4.19.168-rc1 review

This is the start of the stable review cycle for the 4.19.168 release. There are 43 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Sun, 17 Jan 2021 12:19:42 +0000. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

	https://www.kernel.org/pub/linux/kernel/v4.x/stable-review/patch-4.19.168-rc1.gz

or in the git tree and branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-4.19.y
and the diffstat can be found below.

thanks,

greg k-h

-------------
Pseudo-Shortlog of commits:

Greg Kroah-Hartman <[email protected]>

    Linux 4.19.168-rc1

Dan Carpenter <[email protected]>
regmap: debugfs: Fix a reversed if statement in regmap_debugfs_init()

Vasily Averin <[email protected]>
net: drop bogus skb with CHECKSUM_PARTIAL and offset beyond end of trimmed packet

Ming Lei <[email protected]>
block: fix use-after-free in disk_part_iter_next

Marc Zyngier <[email protected]>
KVM: arm64: Don't access PMCR_EL0 when no PMU is available

Arnd Bergmann <[email protected]>
wan: ds26522: select CONFIG_BITREVERSE

Xiaolei Wang <[email protected]>
regmap: debugfs: Fix a memory leak when calling regmap_attach_dev

Dinghao Liu <[email protected]>
net/mlx5e: Fix two double free cases

Dinghao Liu <[email protected]>
net/mlx5e: Fix memleak in mlx5e_create_l2_table_groups

Dinghao Liu <[email protected]>
iommu/intel: Fix memleak in intel_irq_remapping_alloc

Arnd Bergmann <[email protected]>
lightnvm: select CONFIG_CRC32

Arnd Bergmann <[email protected]>
block: rsxx: select CONFIG_CRC32

Arnd Bergmann <[email protected]>

    wil6210: select CONFIG_CRC32

Shravya Kumbham <[email protected]>
dmaengine: xilinx_dma: fix mixed_enum_type coverity warning

Shravya Kumbham <[email protected]>
dmaengine: xilinx_dma: fix incompatible param warning in _child_probe()

Shravya Kumbham <[email protected]>
dmaengine: xilinx_dma: check dma_async_device_register return value

Christophe JAILLET <[email protected]>
dmaengine: mediatek: mtk-hsdma: Fix a resource leak in the error handling path of the probe function

Roman Guskov <[email protected]>
spi: stm32: FIFO threshold level - fix align packet size

Colin Ian King <[email protected]>
cpufreq: powernow-k8: pass policy rather than use cpufreq_cpu_get()

Chunyan Zhang <[email protected]>
i2c: sprd: use a specific timeout to avoid system hang up issue

Andreas Kemnade <[email protected]>
ARM: OMAP2+: omap_device: fix idling of devices during probe

Ping Cheng <[email protected]>
HID: wacom: Fix memory leakage caused by kfifo_alloc

Lorenzo Bianconi <[email protected]>
iio: imu: st_lsm6dsx: fix edge-trigger interrupts

Sean Nyekjaer <[email protected]>
iio: imu: st_lsm6dsx: flip irq return logic

Lukas Wunner <[email protected]>
spi: pxa2xx: Fix use-after-free on unbind

Chris Wilson <[email protected]>
drm/i915: Fix mismatch between misplaced vma check and vma insert

Nick Desaulniers <[email protected]>
vmlinux.lds.h: Add PGO and AutoFDO input sections

Fenghua Yu <[email protected]>
x86/resctrl: Don't move a task to the same resource group

Fenghua Yu <[email protected]>
x86/resctrl: Use an IPI instead of task_work_add() to update PQR_ASSOC MSR

Ayush Sawal <[email protected]>
chtls: Fix chtls resources release sequence

Ayush Sawal <[email protected]>
chtls: Added a check to avoid NULL pointer dereference

Ayush Sawal <[email protected]>
chtls: Replace skb_dequeue with skb_peek

Ayush Sawal <[email protected]>
chtls: Fix panic when route to peer not configured

Ayush Sawal <[email protected]>
chtls: Remove invalid set_tcb call

Ayush Sawal <[email protected]>
chtls: Fix hardware tid leak

Sean Tranchetti <[email protected]>
net: ipv6: fib: flush exceptions when purging route

Florian Westphal <[email protected]>
net: fix pmtu check in nopmtudisc mode

Florian Westphal <[email protected]>
net: ip: always refragment ip defragmented packets

Christophe JAILLET <[email protected]>
net/sonic: Fix some resource leaks in error handling paths

Jakub Kicinski <[email protected]>
net: vlan: avoid leaks on register_vlan_dev() failures

Samuel Holland <[email protected]>
net: stmmac: dwmac-sun8i: Balance internal PHY power

Samuel Holland <[email protected]>
net: stmmac: dwmac-sun8i: Balance internal PHY resource references

Yufeng Mo <[email protected]>
net: hns3: fix the number of queues actually used by ARQ

Jouni K. Seppänen <[email protected]>
net: cdc_ncm: correct overhead in delayed_ndp_size

-------------

Diffstat:

 Makefile                                          |   4 +-
 arch/arm/mach-omap2/omap_device.c                 |   8 +-
 arch/arm64/kvm/sys_regs.c                         |   4 +
 arch/x86/kernel/cpu/intel_rdt_rdtgroup.c          | 113 ++++++++++------------
 block/genhd.c                                     |   9 +-
 drivers/base/regmap/regmap-debugfs.c              |   9 +-
 drivers/block/Kconfig                             |   1 +
 drivers/cpufreq/powernow-k8.c                     |   9 +-
 drivers/crypto/chelsio/chtls/chtls_cm.c           |  68 +++++--------
 drivers/dma/mediatek/mtk-hsdma.c                  |   1 +
 drivers/dma/xilinx/xilinx_dma.c                   |  11 ++-
 drivers/gpu/drm/i915/i915_gem_execbuffer.c        |   2 +-
 drivers/hid/wacom_sys.c                           |  35 ++++++-
 drivers/i2c/busses/i2c-sprd.c                     |   8 +-
 drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c    |  26 ++++-
 drivers/iommu/intel_irq_remapping.c               |   2 +
 drivers/lightnvm/Kconfig                          |   1 +
 drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h   |   4 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_fs.c   |   3 +
 drivers/net/ethernet/natsemi/macsonic.c           |  12 ++-
 drivers/net/ethernet/natsemi/xtsonic.c            |   7 +-
 drivers/net/ethernet/stmicro/stmmac/dwmac-sun8i.c |  58 ++++++++---
 drivers/net/usb/cdc_ncm.c                         |   8 +-
 drivers/net/wan/Kconfig                           |   1 +
 drivers/net/wireless/ath/wil6210/Kconfig          |   1 +
 drivers/spi/spi-pxa2xx.c                          |   3 +-
 drivers/spi/spi-stm32.c                           |   4 +-
 include/asm-generic/vmlinux.lds.h                 |   5 +-
 net/8021q/vlan.c                                  |   3 +-
 net/core/skbuff.c                                 |   6 ++
 net/ipv4/ip_output.c                              |   2 +-
 net/ipv4/ip_tunnel.c                              |  10 +-
 net/ipv6/ip6_fib.c                                |   5 +-
 33 files changed, 268 insertions(+), 175 deletions(-)

