 	force_iret();
@@ -461,6 +460,7 @@ static int __setup_rt_frame(int sig, struct ksignal *ksig,
 {
 	struct rt_sigframe __user *frame;
 	void __user *fp = NULL;
+	unsigned long uc_flags;
 	int err = 0;
 
 	frame = get_sigframe(&ksig->ka, regs, sizeof(struct rt_sigframe), &fp);
@@ -473,9 +473,11 @@ static int __setup_rt_frame(int sig, struct ksignal *ksig,
 			return -EFAULT;
 	}
 
+	uc_flags = frame_uc_flags(regs);
+
 	put_user_try {
 		/* Create the ucontext.  */
-		put_user_ex(frame_uc_flags(regs), &frame->uc.uc_flags);
+		put_user_ex(uc_flags, &frame->uc.uc_flags);
 		put_user_ex(0, &frame->uc.uc_link);
 		save_altstack_ex(&frame->uc.uc_stack, regs->sp);
 
@@ -541,6 +543,7 @@ static int x32_setup_rt_frame(struct ksignal *ksig,
 {
 #ifdef CONFIG_X86_X32_ABI
 	struct rt_sigframe_x32 __user *frame;
+	unsigned long uc_flags;
 	void __user *restorer;
 	int err = 0;
 	void __user *fpstate = NULL;
@@ -555,9 +558,11 @@ static int x32_setup_rt_frame(struct ksignal *ksig,
 			return -EFAULT;
 	}
 
+	uc_flags = frame_uc_flags(regs);
+
 	put_user_try {
 		/* Create the ucontext.  */
-		put_user_ex(frame_uc_flags(regs), &frame->uc.uc_flags);
+		put_user_ex(uc_flags, &frame->uc.uc_flags);
 		put_user_ex(0, &frame->uc.uc_link);
 		compat_save_altstack_ex(&frame->uc.uc_stack, regs->sp);
 		put_user_ex(0, &frame->uc.uc__pad0);
Signature · 1 line
2.20.1
author maintainer
#268 lore raw

PATCH 268/405 x86/ia32: Fix ia32_restore_sigcontext() AC leak

In reply to Greg Kroah-Hartman

[ Upstream commit 67a0514afdbb8b2fc70b771b8c77661a9cb9d3a9 ]

Objtool spotted that we call native_load_gs_index() with AC set. Re-arrange the code to avoid that.

Cc Borislav Petkov <[email protected]>
Cc Josh Poimboeuf <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Thomas Gleixner <[email protected]>
---
 arch/x86/ia32/ia32_signal.c | 29 +++++++++++++++++------------
 1 file changed, 17 insertions(+), 12 deletions(-)
arch/x86/ia32/ia32_signal.c +17 -12
index 321fe5f5d0e96..4d5fcd47ab75a 100644
@@ -61,9 +61,8 @@
 } while (0)
 
 #define RELOAD_SEG(seg)		{		\
-	unsigned int pre = GET_SEG(seg);	\
+	unsigned int pre = (seg) | 3;		\
 	unsigned int cur = get_user_seg(seg);	\
-	pre |= 3;				\
 	if (pre != cur)				\
 		set_user_seg(seg, pre);		\
 }
@@ -72,6 +71,7 @@ static int ia32_restore_sigcontext(struct pt_regs *regs,
 				   struct sigcontext_32 __user *sc)
 {
 	unsigned int tmpflags, err = 0;
+	u16 gs, fs, es, ds;
 	void __user *buf;
 	u32 tmp;
 
@@ -79,16 +79,10 @@ static int ia32_restore_sigcontext(struct pt_regs *regs,
 	current->restart_block.fn = do_no_restart_syscall;
 
 	get_user_try {
-		/*
-		 * Reload fs and gs if they have changed in the signal
-		 * handler.  This does not handle long fs/gs base changes in
-		 * the handler, but does not clobber them at least in the
-		 * normal case.
-		 */
-		RELOAD_SEG(gs);
-		RELOAD_SEG(fs);
-		RELOAD_SEG(ds);
-		RELOAD_SEG(es);
+		gs = GET_SEG(gs);
+		fs = GET_SEG(fs);
+		ds = GET_SEG(ds);
+		es = GET_SEG(es);
 
 		COPY(di); COPY(si); COPY(bp); COPY(sp); COPY(bx);
 		COPY(dx); COPY(cx); COPY(ip); COPY(ax);
@@ -106,6 +100,17 @@ static int ia32_restore_sigcontext(struct pt_regs *regs,
 		buf = compat_ptr(tmp);
 	} get_user_catch(err);
 
+	/*
+	 * Reload fs and gs if they have changed in the signal
+	 * handler.  This does not handle long fs/gs base changes in
+	 * the handler, but does not clobber them at least in the
+	 * normal case.
+	 */
+	RELOAD_SEG(gs);
+	RELOAD_SEG(fs);
+	RELOAD_SEG(ds);
+	RELOAD_SEG(es);
+
 	err |= fpu__restore_sig(buf, 1);
 
 	force_iret();
Signature · 1 line
2.20.1
author maintainer
#269 lore raw

PATCH 269/405 x86/uaccess: Fix up the fixup

In reply to Greg Kroah-Hartman

[ Upstream commit b69656fa7ea2f75e47d7bd5b9430359fa46488af ]

New tooling got confused about this:

arch/x86/lib/memcpy_64.o: warning: objtool: .fixup+0x7: return with UACCESS enabled

While the code isn't wrong, it is tedious (if at all possible) to figure out what function a particular chunk of .fixup belongs to.

This then confuses the objtool uaccess validation. Instead of returning directly from the .fixup, jump back into the right function.

Cc Borislav Petkov <[email protected]>
Cc Josh Poimboeuf <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Thomas Gleixner <[email protected]>
---
 arch/x86/lib/memcpy_64.S | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/x86/lib/memcpy_64.S +2 -1
index 3b24dc05251c7..9d05572370edc 100644
@@ -257,6 +257,7 @@ ENTRY(__memcpy_mcsafe)
 	/* Copy successful. Return zero */
 .L_done_memcpy_trap:
 	xorl %eax, %eax
+.L_done:
 	ret
 ENDPROC(__memcpy_mcsafe)
 EXPORT_SYMBOL_GPL(__memcpy_mcsafe)
@@ -273,7 +274,7 @@ EXPORT_SYMBOL_GPL(__memcpy_mcsafe)
 	addl	%edx, %ecx
 .E_trailing_bytes:
 	mov	%ecx, %eax
-	ret
+	jmp	.L_done
 
 	/*
 	 * For write fault handling, given the destination is unaligned,
Signature · 1 line
2.20.1
author maintainer
#270 lore raw

PATCH 270/405 chardev: add additional check for minor range overlap

In reply to Greg Kroah-Hartman

[ Upstream commit de36e16d1557a0b6eb328bc3516359a12ba5c25c ]

Current overlap checking cannot correctly handle a case which is baseminor < existing baseminor && baseminor + minorct > existing baseminor + minorct.

---
 fs/char_dev.c | 6 ++++++
 1 file changed, 6 insertions(+)
fs/char_dev.c +6 -0
index a279c58fe3606..8a63cfa290053 100644
@@ -159,6 +159,12 @@ __register_chrdev_region(unsigned int major, unsigned int baseminor,
 			ret = -EBUSY;
 			goto out;
 		}
+
+		if (new_min < old_min && new_max > old_max) {
+			ret = -EBUSY;
+			goto out;
+		}
+
 	}
 
 	cd->next = *cp;
Signature · 1 line
2.20.1
author maintainer
#271 lore raw

PATCH 271/405 RDMA/hns: Fix bad endianess of port_pd variable

In reply to Greg Kroah-Hartman

[ Upstream commit 6734b2973565e36659e97e12ab0d0faf1d9f3fbe ]

port_pd is treated as le32 in declaration and read, fix assignment to be in le32 too. This change fixes the following compilation warnings.

drivers/infiniband/hw/hns/hns_roce_ah.c:67:24: warning: incorrect type
in assignment (different base types) drivers/infiniband/hw/hns/hns_roce_ah.c:67:24: expected restricted __le32 [usertype] port_pd drivers/infiniband/hw/hns/hns_roce_ah.c:67:24: got restricted __be32 [usertype]

Fixes 9a4435375cd1 ("IB/hns: Add driver files for hns RoCE driver")
---
 drivers/infiniband/hw/hns/hns_roce_ah.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/hw/hns/hns_roce_ah.c +1 -1
index b3c8c45ec1e3e..64e0c69b69c53 100644
@@ -70,7 +70,7 @@ struct ib_ah *hns_roce_create_ah(struct ib_pd *ibpd,
 			     HNS_ROCE_VLAN_SL_BIT_MASK) <<
 			     HNS_ROCE_VLAN_SL_SHIFT;
 
-	ah->av.port_pd = cpu_to_be32(to_hr_pd(ibpd)->pdn |
+	ah->av.port_pd = cpu_to_le32(to_hr_pd(ibpd)->pdn |
 				     (rdma_ah_get_port_num(ah_attr) <<
 				     HNS_ROCE_PORT_NUM_SHIFT));
 	ah->av.gid_index = grh->sgid_index;
Signature · 1 line
2.20.1
author maintainer
#272 lore raw

PATCH 272/405 sh: sh7786: Add explicit I/O cast to sh7786_mm_sel()

In reply to Greg Kroah-Hartman

[ Upstream commit 8440bb9b944c02222c7a840d406141ed42e945cd ]

When compile-testing on arm:

    arch/sh/include/cpu-sh4/cpu/sh7786.h: In function ‘sh7786_mm_sel’:
    arch/sh/include/cpu-sh4/cpu/sh7786.h:135:21: warning: passing argument 1 of ‘__raw_readl’ makes pointer from integer without a cast [-Wint-conversion]
      return __raw_readl(0xFC400020) & 0x7;
			 ^~~~~~~~~~
    In file included from include/linux/io.h:25:0,
		     from arch/sh/include/cpu-sh4/cpu/sh7786.h:14,
		     from drivers/pinctrl/sh-pfc/pfc-sh7786.c:15:
    arch/arm/include/asm/io.h:113:21: note: expected ‘const volatile void *’ but argument is of type ‘unsigned int’
     #define __raw_readl __raw_readl
			 ^
    arch/arm/include/asm/io.h:114:19: note: in expansion of macro ‘__raw_readl’
     static inline u32 __raw_readl(const volatile void __iomem *addr)
		       ^~~~~~~~~~~

__raw_readl() on SuperH is a macro that casts the passed I/O address to the correct type, while the implementations on most other architectures expect to be passed the correct pointer type.

Add an explicit cast to fix this.

Note that this also gets rid of a sparse warning on SuperH:

    arch/sh/include/cpu-sh4/cpu/sh7786.h:135:16: warning: incorrect type in argument 1 (different base types)
    arch/sh/include/cpu-sh4/cpu/sh7786.h:135:16:    expected void const volatile [noderef] <asn:2>*<noident>
    arch/sh/include/cpu-sh4/cpu/sh7786.h:135:16:    got unsigned int
---
 arch/sh/include/cpu-sh4/cpu/sh7786.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/sh/include/cpu-sh4/cpu/sh7786.h +1 -1
index 8f9bfbf3cdb10..d6cce65b48713 100644
@@ -132,7 +132,7 @@ enum {
 
 static inline u32 sh7786_mm_sel(void)
 {
-	return __raw_readl(0xFC400020) & 0x7;
+	return __raw_readl((const volatile void __iomem *)0xFC400020) & 0x7;
 }
 
 #endif /* __CPU_SH7786_H__ */
Signature · 1 line
2.20.1
author maintainer
#273 lore raw

PATCH 273/405 HID: core: move Usage Page concatenation to Main item

In reply to Greg Kroah-Hartman

[ Upstream commit 58e75155009cc800005629955d3482f36a1e0eec ]

As seen on some USB wireless keyboards manufactured by Primax, the HID parser was using some assumptions that are not always true. In this case it's s the fact that, inside the scope of a main item, an Usage Page will always precede an Usage.