From: "Jouni K. Sepp�nen" <[email protected]>

[ Upstream commit 7a68d725e4ea384977445e0bcaed3d7de83ab5b3 ]

Aligning to tx_ndp_modulus is not sufficient because the next align call can be cdc_ncm_align_tail, which can add up to ctx->tx_modulus + ctx->tx_remainder - 1 bytes. This used to lead to occasional crashes on a Huawei 909s-120 LTE module as follows:

- the condition marked /* if there is a remaining skb [...] */ is true so the swaps happen
- skb_out is set from ctx->tx_curr_skb
- skb_out->len is exactly 0x3f52
- ctx->tx_curr_size is 0x4000 and delayed_ndp_size is 0xac (note that the sum of skb_out->len and delayed_ndp_size is 0x3ffe)
- the for loop over n is executed once
- the cdc_ncm_align_tail call marked /* align beginning of next frame */ increases skb_out->len to 0x3f56 (the sum is now 0x4002)
- the condition marked /* check if we had enough room left [...] */ is false so we break out of the loop
- the condition marked /* If requested, put NDP at end of frame. */ is true so the NDP is written into skb_out
- now skb_out->len is 0x4002, so padding_count is minus two interpreted as an unsigned number, which is used as the length argument to memset, leading to a crash with various symptoms but usually including

The cdc_ncm_align_tail call first aligns on a ctx->tx_modulus boundary (adding at most ctx->tx_modulus-1 bytes), then adds ctx->tx_remainder bytes. Alternatively, the next alignment call can occur in cdc_ncm_ndp16 or cdc_ncm_ndp32, in which case at most ctx->tx_ndp_modulus-1 bytes are added.

A similar problem has occurred before, and the code is nontrivial to reason about, so add a guard before the crashing call. By that time it is too late to prevent any memory corruption (we'll have written past the end of the buffer already) but we can at least try to get a warning written into an on-disk log by avoiding the hard crash caused by padding past the buffer with a huge number of zeros.

---
 drivers/net/usb/cdc_ncm.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

@@ -1127,7 +1127,10 @@ cdc_ncm_fill_tx_frame(struct usbnet *dev
 	 * accordingly. Otherwise, we should check here.
 	 */
 	if (ctx->drvflags & CDC_NCM_FLAG_NDP_TO_END)
-		delayed_ndp_size = ALIGN(ctx->max_ndp_size, ctx->tx_ndp_modulus);
+		delayed_ndp_size = ctx->max_ndp_size +
+			max_t(u32,
+			      ctx->tx_ndp_modulus,
+			      ctx->tx_modulus + ctx->tx_remainder) - 1;
 	else
 		delayed_ndp_size = 0;
 

@@ -1308,7 +1311,8 @@ cdc_ncm_fill_tx_frame(struct usbnet *dev
 	if (!(dev->driver_info->flags & FLAG_SEND_ZLP) &&
 	    skb_out->len > ctx->min_tx_pkt) {
 		padding_count = ctx->tx_curr_size - skb_out->len;
-		skb_put_zero(skb_out, padding_count);
+		if (!WARN_ON(padding_count > ctx->tx_curr_size))
+			skb_put_zero(skb_out, padding_count);
 	} else if (skb_out->len < ctx->tx_curr_size &&
 		   (skb_out->len % dev->maxpacket) == 0) {
 		skb_put_u8(skb_out, 0);	/* force short packet */

From: Yufeng Mo <[email protected]>

[ Upstream commit 65e61e3c2a619c4d4b873885b2d5394025ed117b ]

HCLGE_MBX_MAX_ARQ_MSG_NUM is used to apply memory for the number of queues used by ARQ(Asynchronous Receive Queue), so the head and tail pointers should also use this macro.

---
 drivers/net/ethernet/hisilicon/hns3/hclge_mbx.h |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -102,7 +102,7 @@ struct hclgevf_mbx_arq_ring {
 #define hclge_mbx_ring_ptr_move_crq(crq) \
 	(crq->next_to_use = (crq->next_to_use + 1) % crq->desc_num)
 #define hclge_mbx_tail_ptr_move_arq(arq) \
-	(arq.tail = (arq.tail + 1) % HCLGE_MBX_MAX_ARQ_MSG_SIZE)
+		(arq.tail = (arq.tail + 1) % HCLGE_MBX_MAX_ARQ_MSG_NUM)
 #define hclge_mbx_head_ptr_move_arq(arq) \
-		(arq.head = (arq.head + 1) % HCLGE_MBX_MAX_ARQ_MSG_SIZE)
+		(arq.head = (arq.head + 1) % HCLGE_MBX_MAX_ARQ_MSG_NUM)
 #endif

From: Samuel Holland <[email protected]>

[ Upstream commit 529254216773acd5039c07aa18cf06fd1f9fccdd ]

While stmmac_pltfr_remove calls sun8i_dwmac_exit, the sun8i_dwmac_init and sun8i_dwmac_exit functions are also called by the stmmac_platform suspend/resume callbacks. They may be called many times during the device's lifetime and should not release resources used by the driver.

Furthermore, there was no error handling in case registering the MDIO mux failed during probe, and the EPHY clock was never released at all.

Fix all of these issues by moving the deinitialization code to a driver removal callback. Also ensure the EPHY is powered down before removal.

---
 drivers/net/ethernet/stmicro/stmmac/dwmac-sun8i.c |   27 +++++++++++++++++-----
 1 file changed, 21 insertions(+), 6 deletions(-)

@@ -977,17 +977,12 @@ static void sun8i_dwmac_exit(struct plat
 	struct sunxi_priv_data *gmac = priv;
 
 	if (gmac->variant->soc_has_internal_phy) {
-		/* sun8i_dwmac_exit could be called with mdiomux uninit */
-		if (gmac->mux_handle)
-			mdio_mux_uninit(gmac->mux_handle);
 		if (gmac->internal_phy_powered)
 			sun8i_dwmac_unpower_internal_phy(gmac);
 	}
 
 	sun8i_dwmac_unset_syscon(gmac);
 
-	reset_control_put(gmac->rst_ephy);
-
 	clk_disable_unprepare(gmac->tx_clk);
 
 	if (gmac->regulator)

@@ -1200,12 +1195,32 @@ static int sun8i_dwmac_probe(struct plat
 
 	return ret;
 dwmac_mux:
+	reset_control_put(gmac->rst_ephy);
+	clk_put(gmac->ephy_clk);
 	sun8i_dwmac_unset_syscon(gmac);
 dwmac_exit:
 	stmmac_pltfr_remove(pdev);
 return ret;
 }
 
+static int sun8i_dwmac_remove(struct platform_device *pdev)
+{
+	struct net_device *ndev = platform_get_drvdata(pdev);
+	struct stmmac_priv *priv = netdev_priv(ndev);
+	struct sunxi_priv_data *gmac = priv->plat->bsp_priv;
+
+	if (gmac->variant->soc_has_internal_phy) {
+		mdio_mux_uninit(gmac->mux_handle);
+		sun8i_dwmac_unpower_internal_phy(gmac);
+		reset_control_put(gmac->rst_ephy);
+		clk_put(gmac->ephy_clk);
+	}
+
+	stmmac_pltfr_remove(pdev);
+
+	return 0;
+}
+
 static const struct of_device_id sun8i_dwmac_match[] = {
 	{ .compatible = "allwinner,sun8i-h3-emac",
 		.data = &emac_variant_h3 },