The spec is not pretty clear as 6.2.2.7 states "Any usage that follows is interpreted as a Usage ID and concatenated with the Usage Page". While 6.2.2.8 states "When the parser encounters a main item it concatenates the last declared Usage Page with a Usage to form a complete usage value." Being somewhat contradictory it was decided to match Window's implementation, which follows 6.2.2.8.

In summary, the patch moves the Usage Page concatenation from the local item parsing function to the main item parsing function.

---
 drivers/hid/hid-core.c | 36 ++++++++++++++++++++++++------------
 include/linux/hid.h    |  1 +
 2 files changed, 25 insertions(+), 12 deletions(-)
drivers/hid/hid-core.c +24 -12
index 860e21ec6a492..63a43726cce0f 100644
@@ -218,13 +218,14 @@ static unsigned hid_lookup_collection(struct hid_parser *parser, unsigned type)
  * Add a usage to the temporary parser table.
  */
 
-static int hid_add_usage(struct hid_parser *parser, unsigned usage)
+static int hid_add_usage(struct hid_parser *parser, unsigned usage, u8 size)
 {
 	if (parser->local.usage_index >= HID_MAX_USAGES) {
 		hid_err(parser->device, "usage index exceeded\n");
 		return -1;
 	}
 	parser->local.usage[parser->local.usage_index] = usage;
+	parser->local.usage_size[parser->local.usage_index] = size;
 	parser->local.collection_index[parser->local.usage_index] =
 		parser->collection_stack_ptr ?
 		parser->collection_stack[parser->collection_stack_ptr - 1] : 0;
@@ -486,10 +487,7 @@ static int hid_parser_local(struct hid_parser *parser, struct hid_item *item)
 			return 0;
 		}
 
-		if (item->size <= 2)
-			data = (parser->global.usage_page << 16) + data;
-
-		return hid_add_usage(parser, data);
+		return hid_add_usage(parser, data, item->size);
 
 	case HID_LOCAL_ITEM_TAG_USAGE_MINIMUM:
 
@@ -498,9 +496,6 @@ static int hid_parser_local(struct hid_parser *parser, struct hid_item *item)
 			return 0;
 		}
 
-		if (item->size <= 2)
-			data = (parser->global.usage_page << 16) + data;
-
 		parser->local.usage_minimum = data;
 		return 0;
 
@@ -511,9 +506,6 @@ static int hid_parser_local(struct hid_parser *parser, struct hid_item *item)
 			return 0;
 		}
 
-		if (item->size <= 2)
-			data = (parser->global.usage_page << 16) + data;
-
 		count = data - parser->local.usage_minimum;
 		if (count + parser->local.usage_index >= HID_MAX_USAGES) {
 			/*
@@ -533,7 +525,7 @@ static int hid_parser_local(struct hid_parser *parser, struct hid_item *item)
 		}
 
 		for (n = parser->local.usage_minimum; n <= data; n++)
-			if (hid_add_usage(parser, n)) {
+			if (hid_add_usage(parser, n, item->size)) {
 				dbg_hid("hid_add_usage failed\n");
 				return -1;
 			}
@@ -547,6 +539,22 @@ static int hid_parser_local(struct hid_parser *parser, struct hid_item *item)
 	return 0;
 }
 
+/*
+ * Concatenate Usage Pages into Usages where relevant:
+ * As per specification, 6.2.2.8: "When the parser encounters a main item it
+ * concatenates the last declared Usage Page with a Usage to form a complete
+ * usage value."
+ */
+
+static void hid_concatenate_usage_page(struct hid_parser *parser)
+{
+	int i;
+
+	for (i = 0; i < parser->local.usage_index; i++)
+		if (parser->local.usage_size[i] <= 2)
+			parser->local.usage[i] += parser->global.usage_page << 16;
+}
+
 /*
  * Process a main item.
  */
@@ -556,6 +564,8 @@ static int hid_parser_main(struct hid_parser *parser, struct hid_item *item)
 	__u32 data;
 	int ret;
 
+	hid_concatenate_usage_page(parser);
+
 	data = item_udata(item);
 
 	switch (item->tag) {
@@ -765,6 +775,8 @@ static int hid_scan_main(struct hid_parser *parser, struct hid_item *item)
 	__u32 data;
 	int i;
 
+	hid_concatenate_usage_page(parser);
+
 	data = item_udata(item);
 
 	switch (item->tag) {
include/linux/hid.h +1 -0
index f9707d1dcb584..ac0c70b4ce10a 100644
@@ -417,6 +417,7 @@ struct hid_global {
 
 struct hid_local {
 	unsigned usage[HID_MAX_USAGES]; /* usage array */
+	u8 usage_size[HID_MAX_USAGES]; /* usage size array */
 	unsigned collection_index[HID_MAX_USAGES]; /* collection index array */
 	unsigned usage_index;
 	unsigned usage_minimum;
Signature · 1 line
2.20.1
author maintainer
#274 lore raw

PATCH 274/405 ASoC: eukrea-tlv320: fix a leaked reference by adding missing of_node_put

In reply to Greg Kroah-Hartman

[ Upstream commit b820d52e7eed7b30b2dfef5f4213a2bc3cbea6f3 ]

The call to of_parse_phandle returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./sound/soc/fsl/eukrea-tlv320.c:121:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 102, but without a correspo nding object release within this function.
./sound/soc/fsl/eukrea-tlv320.c:127:3-9: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 102, but without a correspo nding object release within this function.

---
 sound/soc/fsl/eukrea-tlv320.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
sound/soc/fsl/eukrea-tlv320.c +2 -2
index 191426a6d9adf..30a3d68b5c033 100644
@@ -118,13 +118,13 @@ static int eukrea_tlv320_probe(struct platform_device *pdev)
 		if (ret) {
 			dev_err(&pdev->dev,
 				"fsl,mux-int-port node missing or invalid.\n");
-			return ret;
+			goto err;
 		}
 		ret = of_property_read_u32(np, "fsl,mux-ext-port", &ext_port);
 		if (ret) {
 			dev_err(&pdev->dev,
 				"fsl,mux-ext-port node missing or invalid.\n");
-			return ret;
+			goto err;
 		}
 
 		/*
Signature · 1 line
2.20.1
author maintainer
#275 lore raw

PATCH 275/405 ASoC: fsl_utils: fix a leaked reference by adding missing of_node_put

In reply to Greg Kroah-Hartman

[ Upstream commit c705247136a523488eac806bd357c3e5d79a7acd ]

The call to of_parse_phandle returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./sound/soc/fsl/fsl_utils.c:74:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 38, but without a corresponding object release within this function.

---
 sound/soc/fsl/fsl_utils.c | 1 +
 1 file changed, 1 insertion(+)
sound/soc/fsl/fsl_utils.c +1 -0
index 9981668ab5909..040d06b89f00a 100644
@@ -71,6 +71,7 @@ int fsl_asoc_get_dma_channel(struct device_node *ssi_np,
 	iprop = of_get_property(dma_np, "cell-index", NULL);
 	if (!iprop) {
 		of_node_put(dma_np);
+		of_node_put(dma_channel_np);
 		return -EINVAL;
 	}
 	*dma_id = be32_to_cpup(iprop);
Signature · 1 line
2.20.1
author maintainer
#276 lore raw

PATCH 276/405 ASoC: wcd9335: fix a leaked reference by adding missing of_node_put

In reply to Greg Kroah-Hartman

[ Upstream commit 64b92de9603f22b5455da925ee57268ef7fb4e80 ]

The call to of_parse_phandle returns a node pointer with refcount incremented thus it must be explicitly decremented after the last usage.

Detected by coccinelle with the following warnings:
./sound/soc/codecs/wcd9335.c:5193:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 5183, but without a correspon ding object release within this function.

Cc Liam Girdwood <[email protected]>
Cc Mark Brown <[email protected]>
Cc Jaroslav Kysela <[email protected]>
Cc Takashi Iwai <[email protected]>
Cc Srinivas Kandagatla <[email protected]>
Cc Vinod Koul <[email protected]>
Cc Dan Carpenter <[email protected]> (commit_signer:1/11=9%,authored:1/11=9%)
---
 sound/soc/codecs/wcd9335.c | 1 +
 1 file changed, 1 insertion(+)
sound/soc/codecs/wcd9335.c +1 -0
index 981f88a5f6154..a04a7cedd99de 100644
@@ -5188,6 +5188,7 @@ static int wcd9335_slim_status(struct slim_device *sdev,
 
 	wcd->slim = sdev;
 	wcd->slim_ifc_dev = of_slim_get_device(sdev->ctrl, ifc_dev_np);
+	of_node_put(ifc_dev_np);
 	if (!wcd->slim_ifc_dev) {
 		dev_err(dev, "Unable to get SLIM Interface device\n");
 		return -EINVAL;
Signature · 1 line
2.20.1
author maintainer
#277 lore raw

PATCH 277/405 cxgb3/l2t: Fix undefined behaviour

In reply to Greg Kroah-Hartman

[ Upstream commit 76497732932f15e7323dc805e8ea8dc11bb587cf ]

The use of zero-sized array causes undefined behaviour when it is not the last member in a structure. As it happens to be in this case.

Also, the current code makes use of a language extension to the C90 standard, but the preferred mechanism to declare variable-length types such as this one is a flexible array member, introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning in case the flexible array does not occur last. Which is beneficial to cultivate a high-quality code.

Fixes e48f129c2f20 ("[SCSI] cxgb3i: convert cdev->l2opt to use rcu to prevent NULL dereference")
---
 drivers/net/ethernet/chelsio/cxgb3/l2t.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/chelsio/cxgb3/l2t.h +1 -1
index c2fd323c40782..ea75f275023ff 100644
@@ -75,8 +75,8 @@ struct l2t_data {
 	struct l2t_entry *rover;	/* starting point for next allocation */
 	atomic_t nfree;		/* number of free entries */
 	rwlock_t lock;
-	struct l2t_entry l2tab[0];
 	struct rcu_head rcu_head;	/* to handle rcu cleanup */
+	struct l2t_entry l2tab[];
 };
 
 typedef void (*arp_failure_handler_func)(struct t3cdev * dev,
Signature · 1 line
2.20.1
author maintainer
#278 lore raw

PATCH 278/405 clk: renesas: rcar-gen3: Correct parent clock of SYS-DMAC

In reply to Greg Kroah-Hartman

[ Upstream commit 3c772f71a552d343a96868ed9a809f9047be94f5 ]

The clock sources of the AXI BUS clock (266.66 MHz) used for SYS-DMAC DMA transfers are:

Channel R-Car H3 R-Car M3-W R-Car M3-N
-------------------------------------------------
SYS-DMAC0 S0D3 S0D3 S0D3
SYS-DMAC1 S3D1 S3D1 S3D1
SYS-DMAC2 S3D1 S3D1 S3D1

As a result, change the parent clocks of the SYS-DMAC{1,2} module clocks on R-Car H3, R-Car M3-W, and R-Car M3-N to S3D1.