@@ -1223,7 +1238,7 @@ MODULE_DEVICE_TABLE(of, sun8i_dwmac_matc
 
 static struct platform_driver sun8i_dwmac_driver = {
 	.probe  = sun8i_dwmac_probe,
-	.remove = stmmac_pltfr_remove,
+	.remove = sun8i_dwmac_remove,
 	.driver = {
 		.name           = "dwmac-sun8i",
 		.pm		= &stmmac_pltfr_pm_ops,

From: Samuel Holland <[email protected]>

[ Upstream commit b8239638853e3e37b287e4bd4d57b41f14c78550 ]

sun8i_dwmac_exit calls sun8i_dwmac_unpower_internal_phy, but sun8i_dwmac_init did not call sun8i_dwmac_power_internal_phy. This caused PHY power to remain off after a suspend/resume cycle. Fix this by recording if PHY power should be restored, and if so, restoring it.

---
 drivers/net/ethernet/stmicro/stmmac/dwmac-sun8i.c |   31 ++++++++++++++++------
 1 file changed, 23 insertions(+), 8 deletions(-)

@@ -73,6 +73,7 @@ struct emac_variant {
  * @variant:	reference to the current board variant
  * @regmap:	regmap for using the syscon
  * @internal_phy_powered: Does the internal PHY is enabled
+ * @use_internal_phy: Is the internal PHY selected for use
  * @mux_handle:	Internal pointer used by mdio-mux lib
  */
 struct sunxi_priv_data {

@@ -83,6 +84,7 @@ struct sunxi_priv_data {
 	const struct emac_variant *variant;
 	struct regmap_field *regmap_field;
 	bool internal_phy_powered;
+	bool use_internal_phy;
 	void *mux_handle;
 };
 

@@ -518,8 +520,11 @@ static const struct stmmac_dma_ops sun8i
 	.dma_interrupt = sun8i_dwmac_dma_interrupt,
 };
 
+static int sun8i_dwmac_power_internal_phy(struct stmmac_priv *priv);
+
 static int sun8i_dwmac_init(struct platform_device *pdev, void *priv)
 {
+	struct net_device *ndev = platform_get_drvdata(pdev);
 	struct sunxi_priv_data *gmac = priv;
 	int ret;
 

@@ -533,13 +538,25 @@ static int sun8i_dwmac_init(struct platf
 
 	ret = clk_prepare_enable(gmac->tx_clk);
 	if (ret) {
-		if (gmac->regulator)
-			regulator_disable(gmac->regulator);
 		dev_err(&pdev->dev, "Could not enable AHB clock\n");
-		return ret;
+		goto err_disable_regulator;
+	}
+
+	if (gmac->use_internal_phy) {
+		ret = sun8i_dwmac_power_internal_phy(netdev_priv(ndev));
+		if (ret)
+			goto err_disable_clk;
 	}
 
 	return 0;
+
+err_disable_clk:
+	clk_disable_unprepare(gmac->tx_clk);
+err_disable_regulator:
+	if (gmac->regulator)
+		regulator_disable(gmac->regulator);
+
+	return ret;
 }
 
 static void sun8i_dwmac_core_init(struct mac_device_info *hw,

@@ -809,7 +826,6 @@ static int mdio_mux_syscon_switch_fn(int
 	struct sunxi_priv_data *gmac = priv->plat->bsp_priv;
 	u32 reg, val;
 	int ret = 0;
-	bool need_power_ephy = false;
 
 	if (current_child ^ desired_child) {
 		regmap_field_read(gmac->regmap_field, &reg);

@@ -817,13 +833,12 @@ static int mdio_mux_syscon_switch_fn(int
 		case DWMAC_SUN8I_MDIO_MUX_INTERNAL_ID:
 			dev_info(priv->device, "Switch mux to internal PHY");
 			val = (reg & ~H3_EPHY_MUX_MASK) | H3_EPHY_SELECT;
-
-			need_power_ephy = true;
+			gmac->use_internal_phy = true;
 			break;
 		case DWMAC_SUN8I_MDIO_MUX_EXTERNAL_ID:
 			dev_info(priv->device, "Switch mux to external PHY");
 			val = (reg & ~H3_EPHY_MUX_MASK) | H3_EPHY_SHUTDOWN;
-			need_power_ephy = false;
+			gmac->use_internal_phy = false;
 			break;
 		default:
 			dev_err(priv->device, "Invalid child ID %x\n",

@@ -831,7 +846,7 @@ static int mdio_mux_syscon_switch_fn(int
 			return -EINVAL;
 		}
 		regmap_field_write(gmac->regmap_field, val);
-		if (need_power_ephy) {
+		if (gmac->use_internal_phy) {
 			ret = sun8i_dwmac_power_internal_phy(priv);
 			if (ret)
 				return ret;

From: Jakub Kicinski <[email protected]>

[ Upstream commit 55b7ab1178cbf41f979ff83236d3321ad35ed2ad ]

VLAN checks for NETREG_UNINITIALIZED to distinguish between registration failure and unregistration in progress.

Since commit cb626bf566eb ("net-sysfs: Fix reference count leak") registration failure may, however, result in NETREG_UNREGISTERED as well as NETREG_UNINITIALIZED.

This fix is similer to cebb69754f37 ("rtnetlink: Fix memory(net_device) leak when ->newlink fails")

---
 net/8021q/vlan.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -277,7 +277,8 @@ static int register_vlan_device(struct n
 	return 0;
 
 out_free_newdev:
-	if (new_dev->reg_state == NETREG_UNINITIALIZED)
+	if (new_dev->reg_state == NETREG_UNINITIALIZED ||
+	    new_dev->reg_state == NETREG_UNREGISTERED)
 		free_netdev(new_dev);
 	return err;
 }

From: Christophe JAILLET <[email protected]>

[ Upstream commit 0f7ba7bc46fa0b574ccacf5672991b321e028492 ]

A call to dma_alloc_coherent() is wrapped by sonic_alloc_descriptors().

This is correctly freed in the remove function, but not in the error handling path of the probe function. Fix this by adding the missing dma_free_coherent() call.

While at it, rename a label in order to be slightly more informative.