NOTE: This information will be reflected in a future revision of the
R-Car Gen3 Hardware Manual.

---
 drivers/clk/renesas/r8a774a1-cpg-mssr.c | 4 ++--
 drivers/clk/renesas/r8a7795-cpg-mssr.c  | 4 ++--
 drivers/clk/renesas/r8a7796-cpg-mssr.c  | 4 ++--
 drivers/clk/renesas/r8a77965-cpg-mssr.c | 4 ++--
 4 files changed, 8 insertions(+), 8 deletions(-)
drivers/clk/renesas/r8a774a1-cpg-mssr.c +2 -2
index 4d92b27a61538..047599579c651 100644
@@ -123,8 +123,8 @@ static const struct mssr_mod_clk r8a774a1_mod_clks[] __initconst = {
 	DEF_MOD("msiof2",		 209,	R8A774A1_CLK_MSO),
 	DEF_MOD("msiof1",		 210,	R8A774A1_CLK_MSO),
 	DEF_MOD("msiof0",		 211,	R8A774A1_CLK_MSO),
-	DEF_MOD("sys-dmac2",		 217,	R8A774A1_CLK_S0D3),
-	DEF_MOD("sys-dmac1",		 218,	R8A774A1_CLK_S0D3),
+	DEF_MOD("sys-dmac2",		 217,	R8A774A1_CLK_S3D1),
+	DEF_MOD("sys-dmac1",		 218,	R8A774A1_CLK_S3D1),
 	DEF_MOD("sys-dmac0",		 219,	R8A774A1_CLK_S0D3),
 	DEF_MOD("cmt3",			 300,	R8A774A1_CLK_R),
 	DEF_MOD("cmt2",			 301,	R8A774A1_CLK_R),
drivers/clk/renesas/r8a7795-cpg-mssr.c +2 -2
index 86842c9fd314e..eade38e9ed36b 100644
@@ -129,8 +129,8 @@ static struct mssr_mod_clk r8a7795_mod_clks[] __initdata = {
 	DEF_MOD("msiof2",		 209,	R8A7795_CLK_MSO),
 	DEF_MOD("msiof1",		 210,	R8A7795_CLK_MSO),
 	DEF_MOD("msiof0",		 211,	R8A7795_CLK_MSO),
-	DEF_MOD("sys-dmac2",		 217,	R8A7795_CLK_S0D3),
-	DEF_MOD("sys-dmac1",		 218,	R8A7795_CLK_S0D3),
+	DEF_MOD("sys-dmac2",		 217,	R8A7795_CLK_S3D1),
+	DEF_MOD("sys-dmac1",		 218,	R8A7795_CLK_S3D1),
 	DEF_MOD("sys-dmac0",		 219,	R8A7795_CLK_S0D3),
 	DEF_MOD("sceg-pub",		 229,	R8A7795_CLK_CR),
 	DEF_MOD("cmt3",			 300,	R8A7795_CLK_R),
drivers/clk/renesas/r8a7796-cpg-mssr.c +2 -2
index 12c455859f2c2..654f3ea88f335 100644
@@ -126,8 +126,8 @@ static const struct mssr_mod_clk r8a7796_mod_clks[] __initconst = {
 	DEF_MOD("msiof2",		 209,	R8A7796_CLK_MSO),
 	DEF_MOD("msiof1",		 210,	R8A7796_CLK_MSO),
 	DEF_MOD("msiof0",		 211,	R8A7796_CLK_MSO),
-	DEF_MOD("sys-dmac2",		 217,	R8A7796_CLK_S0D3),
-	DEF_MOD("sys-dmac1",		 218,	R8A7796_CLK_S0D3),
+	DEF_MOD("sys-dmac2",		 217,	R8A7796_CLK_S3D1),
+	DEF_MOD("sys-dmac1",		 218,	R8A7796_CLK_S3D1),
 	DEF_MOD("sys-dmac0",		 219,	R8A7796_CLK_S0D3),
 	DEF_MOD("cmt3",			 300,	R8A7796_CLK_R),
 	DEF_MOD("cmt2",			 301,	R8A7796_CLK_R),
drivers/clk/renesas/r8a77965-cpg-mssr.c +2 -2
index eb1cca58a1e1f..13d1f88be04a5 100644
@@ -123,8 +123,8 @@ static const struct mssr_mod_clk r8a77965_mod_clks[] __initconst = {
 	DEF_MOD("msiof2",		209,	R8A77965_CLK_MSO),
 	DEF_MOD("msiof1",		210,	R8A77965_CLK_MSO),
 	DEF_MOD("msiof0",		211,	R8A77965_CLK_MSO),
-	DEF_MOD("sys-dmac2",		217,	R8A77965_CLK_S0D3),
-	DEF_MOD("sys-dmac1",		218,	R8A77965_CLK_S0D3),
+	DEF_MOD("sys-dmac2",		217,	R8A77965_CLK_S3D1),
+	DEF_MOD("sys-dmac1",		218,	R8A77965_CLK_S3D1),
 	DEF_MOD("sys-dmac0",		219,	R8A77965_CLK_S0D3),
 
 	DEF_MOD("cmt3",			300,	R8A77965_CLK_R),
Signature · 1 line
2.20.1
author maintainer
#279 lore raw

PATCH 279/405 block: avoid to break XEN by multi-page bvec

In reply to Greg Kroah-Hartman

[ Upstream commit db5ebd6edd2627d7e81a031643cf43587f63e66c ]

XEN has special page merge requirement, see xen_biovec_phys_mergeable(). We can't merge pages into one bvec simply for XEN.

So move XEN's specific check on page merge into __bio_try_merge_page(), then abvoid to break XEN by multi-page bvec.

---
 block/bio.c | 2 ++
 1 file changed, 2 insertions(+)
block/bio.c +2 -0
index 716510ecd7ffa..a3c80a6c1fe51 100644
@@ -776,6 +776,8 @@ bool __bio_try_merge_page(struct bio *bio, struct page *page,
 
 		if (vec_end_addr + 1 != page_addr + off)
 			return false;
+		if (xen_domain() && !xen_biovec_phys_mergeable(bv, page))
+			return false;
 		if (same_page && (vec_end_addr & PAGE_MASK) != page_addr)
 			return false;
 
Signature · 9 lines
2.20.1




_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
author maintainer
#280 lore raw

PATCH 280/405 block: pass page to xen_biovec_phys_mergeable

In reply to Greg Kroah-Hartman

[ Upstream commit 0383ad4374f7ad7edd925a2ee4753035c3f5508a ]

xen_biovec_phys_mergeable() only needs .bv_page of the 2nd bio bvec for checking if the two bvecs can be merged, so pass page to xen_biovec_phys_mergeable() directly.

No function change.

---
 block/blk.h            | 2 +-
 drivers/xen/biomerge.c | 5 +++--
 include/xen/xen.h      | 4 +++-
 3 files changed, 7 insertions(+), 4 deletions(-)
block/blk.h +1 -1
index 5d636ee416630..e27fd1512e4bb 100644
@@ -75,7 +75,7 @@ static inline bool biovec_phys_mergeable(struct request_queue *q,
 
 	if (addr1 + vec1->bv_len != addr2)
 		return false;
-	if (xen_domain() && !xen_biovec_phys_mergeable(vec1, vec2))
+	if (xen_domain() && !xen_biovec_phys_mergeable(vec1, vec2->bv_page))
 		return false;
 	if ((addr1 | mask) != ((addr2 + vec2->bv_len - 1) | mask))
 		return false;
drivers/xen/biomerge.c +3 -2
index f3fbb700f5697..05a286d24f148 100644
@@ -4,12 +4,13 @@
 #include <xen/xen.h>
 #include <xen/page.h>
 
+/* check if @page can be merged with 'vec1' */
 bool xen_biovec_phys_mergeable(const struct bio_vec *vec1,
-			       const struct bio_vec *vec2)
+			       const struct page *page)
 {
 #if XEN_PAGE_SIZE == PAGE_SIZE
 	unsigned long bfn1 = pfn_to_bfn(page_to_pfn(vec1->bv_page));
-	unsigned long bfn2 = pfn_to_bfn(page_to_pfn(vec2->bv_page));
+	unsigned long bfn2 = pfn_to_bfn(page_to_pfn(page));
 
 	return bfn1 + PFN_DOWN(vec1->bv_offset + vec1->bv_len) == bfn2;
 #else
include/xen/xen.h +3 -1
index 19d032373de5a..19a72f591e2bd 100644
@@ -43,8 +43,10 @@ extern struct hvm_start_info pvh_start_info;
 #endif	/* CONFIG_XEN_DOM0 */
 
 struct bio_vec;
+struct page;
+
 bool xen_biovec_phys_mergeable(const struct bio_vec *vec1,
-		const struct bio_vec *vec2);
+		const struct page *page);
 
 #if defined(CONFIG_MEMORY_HOTPLUG) && defined(CONFIG_XEN_BALLOON)
 extern u64 xen_saved_max_mem_size;
Signature · 9 lines
2.20.1




_______________________________________________
Xen-devel mailing list
[email protected]
https://lists.xenproject.org/mailman/listinfo/xen-devel
author maintainer
#281 lore raw

PATCH 281/405 clk: renesas: rcar-gen3: Correct parent clock of Audio-DMAC

In reply to Greg Kroah-Hartman

[ Upstream commit b9df2ea2b8d09ad850afe4d4a0403cb23d9e0c02 ]

The clock sources of the AXI-bus clock (266.66 MHz) used for Audio-DMAC DMA transfers are:

Channel R-Car H3 R-Car M3-W R-Car M3-N R-Car E3
---------------------------------------------------------------
Audio-DMAC0 S1D2 S1D2 S1D2 S1D2
Audio-DMAC1 S1D2 S1D2 S1D2 -

As a result, change the parent clocks of the Audio-DMAC{0,1} module clocks on R-Car H3, R-Car M3-W, and R-Car M3-N to S1D2, and change the parent clock of the Audio-DMAC0 module on R-Car E3 to S1D2.

NOTE: This information will be reflected in a future revision of the
R-Car Gen3 Hardware Manual.