---
 drivers/net/ethernet/natsemi/macsonic.c |   12 ++++++++++--
 drivers/net/ethernet/natsemi/xtsonic.c  |    7 +++++--
 2 files changed, 15 insertions(+), 4 deletions(-)

@@ -540,10 +540,14 @@ static int mac_sonic_platform_probe(stru
 
 	err = register_netdev(dev);
 	if (err)
-		goto out;
+		goto undo_probe;
 
 	return 0;
 
+undo_probe:
+	dma_free_coherent(lp->device,
+			  SIZEOF_SONIC_DESC * SONIC_BUS_SCALE(lp->dma_bitmode),
+			  lp->descriptors, lp->descriptors_laddr);
 out:
 	free_netdev(dev);
 

@@ -618,12 +622,16 @@ static int mac_sonic_nubus_probe(struct
 
 	err = register_netdev(ndev);
 	if (err)
-		goto out;
+		goto undo_probe;
 
 	nubus_set_drvdata(board, ndev);
 
 	return 0;
 
+undo_probe:
+	dma_free_coherent(lp->device,
+			  SIZEOF_SONIC_DESC * SONIC_BUS_SCALE(lp->dma_bitmode),
+			  lp->descriptors, lp->descriptors_laddr);
 out:
 	free_netdev(ndev);
 	return err;

@@ -265,11 +265,14 @@ int xtsonic_probe(struct platform_device
 	sonic_msg_init(dev);
 
 	if ((err = register_netdev(dev)))
-		goto out1;
+		goto undo_probe1;
 
 	return 0;
 
-out1:
+undo_probe1:
+	dma_free_coherent(lp->device,
+			  SIZEOF_SONIC_DESC * SONIC_BUS_SCALE(lp->dma_bitmode),
+			  lp->descriptors, lp->descriptors_laddr);
 	release_region(dev->base_addr, SONIC_MEM_SIZE);
 out:
 	free_netdev(dev);

From: Florian Westphal <[email protected]>

[ Upstream commit bb4cc1a18856a73f0ff5137df0c2a31f4c50f6cf ]

Conntrack reassembly records the largest fragment size seen in IPCB. However, when this gets forwarded/transmitted, fragmentation will only be forced if one of the fragmented packets had the DF bit set.

In that case, a flag in IPCB will force fragmentation even if the MTU is large enough.

This should work fine, but this breaks with ip tunnels. Consider client that sends a UDP datagram of size X to another host.

The client fragments the datagram, so two packets, of size y and z, are sent. DF bit is not set on any of these packets.

Middlebox netfilter reassembles those packets back to single size-X packet, before routing decision.

packet-size-vs-mtu checks in ip_forward are irrelevant, because DF bit isn't set. At output time, ip refragmentation is skipped as well because x is still smaller than the mtu of the output device.

If ttransmit device is an ip tunnel, the packet size increases to x+overhead.

Also, tunnel might be configured to force DF bit on outer header.

In this case, packet will be dropped (exceeds MTU) and an ICMP error is generated back to sender.

But sender already respects the announced MTU, all the packets that it sent did fit the announced mtu.

Force refragmentation as per original sizes unconditionally so ip tunnel will encapsulate the fragments instead.

The only other solution I see is to place ip refragmentation in the ip_tunnel code to handle this case.

---
 net/ipv4/ip_output.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -312,7 +312,7 @@ static int ip_finish_output(struct net *
 	if (skb_is_gso(skb))
 		return ip_finish_output_gso(net, sk, skb, mtu);
 
-	if (skb->len > mtu || (IPCB(skb)->flags & IPSKB_FRAG_PMTU))
+	if (skb->len > mtu || IPCB(skb)->frag_max_size)
 		return ip_fragment(net, sk, skb, mtu, ip_finish_output2);
 
 	return ip_finish_output2(net, sk, skb);

From: Florian Westphal <[email protected]>

[ Upstream commit 50c661670f6a3908c273503dfa206dfc7aa54c07 ]

For some reason ip_tunnel insist on setting the DF bit anyway when the inner header has the DF bit set, EVEN if the tunnel was configured with 'nopmtudisc'.

This means that the script added in the previous commit cannot be made to work by adding the 'nopmtudisc' flag to the ip tunnel configuration. Doing so breaks connectivity even for the without-conntrack/netfilter scenario.

When nopmtudisc is set, the tunnel will skip the mtu check, so no icmp error is sent to client. Then, because inner header has DF set, the outer header gets added with DF bit set as well.

IP stack then sends an error to itself because the packet exceeds the device MTU.

---
 net/ipv4/ip_tunnel.c |   10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

@@ -736,7 +736,11 @@ void ip_tunnel_xmit(struct sk_buff *skb,
 		goto tx_error;
 	}
 
-	if (tnl_update_pmtu(dev, skb, rt, tnl_params->frag_off, inner_iph)) {
+	df = tnl_params->frag_off;
+	if (skb->protocol == htons(ETH_P_IP) && !tunnel->ignore_df)
+		df |= (inner_iph->frag_off & htons(IP_DF));
+
+	if (tnl_update_pmtu(dev, skb, rt, df, inner_iph)) {
 		ip_rt_put(rt);
 		goto tx_error;
 	}

@@ -764,10 +768,6 @@ void ip_tunnel_xmit(struct sk_buff *skb,
 			ttl = ip4_dst_hoplimit(&rt->dst);
 	}
 
-	df = tnl_params->frag_off;
-	if (skb->protocol == htons(ETH_P_IP) && !tunnel->ignore_df)
-		df |= (inner_iph->frag_off&htons(IP_DF));
-
 	max_headroom = LL_RESERVED_SPACE(rt->dst.dev) + sizeof(struct iphdr)
 			+ rt->dst.header_len + ip_encap_hlen(&tunnel->encap);
 	if (max_headroom > dev->needed_headroom)

From: Sean Tranchetti <[email protected]>

[ Upstream commit d8f5c29653c3f6995e8979be5623d263e92f6b86 ]

Route removal is handled by two code paths. The main removal path is via fib6_del_route() which will handle purging any PMTU exceptions from the cache, removing all per-cpu copies of the DST entry used by the route, and releasing the fib6_info struct.

The second removal location is during fib6_add_rt2node() during a route replacement operation. This path also calls fib6_purge_rt() to handle cleaning up the per-cpu copies of the DST entries and releasing the fib6_info associated with the older route, but it does not flush any PMTU exceptions that the older route had. Since the older route is removed from the tree during the replacement, we lose any way of accessing it again.

As these lingering DSTs and the fib6_info struct are holding references to the underlying netdevice struct as well, unregistering that device from the kernel can never complete.

---
 net/ipv6/ip6_fib.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

@@ -906,6 +906,8 @@ static void fib6_purge_rt(struct fib6_in
 {
 	struct fib6_table *table = rt->fib6_table;
 
+	/* Flush all cached dst in exception table */
+	rt6_flush_exceptions(rt);
 	if (rt->rt6i_pcpu)
 		fib6_drop_pcpu_from(rt, table);
 

@@ -1756,9 +1758,6 @@ static void fib6_del_route(struct fib6_t
 	net->ipv6.rt6_stats->fib_rt_entries--;
 	net->ipv6.rt6_stats->fib_discarded_routes++;
 
-	/* Flush all cached dst in exception table */
-	rt6_flush_exceptions(rt);
-
 	/* Reset round-robin state, if necessary */
 	if (rcu_access_pointer(fn->rr_ptr) == rt)
 		fn->rr_ptr = NULL;

From: Ayush Sawal <[email protected]>

[ Upstream commit 717df0f4cdc9044c415431a3522b3e9ccca5b4a3 ]

send_abort_rpl() is not calculating cpl_abort_req_rss offset and ends up sending wrong TID with abort_rpl WR causng tid leaks. Replaced send_abort_rpl() with chtls_send_abort_rpl() as it is redundant.