[geert: Update R-Car D3, RZ/G2M, and RZ/G2E]
---
 drivers/clk/renesas/r8a774a1-cpg-mssr.c | 4 ++--
 drivers/clk/renesas/r8a774c0-cpg-mssr.c | 2 +-
 drivers/clk/renesas/r8a7795-cpg-mssr.c  | 4 ++--
 drivers/clk/renesas/r8a7796-cpg-mssr.c  | 4 ++--
 drivers/clk/renesas/r8a77965-cpg-mssr.c | 4 ++--
 drivers/clk/renesas/r8a77990-cpg-mssr.c | 2 +-
 drivers/clk/renesas/r8a77995-cpg-mssr.c | 2 +-
 7 files changed, 11 insertions(+), 11 deletions(-)
drivers/clk/renesas/r8a774a1-cpg-mssr.c +2 -2
index 047599579c651..7a4c5957939a5 100644
@@ -143,8 +143,8 @@ static const struct mssr_mod_clk r8a774a1_mod_clks[] __initconst = {
 	DEF_MOD("rwdt",			 402,	R8A774A1_CLK_R),
 	DEF_MOD("intc-ex",		 407,	R8A774A1_CLK_CP),
 	DEF_MOD("intc-ap",		 408,	R8A774A1_CLK_S0D3),
-	DEF_MOD("audmac1",		 501,	R8A774A1_CLK_S0D3),
-	DEF_MOD("audmac0",		 502,	R8A774A1_CLK_S0D3),
+	DEF_MOD("audmac1",		 501,	R8A774A1_CLK_S1D2),
+	DEF_MOD("audmac0",		 502,	R8A774A1_CLK_S1D2),
 	DEF_MOD("hscif4",		 516,	R8A774A1_CLK_S3D1),
 	DEF_MOD("hscif3",		 517,	R8A774A1_CLK_S3D1),
 	DEF_MOD("hscif2",		 518,	R8A774A1_CLK_S3D1),
drivers/clk/renesas/r8a774c0-cpg-mssr.c +1 -1
index 34e274f2a273a..93dacd826fd04 100644
@@ -157,7 +157,7 @@ static const struct mssr_mod_clk r8a774c0_mod_clks[] __initconst = {
 	DEF_MOD("intc-ex",		 407,	R8A774C0_CLK_CP),
 	DEF_MOD("intc-ap",		 408,	R8A774C0_CLK_S0D3),
 
-	DEF_MOD("audmac0",		 502,	R8A774C0_CLK_S3D4),
+	DEF_MOD("audmac0",		 502,	R8A774C0_CLK_S1D2),
 	DEF_MOD("hscif4",		 516,	R8A774C0_CLK_S3D1C),
 	DEF_MOD("hscif3",		 517,	R8A774C0_CLK_S3D1C),
 	DEF_MOD("hscif2",		 518,	R8A774C0_CLK_S3D1C),
drivers/clk/renesas/r8a7795-cpg-mssr.c +2 -2
index eade38e9ed36b..0825cd0ff2866 100644
@@ -153,8 +153,8 @@ static struct mssr_mod_clk r8a7795_mod_clks[] __initdata = {
 	DEF_MOD("rwdt",			 402,	R8A7795_CLK_R),
 	DEF_MOD("intc-ex",		 407,	R8A7795_CLK_CP),
 	DEF_MOD("intc-ap",		 408,	R8A7795_CLK_S0D3),
-	DEF_MOD("audmac1",		 501,	R8A7795_CLK_S0D3),
-	DEF_MOD("audmac0",		 502,	R8A7795_CLK_S0D3),
+	DEF_MOD("audmac1",		 501,	R8A7795_CLK_S1D2),
+	DEF_MOD("audmac0",		 502,	R8A7795_CLK_S1D2),
 	DEF_MOD("drif7",		 508,	R8A7795_CLK_S3D2),
 	DEF_MOD("drif6",		 509,	R8A7795_CLK_S3D2),
 	DEF_MOD("drif5",		 510,	R8A7795_CLK_S3D2),
drivers/clk/renesas/r8a7796-cpg-mssr.c +2 -2
index 654f3ea88f335..997cd956f12bc 100644
@@ -146,8 +146,8 @@ static const struct mssr_mod_clk r8a7796_mod_clks[] __initconst = {
 	DEF_MOD("rwdt",			 402,	R8A7796_CLK_R),
 	DEF_MOD("intc-ex",		 407,	R8A7796_CLK_CP),
 	DEF_MOD("intc-ap",		 408,	R8A7796_CLK_S0D3),
-	DEF_MOD("audmac1",		 501,	R8A7796_CLK_S0D3),
-	DEF_MOD("audmac0",		 502,	R8A7796_CLK_S0D3),
+	DEF_MOD("audmac1",		 501,	R8A7796_CLK_S1D2),
+	DEF_MOD("audmac0",		 502,	R8A7796_CLK_S1D2),
 	DEF_MOD("drif7",		 508,	R8A7796_CLK_S3D2),
 	DEF_MOD("drif6",		 509,	R8A7796_CLK_S3D2),
 	DEF_MOD("drif5",		 510,	R8A7796_CLK_S3D2),
drivers/clk/renesas/r8a77965-cpg-mssr.c +2 -2
index 13d1f88be04a5..afc9c72fa0940 100644
@@ -146,8 +146,8 @@ static const struct mssr_mod_clk r8a77965_mod_clks[] __initconst = {
 	DEF_MOD("intc-ex",		407,	R8A77965_CLK_CP),
 	DEF_MOD("intc-ap",		408,	R8A77965_CLK_S0D3),
 
-	DEF_MOD("audmac1",		501,	R8A77965_CLK_S0D3),
-	DEF_MOD("audmac0",		502,	R8A77965_CLK_S0D3),
+	DEF_MOD("audmac1",		501,	R8A77965_CLK_S1D2),
+	DEF_MOD("audmac0",		502,	R8A77965_CLK_S1D2),
 	DEF_MOD("drif7",		508,	R8A77965_CLK_S3D2),
 	DEF_MOD("drif6",		509,	R8A77965_CLK_S3D2),
 	DEF_MOD("drif5",		510,	R8A77965_CLK_S3D2),
drivers/clk/renesas/r8a77990-cpg-mssr.c +1 -1
index 9a278c75c918c..03f445d47ef69 100644
@@ -152,7 +152,7 @@ static const struct mssr_mod_clk r8a77990_mod_clks[] __initconst = {
 	DEF_MOD("intc-ex",		 407,	R8A77990_CLK_CP),
 	DEF_MOD("intc-ap",		 408,	R8A77990_CLK_S0D3),
 
-	DEF_MOD("audmac0",		 502,	R8A77990_CLK_S3D4),
+	DEF_MOD("audmac0",		 502,	R8A77990_CLK_S1D2),
 	DEF_MOD("drif7",		 508,	R8A77990_CLK_S3D2),
 	DEF_MOD("drif6",		 509,	R8A77990_CLK_S3D2),
 	DEF_MOD("drif5",		 510,	R8A77990_CLK_S3D2),
drivers/clk/renesas/r8a77995-cpg-mssr.c +1 -1
index eee3874865a95..68707277b17b4 100644
@@ -133,7 +133,7 @@ static const struct mssr_mod_clk r8a77995_mod_clks[] __initconst = {
 	DEF_MOD("rwdt",			 402,	R8A77995_CLK_R),
 	DEF_MOD("intc-ex",		 407,	R8A77995_CLK_CP),
 	DEF_MOD("intc-ap",		 408,	R8A77995_CLK_S1D2),
-	DEF_MOD("audmac0",		 502,	R8A77995_CLK_S3D1),
+	DEF_MOD("audmac0",		 502,	R8A77995_CLK_S1D2),
 	DEF_MOD("hscif3",		 517,	R8A77995_CLK_S3D1C),
 	DEF_MOD("hscif0",		 520,	R8A77995_CLK_S3D1C),
 	DEF_MOD("thermal",		 522,	R8A77995_CLK_CP),
Signature · 1 line
2.20.1
author maintainer
#282 lore raw

PATCH 282/405 HID: logitech-hidpp: change low battery level threshold from 31 to 30 percent

In reply to Greg Kroah-Hartman

[ Upstream commit 1f87b0cd32b3456d7efdfb017fcf74d0bfe3ec29 ]

According to hidpp20_batterylevel_get_battery_info my Logitech K270 keyboard reports only 2 battery levels. This matches with what I've seen after testing with batteries at varying level of fullness, it always reports either 5% or 30%.

Windows reports "battery good" for the 30% level. I've captured an USB trace of Windows reading the battery and it is getting the same info as the Linux hidpp code gets.

Now that Linux handles these devices as hidpp devices, it reports the battery as being low as it treats anything under 31% as low, this leads to the user constantly getting a "Keyboard battery is low" warning from GNOME3, which is very annoying.

This commit fixes this by changing the low threshold to anything under 30%, which I assume is what Windows does.

---
 drivers/hid/hid-logitech-hidpp.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/hid/hid-logitech-hidpp.c +5 -1
index ffd30c7492df8..e74fa990ba133 100644
@@ -1021,7 +1021,11 @@ static int hidpp_map_battery_level(int capacity)
 {
 	if (capacity < 11)
 		return POWER_SUPPLY_CAPACITY_LEVEL_CRITICAL;
-	else if (capacity < 31)
+	/*
+	 * The spec says this should be < 31 but some devices report 30
+	 * with brand new batteries and Windows reports 30 as "Good".
+	 */
+	else if (capacity < 30)
 		return POWER_SUPPLY_CAPACITY_LEVEL_LOW;
 	else if (capacity < 81)
 		return POWER_SUPPLY_CAPACITY_LEVEL_NORMAL;
Signature · 1 line
2.20.1
author maintainer
#283 lore raw

PATCH 283/405 spi: tegra114: reset controller on probe

In reply to Greg Kroah-Hartman

[ Upstream commit 019194933339b3e9b486639c8cb3692020844d65 ]

Fixes: SPI driver can be built as module so perform SPI controller reset on probe to make sure it is in valid state before initiating transfer.

---
 drivers/spi/spi-tegra114.c | 32 ++++++++++++++++++--------------
 1 file changed, 18 insertions(+), 14 deletions(-)
drivers/spi/spi-tegra114.c +18 -14
index a76acedd7e2f4..a1888dc6a938a 100644
@@ -1067,27 +1067,19 @@ static int tegra_spi_probe(struct platform_device *pdev)
 
 	spi_irq = platform_get_irq(pdev, 0);
 	tspi->irq = spi_irq;
-	ret = request_threaded_irq(tspi->irq, tegra_spi_isr,
-			tegra_spi_isr_thread, IRQF_ONESHOT,
-			dev_name(&pdev->dev), tspi);
-	if (ret < 0) {
-		dev_err(&pdev->dev, "Failed to register ISR for IRQ %d\n",
-					tspi->irq);
-		goto exit_free_master;
-	}
 
 	tspi->clk = devm_clk_get(&pdev->dev, "spi");
 	if (IS_ERR(tspi->clk)) {
 		dev_err(&pdev->dev, "can not get clock\n");
 		ret = PTR_ERR(tspi->clk);
-		goto exit_free_irq;
+		goto exit_free_master;
 	}
 
 	tspi->rst = devm_reset_control_get_exclusive(&pdev->dev, "spi");
 	if (IS_ERR(tspi->rst)) {
 		dev_err(&pdev->dev, "can not get reset\n");
 		ret = PTR_ERR(tspi->rst);
-		goto exit_free_irq;
+		goto exit_free_master;
 	}
 
 	tspi->max_buf_size = SPI_FIFO_DEPTH << 2;
@@ -1095,7 +1087,7 @@ static int tegra_spi_probe(struct platform_device *pdev)
 
 	ret = tegra_spi_init_dma_param(tspi, true);
 	if (ret < 0)
-		goto exit_free_irq;
+		goto exit_free_master;
 	ret = tegra_spi_init_dma_param(tspi, false);
 	if (ret < 0)
 		goto exit_rx_dma_free;
@@ -1117,18 +1109,32 @@ static int tegra_spi_probe(struct platform_device *pdev)
 		dev_err(&pdev->dev, "pm runtime get failed, e = %d\n", ret);
 		goto exit_pm_disable;
 	}
+
+	reset_control_assert(tspi->rst);
+	udelay(2);
+	reset_control_deassert(tspi->rst);
 	tspi->def_command1_reg  = SPI_M_S;
 	tegra_spi_writel(tspi, tspi->def_command1_reg, SPI_COMMAND1);
 	pm_runtime_put(&pdev->dev);
+	ret = request_threaded_irq(tspi->irq, tegra_spi_isr,
+				   tegra_spi_isr_thread, IRQF_ONESHOT,
+				   dev_name(&pdev->dev), tspi);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "Failed to register ISR for IRQ %d\n",
+			tspi->irq);
+		goto exit_pm_disable;
+	}
 
 	master->dev.of_node = pdev->dev.of_node;
 	ret = devm_spi_register_master(&pdev->dev, master);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "can not register to master err %d\n", ret);
-		goto exit_pm_disable;
+		goto exit_free_irq;
 	}
 	return ret;
 
+exit_free_irq:
+	free_irq(spi_irq, tspi);
 exit_pm_disable:
 	pm_runtime_disable(&pdev->dev);
 	if (!pm_runtime_status_suspended(&pdev->dev))
@@ -1136,8 +1142,6 @@ static int tegra_spi_probe(struct platform_device *pdev)
 	tegra_spi_deinit_dma_param(tspi, false);
 exit_rx_dma_free:
 	tegra_spi_deinit_dma_param(tspi, true);
-exit_free_irq:
-	free_irq(spi_irq, tspi);
 exit_free_master:
 	spi_master_put(master);
 	return ret;
Signature · 1 line
2.20.1
author maintainer
#284 lore raw

PATCH 284/405 habanalabs: prevent CPU soft lockup on Palladium

In reply to Greg Kroah-Hartman

[ Upstream commit e850b89f50d2c1439f58d547b888ee6e43312dea ]

Unmapping ptes in the device MMU on Palladium can take a long time, which can cause a kernel BUG of CPU soft lockup.