---
 drivers/crypto/chelsio/chtls/chtls_cm.c |   39 ++------------------------------
 1 file changed, 3 insertions(+), 36 deletions(-)

@@ -1807,39 +1807,6 @@ static void send_defer_abort_rpl(struct
 	kfree_skb(skb);
 }
 
-static void send_abort_rpl(struct sock *sk, struct sk_buff *skb,
-			   struct chtls_dev *cdev, int status, int queue)
-{
-	struct cpl_abort_req_rss *req = cplhdr(skb);
-	struct sk_buff *reply_skb;
-	struct chtls_sock *csk;
-
-	csk = rcu_dereference_sk_user_data(sk);
-
-	reply_skb = alloc_skb(sizeof(struct cpl_abort_rpl),
-			      GFP_KERNEL);
-
-	if (!reply_skb) {
-		req->status = (queue << 1);
-		send_defer_abort_rpl(cdev, skb);
-		return;
-	}
-
-	set_abort_rpl_wr(reply_skb, GET_TID(req), status);
-	kfree_skb(skb);
-
-	set_wr_txq(reply_skb, CPL_PRIORITY_DATA, queue);
-	if (csk_conn_inline(csk)) {
-		struct l2t_entry *e = csk->l2t_entry;
-
-		if (e && sk->sk_state != TCP_SYN_RECV) {
-			cxgb4_l2t_send(csk->egress_dev, reply_skb, e);
-			return;
-		}
-	}
-	cxgb4_ofld_send(cdev->lldi->ports[0], reply_skb);
-}
-
 /*
  * Add an skb to the deferred skb queue for processing from process context.
  */

@@ -1903,8 +1870,8 @@ static void bl_abort_syn_rcv(struct sock
 
 	skb->sk	= NULL;
 	do_abort_syn_rcv(child, lsk);
-	send_abort_rpl(child, skb, BLOG_SKB_CB(skb)->cdev,
-		       CPL_ABORT_NO_RST, queue);
+	chtls_send_abort_rpl(child, skb, BLOG_SKB_CB(skb)->cdev,
+			     CPL_ABORT_NO_RST, queue);
 }
 
 static int abort_syn_rcv(struct sock *sk, struct sk_buff *skb)

@@ -1935,7 +1902,7 @@ static int abort_syn_rcv(struct sock *sk
 		int queue = csk->txq_idx;
 
 		do_abort_syn_rcv(sk, psk);
-		send_abort_rpl(sk, skb, cdev, CPL_ABORT_NO_RST, queue);
+		chtls_send_abort_rpl(sk, skb, cdev, CPL_ABORT_NO_RST, queue);
 	} else {
 		skb->sk = sk;
 		BLOG_SKB_CB(skb)->backlog_rcv = bl_abort_syn_rcv;

From: Ayush Sawal <[email protected]>

[ Upstream commit 827d329105bfde6701f0077e34a09c4a86e27145 ]

At the time of SYN_RECV, connection information is not initialized at FW, updating tcb flag over uninitialized connection causes adapter crash. We don't need to update the flag during SYN_RECV state, so avoid this.

---
 drivers/crypto/chelsio/chtls/chtls_cm.c |    3 ---
 1 file changed, 3 deletions(-)

@@ -1920,9 +1920,6 @@ static void chtls_abort_req_rss(struct s
 	int queue = csk->txq_idx;
 
 	if (is_neg_adv(req->status)) {
-		if (sk->sk_state == TCP_SYN_RECV)
-			chtls_set_tcb_tflag(sk, 0, 0);
-
 		kfree_skb(skb);
 		return;
 	}

From: Ayush Sawal <[email protected]>

[ Upstream commit 5a5fac9966bb6d513198634b0b1357be7e8447d2 ]

If route to peer is not configured, we might get non tls devices from dst_neigh_lookup() which is invalid, adding a check to avoid it.

---
 drivers/crypto/chelsio/chtls/chtls_cm.c |   14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

@@ -1024,6 +1024,7 @@ static struct sock *chtls_recv_sock(stru
 				    const struct cpl_pass_accept_req *req,
 				    struct chtls_dev *cdev)
 {
+	struct adapter *adap = pci_get_drvdata(cdev->pdev);
 	const struct tcphdr *tcph;
 	struct inet_sock *newinet;
 	const struct iphdr *iph;

@@ -1033,9 +1034,10 @@ static struct sock *chtls_recv_sock(stru
 	struct neighbour *n;
 	struct tcp_sock *tp;
 	struct sock *newsk;
+	bool found = false;
 	u16 port_id;
 	int rxq_idx;
-	int step;
+	int step, i;
 
 	iph = (const struct iphdr *)network_hdr;
 	newsk = tcp_create_openreq_child(lsk, oreq, cdev->askb);

@@ -1048,7 +1050,7 @@ static struct sock *chtls_recv_sock(stru
 
 	tcph = (struct tcphdr *)(iph + 1);
 	n = dst_neigh_lookup(dst, &iph->saddr);
-	if (!n)
+	if (!n || !n->dev)
 		goto free_sk;
 
 	ndev = n->dev;

@@ -1057,6 +1059,13 @@ static struct sock *chtls_recv_sock(stru
 	if (is_vlan_dev(ndev))
 		ndev = vlan_dev_real_dev(ndev);
 
+	for_each_port(adap, i)
+		if (cdev->ports[i] == ndev)
+			found = true;
+
+	if (!found)
+		goto free_dst;
+
 	port_id = cxgb4_port_idx(ndev);
 
 	csk = chtls_sock_create(cdev);

@@ -1108,6 +1117,7 @@ static struct sock *chtls_recv_sock(stru
 free_csk:
 	chtls_sock_release(&csk->kref);
 free_dst:
+	neigh_release(n);
 	dst_release(dst);
 free_sk:
 	inet_csk_prepare_forced_close(newsk);

From: Ayush Sawal <[email protected]>

[ Upstream commit a84b2c0d5fa23da6d6c8c0d5f5c93184a2744d3e ]

The skb is unlinked twice, one in __skb_dequeue in function chtls_reset_synq() and another in cleanup_syn_rcv_conn(). So in this patch using skb_peek() instead of __skb_dequeue(), so that unlink will be handled only in cleanup_syn_rcv_conn().

---
 drivers/crypto/chelsio/chtls/chtls_cm.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -581,7 +581,7 @@ static void chtls_reset_synq(struct list
 
 	while (!skb_queue_empty(&listen_ctx->synq)) {
 		struct chtls_sock *csk =
-			container_of((struct synq *)__skb_dequeue
+			container_of((struct synq *)skb_peek
 				(&listen_ctx->synq), struct chtls_sock, synq);
 		struct sock *child = csk->sk;
 

From: Ayush Sawal <[email protected]>

[ Upstream commit eade1e0a4fb31d48eeb1589d9bb859ae4dd6181d ]

In case of server removal lookup_stid() may return NULL pointer, which is used as listen_ctx. So added a check before accessing this pointer.