This patch minimize the chances for this bug by sleeping a little between unmapping ptes.

---
 drivers/misc/habanalabs/memory.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
drivers/misc/habanalabs/memory.c +9 -2
index ce1fda40a8b81..fadaf557603f5 100644
@@ -1046,10 +1046,17 @@ static int unmap_device_va(struct hl_ctx *ctx, u64 vaddr)
 
 	mutex_lock(&ctx->mmu_lock);
 
-	for (i = 0 ; i < phys_pg_pack->npages ; i++, next_vaddr += page_size)
+	for (i = 0 ; i < phys_pg_pack->npages ; i++, next_vaddr += page_size) {
 		if (hl_mmu_unmap(ctx, next_vaddr, page_size))
 			dev_warn_ratelimited(hdev->dev,
-				"unmap failed for vaddr: 0x%llx\n", next_vaddr);
+			"unmap failed for vaddr: 0x%llx\n", next_vaddr);
+
+		/* unmapping on Palladium can be really long, so avoid a CPU
+		 * soft lockup bug by sleeping a little between unmapping pages
+		 */
+		if (hdev->pldm)
+			usleep_range(500, 1000);
+	}
 
 	hdev->asic_funcs->mmu_invalidate_cache(hdev, true);
 
Signature · 1 line
2.20.1
author maintainer
#285 lore raw

PATCH 285/405 kobject: Dont trigger kobject_uevent(KOBJ_REMOVE) twice.

In reply to Greg Kroah-Hartman

[ Upstream commit c03a0fd0b609e2f5c669c2b7f27c8e1928e9196e ]

syzbot is hitting use-after-free bug in uinput module [1]. This is because kobject_uevent(KOBJ_REMOVE) is called again due to commit 0f4dafc0563c6c49 ("Kobject: auto-cleanup on final unref") after memory allocation fault injection made kobject_uevent(KOBJ_REMOVE) from device_del() from input_unregister_device() fail, while uinput_destroy_device() is expecting that kobject_uevent(KOBJ_REMOVE) is not called after device_del() from input_unregister_device() completed.

That commit intended to catch cases where nobody even attempted to send "remove" uevents. But there is no guarantee that an event will ultimately be sent. We are at the point of no return as far as the rest of the kernel is concerned; there are no repeats or do-overs.

Also, it is not clear whether some subsystem depends on that commit. If no subsystem depends on that commit, it will be better to remove the state_{add,remove}_uevent_sent logic. But we don't want to risk a regression (in a patch which will be backported) by trying to remove that logic. Therefore, as a first step, let's avoid the use-after-free bug by making sure that kobject_uevent(KOBJ_REMOVE) won't be triggered twice.

[1] https://syzkaller.appspot.com/bug?id=8b17c134fe938bbddd75a45afaa9e68af43a362d

Fixes 0f4dafc0563c6c49 ("Kobject: auto-cleanup on final unref")
Cc Kay Sievers <[email protected]>
---
 lib/kobject_uevent.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
lib/kobject_uevent.c +7 -4
index f05802687ba4d..7998affa45d49 100644
@@ -466,6 +466,13 @@ int kobject_uevent_env(struct kobject *kobj, enum kobject_action action,
 	int i = 0;
 	int retval = 0;
 
+	/*
+	 * Mark "remove" event done regardless of result, for some subsystems
+	 * do not want to re-trigger "remove" event via automatic cleanup.
+	 */
+	if (action == KOBJ_REMOVE)
+		kobj->state_remove_uevent_sent = 1;
+
 	pr_debug("kobject: '%s' (%p): %s\n",
 		 kobject_name(kobj), kobj, __func__);
 
@@ -567,10 +574,6 @@ int kobject_uevent_env(struct kobject *kobj, enum kobject_action action,
 		kobj->state_add_uevent_sent = 1;
 		break;
 
-	case KOBJ_REMOVE:
-		kobj->state_remove_uevent_sent = 1;
-		break;
-
 	case KOBJ_UNBIND:
 		zap_modalias_env(env);
 		break;
Signature · 1 line
2.20.1
author maintainer
#286 lore raw

PATCH 286/405 media: video-mux: fix null pointer dereferences

In reply to Greg Kroah-Hartman

[ Upstream commit aeb0d0f581e2079868e64a2e5ee346d340376eae ]

devm_kcalloc may fail and return a null pointer. The fix returns -ENOMEM upon failures to avoid null pointer dereferences.

---
 drivers/media/platform/video-mux.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/media/platform/video-mux.c +5 -0
index 0ba30756e1e40..d8cd5f5cb10d6 100644
@@ -419,9 +419,14 @@ static int video_mux_probe(struct platform_device *pdev)
 	vmux->active = -1;
 	vmux->pads = devm_kcalloc(dev, num_pads, sizeof(*vmux->pads),
 				  GFP_KERNEL);
+	if (!vmux->pads)
+		return -ENOMEM;
+
 	vmux->format_mbus = devm_kcalloc(dev, num_pads,
 					 sizeof(*vmux->format_mbus),
 					 GFP_KERNEL);
+	if (!vmux->format_mbus)
+		return -ENOMEM;
 
 	for (i = 0; i < num_pads; i++) {
 		vmux->pads[i].flags = (i < num_pads - 1) ? MEDIA_PAD_FL_SINK
Signature · 1 line
2.20.1
author maintainer
#287 lore raw

PATCH 287/405 media: wl128x: prevent two potential buffer overflows

In reply to Greg Kroah-Hartman

[ Upstream commit 9c2ccc324b3a6cbc865ab8b3e1a09e93d3c8ade9 ]

Smatch marks skb->data as untrusted so it warns that "evt_hdr->dlen" can copy up to 255 bytes and we only have room for two bytes. Even if this comes from the firmware and we trust it, the new policy generally is just to fix it as kernel hardenning.

I can't test this code so I tried to be very conservative. I considered not allowing "evt_hdr->dlen == 1" because it doesn't initialize the whole variable but in the end I decided to allow it and manually initialized "asic_id" and "asic_ver" to zero.

Fixes e8454ff7b9a4 ("[media] drivers:media:radio: wl128x: FM Driver Common sources")
---
 drivers/media/radio/wl128x/fmdrv_common.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/media/radio/wl128x/fmdrv_common.c +5 -2
index 3c8987af37725..ac5706b4cab84 100644
@@ -489,7 +489,8 @@ int fmc_send_cmd(struct fmdev *fmdev, u8 fm_op, u16 type, void *payload,
 		return -EIO;
 	}
 	/* Send response data to caller */
-	if (response != NULL && response_len != NULL && evt_hdr->dlen) {
+	if (response != NULL && response_len != NULL && evt_hdr->dlen &&
+	    evt_hdr->dlen <= payload_len) {
 		/* Skip header info and copy only response data */
 		skb_pull(skb, sizeof(struct fm_event_msg_hdr));
 		memcpy(response, skb->data, evt_hdr->dlen);
@@ -583,6 +584,8 @@ static void fm_irq_handle_flag_getcmd_resp(struct fmdev *fmdev)
 		return;
 
 	fm_evt_hdr = (void *)skb->data;
+	if (fm_evt_hdr->dlen > sizeof(fmdev->irq_info.flag))
+		return;
 
 	/* Skip header info and copy only response data */
 	skb_pull(skb, sizeof(struct fm_event_msg_hdr));
@@ -1308,7 +1311,7 @@ static int load_default_rx_configuration(struct fmdev *fmdev)
 static int fm_power_up(struct fmdev *fmdev, u8 mode)
 {
 	u16 payload;
-	__be16 asic_id, asic_ver;
+	__be16 asic_id = 0, asic_ver = 0;
 	int resp_len, ret;
 	u8 fw_name[50];
 
Signature · 1 line
2.20.1
author maintainer
#288 lore raw

PATCH 288/405 media: gspca: Kill URBs on USB device disconnect

In reply to Greg Kroah-Hartman

[ Upstream commit 9b9ea7c2b57a0c9c3341fc6db039d1f7971a432e ]

In order to prevent ISOC URBs from being infinitely resubmitted, the driver's USB disconnect handler must kill all the in-flight URBs.

While here, change the URB packet status message to a debug level, to avoid spamming the console too much.

This commit fixes a lockup caused by an interrupt storm coming from the URB completion handler.

---
 drivers/media/usb/gspca/gspca.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/media/usb/gspca/gspca.c +3 -1
index ac70b36d67b7b..128935f2a217e 100644
@@ -294,7 +294,7 @@ static void fill_frame(struct gspca_dev *gspca_dev,
 		/* check the packet status and length */
 		st = urb->iso_frame_desc[i].status;
 		if (st) {
-			pr_err("ISOC data error: [%d] len=%d, status=%d\n",
+			gspca_dbg(gspca_dev, D_PACK, "ISOC data error: [%d] len=%d, status=%d\n",
 			       i, len, st);
 			gspca_dev->last_packet_type = DISCARD_PACKET;
 			continue;
@@ -1638,6 +1638,8 @@ void gspca_disconnect(struct usb_interface *intf)
 
 	mutex_lock(&gspca_dev->usb_lock);
 	gspca_dev->present = false;
+	destroy_urbs(gspca_dev);
+	gspca_input_destroy_urb(gspca_dev);
 
 	vb2_queue_error(&gspca_dev->queue);
 
Signature · 1 line
2.20.1
author maintainer
#289 lore raw

PATCH 289/405 efifb: Omit memory map check on legacy boot

In reply to Greg Kroah-Hartman

[ Upstream commit c2999c281ea2d2ebbdfce96cecc7b52e2ae7c406 ]

Since the following commit:

38ac0287b7f4 ("fbdev/efifb: Honour UEFI memory map attributes when mapping the FB")

efifb_probe() checks its memory range via efi_mem_desc_lookup(), and this leads to a spurious error message:

EFI_MEMMAP is not enabled

at every boot on KVM. This is quite annoying since the error message appears even if you set "quiet" boot option.

Since this happens on legacy boot, which strangely enough exposes a EFI framebuffer via screen_info, let's double check that we are doing an EFI boot before attempting to access the EFI memory map.

---
 drivers/video/fbdev/efifb.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/video/fbdev/efifb.c +2 -1
index fd02e8a4841d6..9f39f0c360e0c 100644
@@ -464,7 +464,8 @@ static int efifb_probe(struct platform_device *dev)
 	info->apertures->ranges[0].base = efifb_fix.smem_start;
 	info->apertures->ranges[0].size = size_remap;
 
-	if (!efi_mem_desc_lookup(efifb_fix.smem_start, &md)) {
+	if (efi_enabled(EFI_BOOT) &&
+	    !efi_mem_desc_lookup(efifb_fix.smem_start, &md)) {
 		if ((efifb_fix.smem_start + efifb_fix.smem_len) >
 		    (md.phys_addr + (md.num_pages << EFI_PAGE_SHIFT))) {
 			pr_err("efifb: video memory @ 0x%lx spans multiple EFI memory regions\n",
Signature · 1 line
2.20.1
author maintainer
#290 lore raw

PATCH 290/405 media: mtk-vcodec: fix access to incorrect planes member

In reply to Greg Kroah-Hartman

[ Upstream commit 52fafc58c3535c9f4f53864686dbaee3bcbadcb4 ]

Commit 0650a91499e0 ("media: mtk-vcodec: Correct return type for mem2mem buffer helpers") fixed the return types for mem2mem buffer helper functions by changing a few local variables from vb2_buffer to vb2_v4l2_buffer. However, it left a few accesses to vb2_buffer::planes as-is, accidentally turning them into accesses to vb2_v4l2_buffer::planes and resulting in values being read from/written to the wrong place.

Fix this by inserting vb2_buf into these accesses so they mimic their original behavior.

Fixes 0650a91499e0 ("media: mtk-vcodec: Correct return type for mem2mem buffer helpers")
---
 drivers/media/platform/mtk-vcodec/mtk_vcodec_dec.c |  4 ++--
 drivers/media/platform/mtk-vcodec/mtk_vcodec_enc.c | 10 +++++-----
 2 files changed, 7 insertions(+), 7 deletions(-)
drivers/media/platform/mtk-vcodec/mtk_vcodec_dec.c +2 -2
index d022c65bb34c2..49babf994cb75 100644
@@ -388,7 +388,7 @@ static void mtk_vdec_worker(struct work_struct *work)
 	}
 	buf.va = vb2_plane_vaddr(&src_buf->vb2_buf, 0);
 	buf.dma_addr = vb2_dma_contig_plane_dma_addr(&src_buf->vb2_buf, 0);
-	buf.size = (size_t)src_buf->planes[0].bytesused;
+	buf.size = (size_t)src_buf->vb2_buf.planes[0].bytesused;
 	if (!buf.va) {
 		v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx);
 		mtk_v4l2_err("[%d] id=%d src_addr is NULL!!",
@@ -1155,7 +1155,7 @@ static void vb2ops_vdec_buf_queue(struct vb2_buffer *vb)
 
 	src_mem.va = vb2_plane_vaddr(&src_buf->vb2_buf, 0);
 	src_mem.dma_addr = vb2_dma_contig_plane_dma_addr(&src_buf->vb2_buf, 0);
-	src_mem.size = (size_t)src_buf->planes[0].bytesused;
+	src_mem.size = (size_t)src_buf->vb2_buf.planes[0].bytesused;
 	mtk_v4l2_debug(2,
 			"[%d] buf id=%d va=%p dma=%pad size=%zx",
 			ctx->id, src_buf->index,
drivers/media/platform/mtk-vcodec/mtk_vcodec_enc.c +5 -5
index c6b48b5925fbe..50351adafc470 100644
@@ -894,7 +894,7 @@ static void vb2ops_venc_stop_streaming(struct vb2_queue *q)
 
 	if (q->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) {
 		while ((dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx))) {
-			dst_buf->planes[0].bytesused = 0;
+			dst_buf->vb2_buf.planes[0].bytesused = 0;
 			v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_ERROR);
 		}
 	} else {
@@ -947,7 +947,7 @@ static int mtk_venc_encode_header(void *priv)
 
 	bs_buf.va = vb2_plane_vaddr(&dst_buf->vb2_buf, 0);
 	bs_buf.dma_addr = vb2_dma_contig_plane_dma_addr(&dst_buf->vb2_buf, 0);
-	bs_buf.size = (size_t)dst_buf->planes[0].length;
+	bs_buf.size = (size_t)dst_buf->vb2_buf.planes[0].length;
 
 	mtk_v4l2_debug(1,
 			"[%d] buf id=%d va=0x%p dma_addr=0x%llx size=%zu",
@@ -976,7 +976,7 @@ static int mtk_venc_encode_header(void *priv)
 	}
 
 	ctx->state = MTK_STATE_HEADER;
-	dst_buf->planes[0].bytesused = enc_result.bs_size;
+	dst_buf->vb2_buf.planes[0].bytesused = enc_result.bs_size;
 	v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_DONE);
 
 	return 0;
@@ -1107,12 +1107,12 @@ static void mtk_venc_worker(struct work_struct *work)
 
 	if (ret) {
 		v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_ERROR);
-		dst_buf->planes[0].bytesused = 0;
+		dst_buf->vb2_buf.planes[0].bytesused = 0;
 		v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_ERROR);
 		mtk_v4l2_err("venc_if_encode failed=%d", ret);
 	} else {
 		v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_DONE);
-		dst_buf->planes[0].bytesused = enc_result.bs_size;
+		dst_buf->vb2_buf.planes[0].bytesused = enc_result.bs_size;
 		v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_DONE);
 		mtk_v4l2_debug(2, "venc_if_encode bs size=%d",
 				 enc_result.bs_size);
Signature · 1 line
2.20.1
author maintainer
#291 lore raw

PATCH 291/405 thunderbolt: property: Fix a missing check of kzalloc

In reply to Greg Kroah-Hartman

[ Upstream commit 6183d5a51866f3acdeeb66b75e87d44025b01a55 ]

No check is enforced for the return value of kzalloc, which may lead to NULL-pointer dereference.

The patch fixes this issue.

---
 drivers/thunderbolt/property.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/thunderbolt/property.c +6 -1
index b2f0d6386ceea..ead18c532b53d 100644
@@ -578,7 +578,12 @@ int tb_property_add_text(struct tb_property_dir *parent, const char *key,
 		return -ENOMEM;
 
 	property->length = size / 4;
-	property->value.data = kzalloc(size, GFP_KERNEL);
+	property->value.text = kzalloc(size, GFP_KERNEL);
+	if (!property->value.text) {
+		kfree(property);
+		return -ENOMEM;
+	}
+
 	strcpy(property->value.text, text);
 
 	list_add_tail(&property->list, &parent->properties);
Signature · 1 line
2.20.1
author maintainer
#292 lore raw

PATCH 292/405 thunderbolt: Fix to check the return value of kmemdup

In reply to Greg Kroah-Hartman

[ Upstream commit fd21b79e541e4666c938a344f3ad2df74b4f5120 ]

uuid in add_switch is allocted via kmemdup which can fail. The patch logs the error and cleans up the allocated memory for switch.

---
 drivers/thunderbolt/icm.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/thunderbolt/icm.c +5 -0
index e3fc920af6825..8b7f9131e9d12 100644
@@ -473,6 +473,11 @@ static void add_switch(struct tb_switch *parent_sw, u64 route,
 		goto out;
 
 	sw->uuid = kmemdup(uuid, sizeof(*uuid), GFP_KERNEL);
+	if (!sw->uuid) {
+		tb_sw_warn(sw, "cannot allocate memory for switch\n");
+		tb_switch_put(sw);
+		goto out;
+	}
 	sw->connection_id = connection_id;
 	sw->connection_key = connection_key;
 	sw->link = link;
Signature · 1 line
2.20.1
author maintainer
#293 lore raw

PATCH 293/405 drm: rcar-du: lvds: Set LVEN and LVRES bits together on D3

In reply to Greg Kroah-Hartman

[ Upstream commit 00d082cc4ea6e42ec4fed832a1020231bb1ca150 ]

On the D3 SoC the LVDS PHY must be enabled in the same register write that enables the LVDS output. Skip writing the LVEN bit independently on that platform, it will be set by the write that sets LVRES.

---
 drivers/gpu/drm/rcar-du/rcar_lvds.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/gpu/drm/rcar-du/rcar_lvds.c +6 -2
index 7ef97b2a6edaa..f0314790333ba 100644
@@ -485,9 +485,13 @@ static void rcar_lvds_enable(struct drm_bridge *bridge)
 	}
 
 	if (lvds->info->quirks & RCAR_LVDS_QUIRK_GEN3_LVEN) {
-		/* Turn on the LVDS PHY. */
+		/*
+		 * Turn on the LVDS PHY. On D3, the LVEN and LVRES bit must be
+		 * set at the same time, so don't write the register yet.
+		 */
 		lvdcr0 |= LVDCR0_LVEN;
-		rcar_lvds_write(lvds, LVDCR0, lvdcr0);
+		if (!(lvds->info->quirks & RCAR_LVDS_QUIRK_PWD))
+			rcar_lvds_write(lvds, LVDCR0, lvdcr0);
 	}
 
 	if (!(lvds->info->quirks & RCAR_LVDS_QUIRK_EXT_PLL)) {
Signature · 1 line
2.20.1
author maintainer
#294 lore raw

PATCH 294/405 drm: rcar-du: lvds: Fix post-DLL divider calculation

In reply to Greg Kroah-Hartman

[ Upstream commit 167e535438ecc73d299340bb1269616432020dfb ]

The PLL parameters are computed by looping over the range of acceptable M, N and E values, and selecting the combination that produces the output frequency closest to the target. The internal frequency constraints are taken into account by restricting the tested values for the PLL parameters, reducing the search space. The target frequency, however, is only taken into account when computing the post-PLL divider, which can result in a 0 value for the divider when the PLL output frequency being tested is lower than half of the target frequency. Subsequent loops will produce a better set of PLL parameters, but for some of the iterations this can result in a division by 0.

Fix it by clamping the divider value. We could instead restrict the E values being tested in the inner loop, but that would require additional calculation that would likely be less efficient as the E parameter can only take three different values.

Fixes c25c01361199 ("drm: rcar-du: lvds: D3/E3 support")
---
 drivers/gpu/drm/rcar-du/rcar_lvds.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/rcar-du/rcar_lvds.c +1 -1
index f0314790333ba..033f44e46daf4 100644
@@ -283,7 +283,7 @@ static void rcar_lvds_d3_e3_pll_calc(struct rcar_lvds *lvds, struct clk *clk,
 				 * divider.
 				 */
 				fout = fvco / (1 << e) / div7;
-				div = DIV_ROUND_CLOSEST(fout, target);
+				div = max(1UL, DIV_ROUND_CLOSEST(fout, target));
 				diff = abs(fout / div - target);
 
 				if (diff < pll->diff) {
Signature · 1 line
2.20.1
author maintainer
#295 lore raw

PATCH 295/405 timekeeping: Force upper bound for setting CLOCK_REALTIME

In reply to Greg Kroah-Hartman

[ Upstream commit 7a8e61f8478639072d402a26789055a4a4de8f77 ]

Several people reported testing failures after setting CLOCK_REALTIME close to the limits of the kernel internal representation in nanoseconds, i.e. year 2262.