---
 drivers/crypto/chelsio/chtls/chtls_cm.c |    5 +++++
 1 file changed, 5 insertions(+)

@@ -1432,6 +1432,11 @@ static int chtls_pass_establish(struct c
 			sk_wake_async(sk, 0, POLL_OUT);
 
 		data = lookup_stid(cdev->tids, stid);
+		if (!data) {
+			/* listening server close */
+			kfree_skb(skb);
+			goto unlock;
+		}
 		lsk = ((struct listen_ctx *)data)->lsk;
 
 		bh_lock_sock(lsk);

From: Ayush Sawal <[email protected]>

[ Upstream commit 15ef6b0e30b354253e2c10b3836bc59767eb162b ]

CPL_ABORT_RPL is sent after releasing the resources by calling chtls_release_resources(sk); and chtls_conn_done(sk); eventually causing kernel panic. Fixing it by calling release in appropriate order.

---
 drivers/crypto/chelsio/chtls/chtls_cm.c |    9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

@@ -1884,9 +1884,9 @@ static void bl_abort_syn_rcv(struct sock
 	queue = csk->txq_idx;
 
 	skb->sk	= NULL;
-	do_abort_syn_rcv(child, lsk);
 	chtls_send_abort_rpl(child, skb, BLOG_SKB_CB(skb)->cdev,
 			     CPL_ABORT_NO_RST, queue);
+	do_abort_syn_rcv(child, lsk);
 }
 
 static int abort_syn_rcv(struct sock *sk, struct sk_buff *skb)

@@ -1916,8 +1916,8 @@ static int abort_syn_rcv(struct sock *sk
 	if (!sock_owned_by_user(psk)) {
 		int queue = csk->txq_idx;
 
-		do_abort_syn_rcv(sk, psk);
 		chtls_send_abort_rpl(sk, skb, cdev, CPL_ABORT_NO_RST, queue);
+		do_abort_syn_rcv(sk, psk);
 	} else {
 		skb->sk = sk;
 		BLOG_SKB_CB(skb)->backlog_rcv = bl_abort_syn_rcv;

@@ -1960,12 +1960,11 @@ static void chtls_abort_req_rss(struct s
 
 		if (sk->sk_state == TCP_SYN_RECV && !abort_syn_rcv(sk, skb))
 			return;
-
-		chtls_release_resources(sk);
-		chtls_conn_done(sk);
 	}
 
 	chtls_send_abort_rpl(sk, skb, csk->cdev, rst_status, queue);
+	chtls_release_resources(sk);
+	chtls_conn_done(sk);
 }
 
 static void chtls_abort_rpl_rss(struct sock *sk, struct sk_buff *skb)

From: Fenghua Yu <[email protected]>

commit ae28d1aae48a1258bd09a6f707ebb4231d79a761 upstream

Currently, when moving a task to a resource group the PQR_ASSOC MSR is updated with the new closid and rmid in an added task callback. If the task is running, the work is run as soon as possible. If the task is not running, the work is executed later in the kernel exit path when the kernel returns to the task again.

Updating the PQR_ASSOC MSR as soon as possible on the CPU a moved task is running is the right thing to do. Queueing work for a task that is not running is unnecessary (the PQR_ASSOC MSR is already updated when the task is scheduled in) and causing system resource waste with the way in which it is implemented: Work to update the PQR_ASSOC register is queued every time the user writes a task id to the "tasks" file, even if the task already belongs to the resource group.

This could result in multiple pending work items associated with a single task even if they are all identical and even though only a single update with most recent values is needed. Specifically, even if a task is moved between different resource groups while it is sleeping then it is only the last move that is relevant but yet a work item is queued during each move.

This unnecessary queueing of work items could result in significant system resource waste, especially on tasks sleeping for a long time. For example, as demonstrated by Shakeel Butt in [1] writing the same task id to the "tasks" file can quickly consume significant memory. The same problem (wasted system resources) occurs when moving a task between different resource groups.

As pointed out by Valentin Schneider in [2] there is an additional issue with the way in which the queueing of work is done in that the task_struct update is currently done after the work is queued, resulting in a race with the register update possibly done before the data needed by the update is available.

To solve these issues, update the PQR_ASSOC MSR in a synchronous way right after the new closid and rmid are ready during the task movement, only if the task is running. If a moved task is not running nothing is done since the PQR_ASSOC MSR will be updated next time the task is scheduled. This is the same way used to update the register when tasks are moved as part of resource group removal.

[1] https://lore.kernel.org/lkml/CALvZod7E9zzHwenzf7objzGKsdBmVwTgEJ0nPgs0LUFU3SN5Pw@mail.gmail.com/
[2] https://lore.kernel.org/lkml/[email protected]

[ bp: Massage commit message and drop the two update_task_closid_rmid()
variants. ]

Backporting notes:

Since upstream commit fa7d949337cc ("x86/resctrl: Rename and move rdt files to a separate directory"), the file arch/x86/kernel/cpu/intel_rdt_rdtgroup.c has been renamed and moved to arch/x86/kernel/cpu/resctrl/rdtgroup.c.
Apply the change against file arch/x86/kernel/cpu/intel_rdt_rdtgroup.c for older stable trees.

Since upstream commit 352940ececaca ("x86/resctrl: Rename the RDT functions and definitions"), resctrl functions received more generic names. Specifically related to this backport, intel_rdt_sched_in() was renamed to rescrl_sched_in().

---
 arch/x86/kernel/cpu/intel_rdt_rdtgroup.c |  108 ++++++++++++-------------------
 1 file changed, 43 insertions(+), 65 deletions(-)

@@ -533,85 +533,63 @@ static void rdtgroup_remove(struct rdtgr
 	kfree(rdtgrp);
 }
 
-struct task_move_callback {
-	struct callback_head	work;
-	struct rdtgroup		*rdtgrp;
-};
-
-static void move_myself(struct callback_head *head)
+static void _update_task_closid_rmid(void *task)
 {
-	struct task_move_callback *callback;
-	struct rdtgroup *rdtgrp;
-
-	callback = container_of(head, struct task_move_callback, work);
-	rdtgrp = callback->rdtgrp;
-
 	/*
-	 * If resource group was deleted before this task work callback
-	 * was invoked, then assign the task to root group and free the
-	 * resource group.
+	 * If the task is still current on this CPU, update PQR_ASSOC MSR.
+	 * Otherwise, the MSR is updated when the task is scheduled in.
 	 */
-	if (atomic_dec_and_test(&rdtgrp->waitcount) &&
-	    (rdtgrp->flags & RDT_DELETED)) {
-		current->closid = 0;
-		current->rmid = 0;
-		rdtgroup_remove(rdtgrp);
-	}
-
-	preempt_disable();
-	/* update PQR_ASSOC MSR to make resource group go into effect */
-	intel_rdt_sched_in();
-	preempt_enable();
+	if (task == current)
+		intel_rdt_sched_in();
+}
 