The failures are exposed in subsequent operations, i.e. when arming timers or when the advancing CLOCK_MONOTONIC makes the calculation of CLOCK_REALTIME overflow into negative space.

Now people start to paper over the underlying problem by clamping calculations to the valid range, but that's just wrong because such workarounds will prevent detection of real issues as well.

It is reasonable to force an upper bound for the various methods of setting CLOCK_REALTIME. Year 2262 is the absolute upper bound. Assume a maximum uptime of 30 years which is plenty enough even for esoteric embedded systems. That results in an upper bound of year 2232 for setting the time.

Once that limit is reached in reality this limit is only a small part of the problem space. But until then this stops people from trying to paper over the problem at the wrong places.

---
 include/linux/time64.h    | 21 +++++++++++++++++++++
 kernel/time/time.c        |  2 +-
 kernel/time/timekeeping.c |  6 +++---
 3 files changed, 25 insertions(+), 4 deletions(-)
include/linux/time64.h +21 -0
index f38d382ffec13..a620ee610b9f3 100644
@@ -33,6 +33,17 @@ struct itimerspec64 {
 #define KTIME_MAX			((s64)~((u64)1 << 63))
 #define KTIME_SEC_MAX			(KTIME_MAX / NSEC_PER_SEC)
 
+/*
+ * Limits for settimeofday():
+ *
+ * To prevent setting the time close to the wraparound point time setting
+ * is limited so a reasonable uptime can be accomodated. Uptime of 30 years
+ * should be really sufficient, which means the cutoff is 2232. At that
+ * point the cutoff is just a small part of the larger problem.
+ */
+#define TIME_UPTIME_SEC_MAX		(30LL * 365 * 24 *3600)
+#define TIME_SETTOD_SEC_MAX		(KTIME_SEC_MAX - TIME_UPTIME_SEC_MAX)
+
 static inline int timespec64_equal(const struct timespec64 *a,
 				   const struct timespec64 *b)
 {
@@ -100,6 +111,16 @@ static inline bool timespec64_valid_strict(const struct timespec64 *ts)
 	return true;
 }
 
+static inline bool timespec64_valid_settod(const struct timespec64 *ts)
+{
+	if (!timespec64_valid(ts))
+		return false;
+	/* Disallow values which cause overflow issues vs. CLOCK_REALTIME */
+	if ((unsigned long long)ts->tv_sec >= TIME_SETTOD_SEC_MAX)
+		return false;
+	return true;
+}
+
 /**
  * timespec64_to_ns - Convert timespec64 to nanoseconds
  * @ts:		pointer to the timespec64 variable to be converted
kernel/time/time.c +1 -1
index c3f756f8534bb..86656bbac232e 100644
@@ -171,7 +171,7 @@ int do_sys_settimeofday64(const struct timespec64 *tv, const struct timezone *tz
 	static int firsttime = 1;
 	int error = 0;
 
-	if (tv && !timespec64_valid(tv))
+	if (tv && !timespec64_valid_settod(tv))
 		return -EINVAL;
 
 	error = security_settime64(tv, tz);
kernel/time/timekeeping.c +3 -3
index f986e1918d129..f136c56c28057 100644
@@ -1221,7 +1221,7 @@ int do_settimeofday64(const struct timespec64 *ts)
 	unsigned long flags;
 	int ret = 0;
 
-	if (!timespec64_valid_strict(ts))
+	if (!timespec64_valid_settod(ts))
 		return -EINVAL;
 
 	raw_spin_lock_irqsave(&timekeeper_lock, flags);
@@ -1278,7 +1278,7 @@ static int timekeeping_inject_offset(const struct timespec64 *ts)
 	/* Make sure the proposed value is valid */
 	tmp = timespec64_add(tk_xtime(tk), *ts);
 	if (timespec64_compare(&tk->wall_to_monotonic, ts) > 0 ||
-	    !timespec64_valid_strict(&tmp)) {
+	    !timespec64_valid_settod(&tmp)) {
 		ret = -EINVAL;
 		goto error;
 	}
@@ -1527,7 +1527,7 @@ void __init timekeeping_init(void)
 	unsigned long flags;
 
 	read_persistent_wall_and_boot_offset(&wall_time, &boot_offset);
-	if (timespec64_valid_strict(&wall_time) &&
+	if (timespec64_valid_settod(&wall_time) &&
 	    timespec64_to_ns(&wall_time) > 0) {
 		persistent_clock_exists = true;
 	} else if (timespec64_to_ns(&wall_time) != 0) {
Signature · 1 line
2.20.1
author maintainer
#296 lore raw

PATCH 296/405 IB/mlx5: Compare only index part of a memory window rkey

In reply to Greg Kroah-Hartman

[ Upstream commit d623dfd2836114507d647c9793a80d213d8bffe8 ]

The InfiniBand Architecture Specification section 10.6.7.2.4 TYPE 2 MEMORY WINDOWS says that if the CI supports the Base Memory Management Extensions defined in this specification, the R_Key format for a Type 2 Memory Window must consist of:

* 24 bit index in the most significant bits of the R_Key, which is owned by the CI, and
* 8 bit key in the least significant bits of the R_Key, which is owned by the Consumer.

This means that the kernel should compare only the index part of a R_Key to determine equality with another R_Key.

Fixes db570d7deafb ("IB/mlx5: Add ODP support to MW")
---
 drivers/infiniband/hw/mlx5/odp.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/infiniband/hw/mlx5/odp.c +10 -1
index 0aa10ebda5d9a..91669e35c6ca8 100644
@@ -711,6 +711,15 @@ struct pf_frame {
 	int depth;
 };
 
+static bool mkey_is_eq(struct mlx5_core_mkey *mmkey, u32 key)
+{
+	if (!mmkey)
+		return false;
+	if (mmkey->type == MLX5_MKEY_MW)
+		return mlx5_base_mkey(mmkey->key) == mlx5_base_mkey(key);
+	return mmkey->key == key;
+}
+
 static int get_indirect_num_descs(struct mlx5_core_mkey *mmkey)
 {
 	struct mlx5_ib_mw *mw;
@@ -760,7 +769,7 @@ static int pagefault_single_data_segment(struct mlx5_ib_dev *dev,
 
 next_mr:
 	mmkey = __mlx5_mr_lookup(dev->mdev, mlx5_base_mkey(key));
-	if (!mmkey || mmkey->key != key) {
+	if (!mkey_is_eq(mmkey, key)) {
 		mlx5_ib_dbg(dev, "failed to find mkey %x\n", key);
 		ret = -EFAULT;
 		goto srcu_unlock;
Signature · 1 line
2.20.1
author maintainer
#297 lore raw

PATCH 297/405 scsi: qedf: Add missing return in qedf_post_io_req() in the fcport offload check

In reply to Greg Kroah-Hartman

[ Upstream commit c5e06ba2f76809ad1492fdad312e81335df46bc5 ]

Fixes the following crash as the return was missing from the check if an fcport is offloaded. If we hit this code we continue to try to post an invalid task which can lead to the crash:

[30259.616411] [0000:61:00.3]:[qedf_post_io_req:989]:3: Session not offloaded yet.
[30259.616413] [0000:61:00.3]:[qedf_upload_connection:1340]:3: Uploading connection port_id=490020.
[30259.623769] BUG: unable to handle kernel NULL pointer dereference at 0000000000000198
[30259.631645] IP: [<ffffffffc035b1ed>] qedf_init_task.isra.16+0x3d/0x450 [qedf]
[30259.638816] PGD 0
[30259.640841] Oops: 0000 [#1] SMP
[30259.644098] Modules linked in: fuse xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge stp llc ebtable_filter ebtables devlink ip6table_filter ip6_tables iptable_filter vfat fat ib_isert iscsi_target_mod ib_srpt target_core_mod ib_srp scsi_transport_srp ib_ipoib ib_ucm ib_umad dm_service_time skx_edac intel_powerclamp coretemp intel_rapl iosf_mbi kvm_intel kvm irqbypass crc32_pclmul ghash_clmulni_intel aesni_intel rpcrdma sunrpc rdma_ucm ib_uverbs lrw gf128mul ib_iser rdma_cm iw_cm ib_cm libiscsi scsi_transport_iscsi qedr(OE) glue_helper ablk_helper cryptd ib_core dm_round_robin joydev pcspkr ipmi_ssif ses enclosure ipmi_si ipmi_devintf ipmi_msghandler mei_me
[30259.715529]  mei sg hpilo hpwdt shpchp wmi lpc_ich acpi_power_meter dm_multipath ip_tables xfs libcrc32c sd_mod crc_t10dif crct10dif_generic uas usb_storage mgag200 qedf(OE) i2c_algo_bit libfcoe drm_kms_helper libfc syscopyarea sysfillrect scsi_transport_fc qede(OE) sysimgblt fb_sys_fops ptp ttm pps_core drm qed(OE) smartpqi crct10dif_pclmul crct10dif_common crc32c_intel i2c_core scsi_transport_sas scsi_tgt dm_mirror dm_region_hash dm_log dm_mod
[30259.754237] CPU: 9 PID: 977 Comm: kdmwork-253:7 Kdump: loaded Tainted: G        W  OE  ------------   3.10.0-862.el7.x86_64 #1
[30259.765664] Hardware name: HPE Synergy 480 Gen10/Synergy 480 Gen10 Compute Module, BIOS I42 04/04/2018
[30259.775000] task: ffff8c801efd0000 ti: ffff8c801efd8000 task.ti: ffff8c801efd8000
[30259.782505] RIP: 0010:[<ffffffffc035b1ed>]  [<ffffffffc035b1ed>] qedf_init_task.isra.16+0x3d/0x450 [qedf]
[30259.792116] RSP: 0018:ffff8c801efdbbb0  EFLAGS: 00010046
[30259.797444] RAX: 0000000000000000 RBX: ffffa7f1450948d8 RCX: ffff8c7fe5bc40c8
[30259.804600] RDX: ffff8c800715b300 RSI: ffffa7f1450948d8 RDI: ffff8c80169c2480
[30259.811755] RBP: ffff8c801efdbc30 R08: 00000000000000ae R09: ffff8c800a314540
[30259.818911] R10: ffff8c7fe5bc40c8 R11: ffff8c801efdb8ae R12: 0000000000000000
[30259.826068] R13: ffff8c800715b300 R14: ffff8c80169c2480 R15: ffff8c8005da28e0
[30259.833223] FS:  0000000000000000(0000) GS:ffff8c803f840000(0000) knlGS:0000000000000000
[30259.841338] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[30259.847100] CR2: 0000000000000198 CR3: 000000081242e000 CR4: 00000000007607e0
[30259.854256] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[30259.861412] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[30259.868568] PKRU: 00000000
[30259.871278] Call Trace:
[30259.873737]  [<ffffffffc035c948>] qedf_post_io_req+0x148/0x680 [qedf]
[30259.880201]  [<ffffffffc035d070>] qedf_queuecommand+0x1f0/0x240 [qedf]
[30259.886749]  [<ffffffffa329b050>] scsi_dispatch_cmd+0xb0/0x240
[30259.892600]  [<ffffffffa32a45bc>] scsi_request_fn+0x4cc/0x680
[30259.898364]  [<ffffffffa3118ad9>] __blk_run_queue+0x39/0x50
[30259.903954]  [<ffffffffa3114393>] __elv_add_request+0xd3/0x260
[30259.909805]  [<ffffffffa311baf0>] blk_insert_cloned_request+0xf0/0x1b0
[30259.916358]  [<ffffffffc010b622>] map_request+0x142/0x220 [dm_mod]
[30259.922560]  [<ffffffffc010b716>] map_tio_request+0x16/0x40 [dm_mod]
[30259.928932]  [<ffffffffa2ebb1f5>] kthread_worker_fn+0x85/0x180
[30259.934782]  [<ffffffffa2ebb170>] ? kthread_stop+0xf0/0xf0
[30259.940284]  [<ffffffffa2ebae31>] kthread+0xd1/0xe0
[30259.945176]  [<ffffffffa2ebad60>] ? insert_kthread_work+0x40/0x40
[30259.951290]  [<ffffffffa351f61d>] ret_from_fork_nospec_begin+0x7/0x21
[30259.957750]  [<ffffffffa2ebad60>] ? insert_kthread_work+0x40/0x40
[30259.963860] Code: fe 41 55 49 89 d5 41 54 53 48 89 f3 48 83 ec 58 4c 8b 67 28 4c 8b 4e 18 65 48 8b 04 25 28 00 00 00 48 89 45 d0 31 c0 4c 8b 7e 58 <49> 8b 84 24 98 01 00 00 48 8b 00 f6 80 31 01 00 00 10 0f 85 0b
[30259.983372] RIP  [<ffffffffc035b1ed>] qedf_init_task.isra.16+0x3d/0x450 [qedf]
[30259.990630]  RSP <ffff8c801efdbbb0>
[30259.994127] CR2: 0000000000000198
---
 drivers/scsi/qedf/qedf_io.c | 1 +
 1 file changed, 1 insertion(+)
drivers/scsi/qedf/qedf_io.c +1 -0
index 6ca583bdde23c..29b51c4667217 100644
@@ -902,6 +902,7 @@ int qedf_post_io_req(struct qedf_rport *fcport, struct qedf_ioreq *io_req)
 	if (!test_bit(QEDF_RPORT_SESSION_READY, &fcport->flags)) {
 		QEDF_ERR(&(qedf->dbg_ctx), "Session not offloaded yet.\n");
 		kref_put(&io_req->refcount, qedf_release_cmd);
+		return -EINVAL;
 	}
 