-	kfree(callback);
+static void update_task_closid_rmid(struct task_struct *t)
+{
+	if (IS_ENABLED(CONFIG_SMP) && task_curr(t))
+		smp_call_function_single(task_cpu(t), _update_task_closid_rmid, t, 1);
+	else
+		_update_task_closid_rmid(t);
 }
 
 static int __rdtgroup_move_task(struct task_struct *tsk,
 				struct rdtgroup *rdtgrp)
 {
-	struct task_move_callback *callback;
-	int ret;
-
-	callback = kzalloc(sizeof(*callback), GFP_KERNEL);
-	if (!callback)
-		return -ENOMEM;
-	callback->work.func = move_myself;
-	callback->rdtgrp = rdtgrp;
-
 	/*
-	 * Take a refcount, so rdtgrp cannot be freed before the
-	 * callback has been invoked.
+	 * Set the task's closid/rmid before the PQR_ASSOC MSR can be
+	 * updated by them.
+	 *
+	 * For ctrl_mon groups, move both closid and rmid.
+	 * For monitor groups, can move the tasks only from
+	 * their parent CTRL group.
 	 */
-	atomic_inc(&rdtgrp->waitcount);
-	ret = task_work_add(tsk, &callback->work, true);
-	if (ret) {
-		/*
-		 * Task is exiting. Drop the refcount and free the callback.
-		 * No need to check the refcount as the group cannot be
-		 * deleted before the write function unlocks rdtgroup_mutex.
-		 */
-		atomic_dec(&rdtgrp->waitcount);
-		kfree(callback);
-		rdt_last_cmd_puts("task exited\n");
-	} else {
-		/*
-		 * For ctrl_mon groups move both closid and rmid.
-		 * For monitor groups, can move the tasks only from
-		 * their parent CTRL group.
-		 */
-		if (rdtgrp->type == RDTCTRL_GROUP) {
-			tsk->closid = rdtgrp->closid;
+
+	if (rdtgrp->type == RDTCTRL_GROUP) {
+		tsk->closid = rdtgrp->closid;
+		tsk->rmid = rdtgrp->mon.rmid;
+	} else if (rdtgrp->type == RDTMON_GROUP) {
+		if (rdtgrp->mon.parent->closid == tsk->closid) {
 			tsk->rmid = rdtgrp->mon.rmid;
-		} else if (rdtgrp->type == RDTMON_GROUP) {
-			if (rdtgrp->mon.parent->closid == tsk->closid) {
-				tsk->rmid = rdtgrp->mon.rmid;
-			} else {
-				rdt_last_cmd_puts("Can't move task to different control group\n");
-				ret = -EINVAL;
-			}
+		} else {
+			rdt_last_cmd_puts("Can't move task to different control group\n");
+			return -EINVAL;
 		}
 	}
-	return ret;
+
+	/*
+	 * Ensure the task's closid and rmid are written before determining if
+	 * the task is current that will decide if it will be interrupted.
+	 */
+	barrier();
+
+	/*
+	 * By now, the task's closid and rmid are set. If the task is current
+	 * on a CPU, the PQR_ASSOC MSR needs to be updated to make the resource
+	 * group go into effect. If the task is not current, the MSR will be
+	 * updated when the task is scheduled in.
+	 */
+	update_task_closid_rmid(tsk);
+
+	return 0;
 }
 
 /**

From: Fenghua Yu <[email protected]>

commit a0195f314a25582b38993bf30db11c300f4f4611 upstream

Shakeel Butt reported in [1] that a user can request a task to be moved to a resource group even if the task is already in the group. It just wastes time to do the move operation which could be costly to send IPI to a different CPU.

Add a sanity check to ensure that the move operation only happens when the task is not already in the resource group.

[1] https://lore.kernel.org/lkml/CALvZod7E9zzHwenzf7objzGKsdBmVwTgEJ0nPgs0LUFU3SN5Pw@mail.gmail.com/

Backporting notes:

Since upstream commit fa7d949337cc ("x86/resctrl: Rename and move rdt files to a separate directory"), the file arch/x86/kernel/cpu/intel_rdt_rdtgroup.c has been renamed and moved to arch/x86/kernel/cpu/resctrl/rdtgroup.c.
Apply the change against file arch/x86/kernel/cpu/intel_rdt_rdtgroup.c for older stable trees.

---
 arch/x86/kernel/cpu/intel_rdt_rdtgroup.c |    7 +++++++
 1 file changed, 7 insertions(+)

@@ -554,6 +554,13 @@ static void update_task_closid_rmid(stru
 static int __rdtgroup_move_task(struct task_struct *tsk,
 				struct rdtgroup *rdtgrp)
 {
+	/* If the task is already in rdtgrp, no need to move the task. */
+	if ((rdtgrp->type == RDTCTRL_GROUP && tsk->closid == rdtgrp->closid &&
+	     tsk->rmid == rdtgrp->mon.rmid) ||
+	    (rdtgrp->type == RDTMON_GROUP && tsk->rmid == rdtgrp->mon.rmid &&
+	     tsk->closid == rdtgrp->mon.parent->closid))
+		return 0;
+
 	/*
 	 * Set the task's closid/rmid before the PQR_ASSOC MSR can be
 	 * updated by them.

From: Nick Desaulniers <[email protected]>

commit eff8728fe69880d3f7983bec3fb6cea4c306261f upstream.

Basically, consider .text.{hot|unlikely|unknown}.* part of .text, too.

When compiling with profiling information (collected via PGO instrumentations or AutoFDO sampling), Clang will separate code into .text.hot, .text.unlikely, or .text.unknown sections based on profiling information. After D79600 (clang-11), these sections will have a trailing . suffix, ie. .text.hot., .text.unlikely., .text.unknown..

When using -ffunction-sections together with profiling infomation, either explicitly (FGKASLR) or implicitly (LTO), code may be placed in sections following the convention:
.text.hot.<foo>, .text.unlikely.<bar>, .text.unknown.<baz> where <foo>, <bar>, and <baz> are functions. (This produces one section per function; we generally try to merge these all back via linker script so that we don't have 50k sections).

For the above cases, we need to teach our linker scripts that such sections might exist and that we'd explicitly like them grouped together, otherwise we can wind up with code outside of the _stext/_etext boundaries that might not be mapped properly for some architectures, resulting in boot failures.

If the linker script is not told about possible input sections, then where the section is placed as output is a heuristic-laiden mess that's non-portable between linkers (ie. BFD and LLD), and has resulted in many hard to debug bugs. Kees Cook is working on cleaning this up by adding --orphan-handling=warn linker flag used in ARCH=powerpc to additional architectures. In the case of linker scripts, borrowing from the Zen of
Python: explicit is better than implicit.

Also, ld.bfd's internal linker script considers .text.hot AND .text.hot.* to be part of .text, as well as .text.unlikely and .text.unlikely.*. I didn't see support for .text.unknown.*, and didn't see Clang producing such code in our kernel builds, but I see code in LLVM that can produce such section names if profiling information is missing. That may point to a larger issue with generating or collecting profiles, but I would much rather be safe and explicit than have to debug yet another issue related to orphan section placement.