 	/* Obtain free SQE */
Signature · 1 line
2.20.1
author maintainer
#298 lore raw

PATCH 298/405 misc: fastrpc: consider address offset before sending to DSP

In reply to Greg Kroah-Hartman

[ Upstream commit 80f3afd72bd4149c57daf852905476b43bb47647 ]

While passing address phy address to DSP, take care of the offset calculated from virtual address vma.

Fixes c68cfb718c8f ("misc: fastrpc: Add support for context Invoke method")
---
 drivers/misc/fastrpc.c | 8 ++++++++
 1 file changed, 8 insertions(+)
drivers/misc/fastrpc.c +8 -0
index 36d0d5c9cfbad..9996c83ba5cb9 100644
@@ -667,8 +667,16 @@ static int fastrpc_get_args(u32 kernel, struct fastrpc_invoke_ctx *ctx)
 		pages[i].size = roundup(len, PAGE_SIZE);
 
 		if (ctx->maps[i]) {
+			struct vm_area_struct *vma = NULL;
+
 			rpra[i].pv = (u64) ctx->args[i].ptr;
 			pages[i].addr = ctx->maps[i]->phys;
+
+			vma = find_vma(current->mm, ctx->args[i].ptr);
+			if (vma)
+				pages[i].addr += ctx->args[i].ptr -
+						 vma->vm_start;
+
 		} else {
 			rlen -= ALIGN(args, FASTRPC_ALIGN) - args;
 			args = ALIGN(args, FASTRPC_ALIGN);
Signature · 1 line
2.20.1
author maintainer
#299 lore raw

PATCH 299/405 misc: fastrpc: make sure memory read and writes are visible

In reply to Greg Kroah-Hartman

[ Upstream commit 415a0729bd1225f0ffbc0ba82888dd65772554f7 ]

dma_alloc_coherent buffers could have writes queued in store buffers so commit them before sending buffer to DSP using correct dma barriers. Same with vice-versa.

Fixes c68cfb718c8f ("misc: fastrpc: Add support for context Invoke method")
---
 drivers/misc/fastrpc.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/misc/fastrpc.c +5 -0
index 9996c83ba5cb9..a10937652ca73 100644
@@ -790,6 +790,9 @@ static int fastrpc_internal_invoke(struct fastrpc_user *fl,  u32 kernel,
 		if (err)
 			goto bail;
 	}
+
+	/* make sure that all CPU memory writes are seen by DSP */
+	dma_wmb();
 	/* Send invoke buffer to remote dsp */
 	err = fastrpc_invoke_send(fl->sctx, ctx, kernel, handle);
 	if (err)
@@ -806,6 +809,8 @@ static int fastrpc_internal_invoke(struct fastrpc_user *fl,  u32 kernel,
 		goto bail;
 
 	if (ctx->nscalars) {
+		/* make sure that all memory writes by DSP are seen by CPU */
+		dma_rmb();
 		/* populate all the output buffers with results */
 		err = fastrpc_put_args(ctx, kernel);
 		if (err)
Signature · 1 line
2.20.1
author maintainer
#300 lore raw

PATCH 300/405 misc: fastrpc: Fix a possible double free

In reply to Greg Kroah-Hartman

[ Upstream commit b49f6d83e290f17e20f4e5cf31288d3bb4955ea6 ]

This patch fixes the error exit path of fastrpc_init_create_process(). If the DMA allocation or the DSP invoke fails the fastrpc_map was freed but not removed from the mapping list leading to a double free once the mapping list is emptied in fastrpc_device_release().

[srinivas kandagatla]: Cleaned up error path labels and reset init mem to NULL after free

Fixes d73f71c7c6ee("misc: fastrpc: Add support for create remote init process")
---
 drivers/misc/fastrpc.c | 31 ++++++++++++++++++++-----------
 1 file changed, 20 insertions(+), 11 deletions(-)
drivers/misc/fastrpc.c +20 -11
index a10937652ca73..35be1cc11dd85 100644
@@ -856,12 +856,12 @@ static int fastrpc_init_create_process(struct fastrpc_user *fl,
 
 	if (copy_from_user(&init, argp, sizeof(init))) {
 		err = -EFAULT;
-		goto bail;
+		goto err;
 	}
 
 	if (init.filelen > INIT_FILELEN_MAX) {
 		err = -EINVAL;
-		goto bail;
+		goto err;
 	}
 
 	inbuf.pgid = fl->tgid;
@@ -875,17 +875,15 @@ static int fastrpc_init_create_process(struct fastrpc_user *fl,
 	if (init.filelen && init.filefd) {
 		err = fastrpc_map_create(fl, init.filefd, init.filelen, &map);
 		if (err)
-			goto bail;
+			goto err;
 	}
 
 	memlen = ALIGN(max(INIT_FILELEN_MAX, (int)init.filelen * 4),
 		       1024 * 1024);
 	err = fastrpc_buf_alloc(fl, fl->sctx->dev, memlen,
 				&imem);
-	if (err) {
-		fastrpc_map_put(map);
-		goto bail;
-	}
+	if (err)
+		goto err_alloc;
 
 	fl->init_mem = imem;
 	args[0].ptr = (u64)(uintptr_t)&inbuf;
@@ -921,13 +919,24 @@ static int fastrpc_init_create_process(struct fastrpc_user *fl,
 
 	err = fastrpc_internal_invoke(fl, true, FASTRPC_INIT_HANDLE,
 				      sc, args);
+	if (err)
+		goto err_invoke;
 
-	if (err) {
+	kfree(args);
+
+	return 0;
+
+err_invoke:
+	fl->init_mem = NULL;
+	fastrpc_buf_free(imem);
+err_alloc:
+	if (map) {
+		spin_lock(&fl->lock);
+		list_del(&map->node);
+		spin_unlock(&fl->lock);
 		fastrpc_map_put(map);
-		fastrpc_buf_free(imem);
 	}
-
-bail:
+err:
 	kfree(args);
 
 	return err;
Signature · 1 line
2.20.1
author maintainer
#301 lore raw

PATCH 301/405 virtio_console: initialize vtermno value for ports

In reply to Greg Kroah-Hartman

[ Upstream commit 4b0a2c5ff7215206ea6135a405f17c5f6fca7d00 ]

For regular serial ports we do not initialize value of vtermno variable. A garbage value is assigned for non console ports. The value can be observed as a random integer with [1].

[1] vim /sys/kernel/debug/virtio-ports/vport*p*

This patch initialize the value of vtermno for console serial ports to '1' and regular serial ports are initiaized to '0'.

---
 drivers/char/virtio_console.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/char/virtio_console.c +2 -1
index fbeb71953526a..05dbfdb9f4aff 100644
@@ -75,7 +75,7 @@ struct ports_driver_data {
 	/* All the console devices handled by this driver */
 	struct list_head consoles;
 };
-static struct ports_driver_data pdrvdata;
+static struct ports_driver_data pdrvdata = { .next_vtermno = 1};
 
 static DEFINE_SPINLOCK(pdrvdata_lock);
 static DECLARE_COMPLETION(early_console_added);
@@ -1394,6 +1394,7 @@ static int add_port(struct ports_device *portdev, u32 id)
 	port->async_queue = NULL;
 
 	port->cons.ws.ws_row = port->cons.ws.ws_col = 0;
+	port->cons.vtermno = 0;
 
 	port->host_connected = port->guest_connected = false;
 	port->stats = (struct port_stats) { 0 };
Signature · 1 line
2.20.1
author maintainer
#302 lore raw

PATCH 302/405 tty: ipwireless: fix missing checks for ioremap

In reply to Greg Kroah-Hartman

[ Upstream commit 1bbb1c318cd8a3a39e8c3e2e83d5e90542d6c3e3 ]

ipw->attr_memory and ipw->common_memory are assigned with the return value of ioremap. ioremap may fail, but no checks are enforced. The fix inserts the checks to avoid potential NULL pointer dereferences.

---
 drivers/tty/ipwireless/main.c | 8 ++++++++
 1 file changed, 8 insertions(+)
drivers/tty/ipwireless/main.c +8 -0
index 3475e841ef5c1..4c18bbfe1a92e 100644
@@ -114,6 +114,10 @@ static int ipwireless_probe(struct pcmcia_device *p_dev, void *priv_data)
 
 	ipw->common_memory = ioremap(p_dev->resource[2]->start,
 				resource_size(p_dev->resource[2]));
+	if (!ipw->common_memory) {
+		ret = -ENOMEM;
+		goto exit1;
+	}
 	if (!request_mem_region(p_dev->resource[2]->start,
 				resource_size(p_dev->resource[2]),
 				IPWIRELESS_PCCARD_NAME)) {
@@ -134,6 +138,10 @@ static int ipwireless_probe(struct pcmcia_device *p_dev, void *priv_data)
 
 	ipw->attr_memory = ioremap(p_dev->resource[3]->start,
 				resource_size(p_dev->resource[3]));
+	if (!ipw->attr_memory) {
+		ret = -ENOMEM;
+		goto exit3;
+	}
 	if (!request_mem_region(p_dev->resource[3]->start,
 				resource_size(p_dev->resource[3]),
 				IPWIRELESS_PCCARD_NAME)) {
Signature · 1 line
2.20.1