---
 include/asm-generic/vmlinux.lds.h |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

@@ -492,7 +492,10 @@
  */
 #define TEXT_TEXT							\
 		ALIGN_FUNCTION();					\
-		*(.text.hot TEXT_MAIN .text.fixup .text.unlikely)	\
+		*(.text.hot .text.hot.*)				\
+		*(TEXT_MAIN .text.fixup)				\
+		*(.text.unlikely .text.unlikely.*)			\
+		*(.text.unknown .text.unknown.*)			\
 		*(.text..refcount)					\
 		*(.ref.text)						\
 	MEM_KEEP(init.text*)						\

From: Chris Wilson <[email protected]>

commit 0e53656ad8abc99e0a80c3de611e593ebbf55829 upstream

When inserting a VMA, we restrict the placement to the low 4G unless the caller opts into using the full range. This was done to allow usersapce the opportunity to transition slowly from a 32b address space, and to avoid breaking inherent 32b assumptions of some commands.

However, for insert we limited ourselves to 4G-4K, but on verification we allowed the full 4G. This causes some attempts to bind a new buffer to sporadically fail with -ENOSPC, but at other times be bound successfully.

commit 48ea1e32c39d ("drm/i915/gen9: Set PIN_ZONE_4G end to 4GB - 1 page") suggests that there is a genuine problem with stateless addressing that cannot utilize the last page in 4G and so we purposefully excluded it. This means that the quick pin pass may cause us to utilize a buggy placement.

Testcase: igt/gem_exec_params/larger-than-life-batch

---
 drivers/gpu/drm/i915/i915_gem_execbuffer.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -380,7 +380,7 @@ eb_vma_misplaced(const struct drm_i915_g
 		return true;
 
 	if (!(flags & EXEC_OBJECT_SUPPORTS_48B_ADDRESS) &&
-	    (vma->node.start + vma->node.size - 1) >> 32)
+	    (vma->node.start + vma->node.size + 4095) >> 32)
 		return true;
 
 	if (flags & __EXEC_OBJECT_NEEDS_MAP &&

From: Lukas Wunner <[email protected]>

commit 5626308bb94d9f930aa5f7c77327df4c6daa7759 upstream

pxa2xx_spi_remove() accesses the driver's private data after calling spi_unregister_controller() even though that function releases the last reference on the spi_controller and thereby frees the private data.

Fix by switching over to the new devm_spi_alloc_master/slave() helper which keeps the private data accessible until the driver has unbound.

---
 drivers/spi/spi-pxa2xx.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

@@ -1572,7 +1572,7 @@ static int pxa2xx_spi_probe(struct platf
 		return -ENODEV;
 	}
 
-	master = spi_alloc_master(dev, sizeof(struct driver_data));
+	master = devm_spi_alloc_master(dev, sizeof(*drv_data));
 	if (!master) {
 		dev_err(&pdev->dev, "cannot alloc spi_master\n");
 		pxa_ssp_free(ssp);

@@ -1759,7 +1759,6 @@ out_error_dma_irq_alloc:
 	free_irq(ssp->irq, drv_data);
 
 out_error_master_alloc:
-	spi_controller_put(master);
 	pxa_ssp_free(ssp);
 	return status;
 }

From: Sean Nyekjaer <[email protected]>

commit ec76d918f23034f9f662539ca9c64e2ae3ba9fba upstream

No need for using reverse logic in the irq return, fix this by flip things around.

---
 drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -481,7 +481,7 @@ static irqreturn_t st_lsm6dsx_handler_th
 	count = st_lsm6dsx_read_fifo(hw);
 	mutex_unlock(&hw->fifo_lock);
 
-	return !count ? IRQ_NONE : IRQ_HANDLED;
+	return count ? IRQ_HANDLED : IRQ_NONE;
 }
 
 static int st_lsm6dsx_buffer_preenable(struct iio_dev *iio_dev)

From: Lorenzo Bianconi <[email protected]>

commit 3f9bce7a22a3f8ac9d885c9d75bc45569f24ac8b upstream

If we are using edge IRQs, new samples can arrive while processing current interrupt since there are no hw guarantees the irq line stays "low" long enough to properly detect the new interrupt. In this case the new sample will be missed.
Polling FIFO status register in st_lsm6dsx_handler_thread routine allow us to read new samples even if the interrupt arrives while processing previous data and the timeslot where the line is "low" is too short to be properly detected.

---
 drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_buffer.c |   26 ++++++++++++++++++++-----
 1 file changed, 21 insertions(+), 5 deletions(-)

@@ -475,13 +475,29 @@ static irqreturn_t st_lsm6dsx_handler_ir
 static irqreturn_t st_lsm6dsx_handler_thread(int irq, void *private)
 {
 	struct st_lsm6dsx_hw *hw = private;
-	int count;
+	int fifo_len = 0, len;
 
-	mutex_lock(&hw->fifo_lock);
-	count = st_lsm6dsx_read_fifo(hw);
-	mutex_unlock(&hw->fifo_lock);
+	/*
+	 * If we are using edge IRQs, new samples can arrive while
+	 * processing current interrupt since there are no hw
+	 * guarantees the irq line stays "low" long enough to properly
+	 * detect the new interrupt. In this case the new sample will
+	 * be missed.
+	 * Polling FIFO status register allow us to read new
+	 * samples even if the interrupt arrives while processing
+	 * previous data and the timeslot where the line is "low" is
+	 * too short to be properly detected.
+	 */
+	do {
+		mutex_lock(&hw->fifo_lock);
+		len = st_lsm6dsx_read_fifo(hw);
+		mutex_unlock(&hw->fifo_lock);
 
-	return count ? IRQ_HANDLED : IRQ_NONE;
+		if (len > 0)
+			fifo_len += len;
+	} while (len > 0);
+
+	return fifo_len ? IRQ_HANDLED : IRQ_NONE;
 }
 
 static int st_lsm6dsx_buffer_preenable(struct iio_dev *iio_dev)

From: Ping Cheng <[email protected]>

commit 37309f47e2f5674f3e86cb765312ace42cfcedf5 upstream.

As reported by syzbot below, kfifo_alloc'd memory would not be freed if a non-zero return value is triggered in wacom_probe. This patch creates and uses devm_kfifo_alloc to allocate and free itself.

BUG: memory leak

unreferenced object 0xffff88810dc44a00 (size 512):
comm "kworker/1:2", pid 3674, jiffies 4294943617 (age 14.100s) hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................ 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:

   [<0000000023e1afac>] kmalloc_array include/linux/slab.h:592 [inline]
   [<0000000023e1afac>] __kfifo_alloc+0xad/0x100 lib/kfifo.c:43
   [<00000000c477f737>] wacom_probe+0x1a1/0x3b0 drivers/hid/wacom_sys.c:2727
   [<00000000b3109aca>] hid_device_probe+0x16b/0x210 drivers/hid/hid-core.c:2281
   [<00000000aff7c640>] really_probe+0x159/0x480 drivers/base/dd.c:554
   [<00000000778d0bc3>] driver_probe_device+0x84/0x100 drivers/base/dd.c:738
   [<000000005108dbb5>] __device_attach_driver+0xee/0x110 drivers/base/dd.c:844
   [<00000000efb7c59e>] bus_for_each_drv+0xb7/0x100 drivers/base/bus.c:431
   [<0000000024ab1590>] __device_attach+0x122/0x250 drivers/base/dd.c:912
   [<000000004c7ac048>] bus_probe_device+0xc6/0xe0 drivers/base/bus.c:491
   [<00000000b93050a3>] device_add+0x5ac/0xc30 drivers/base/core.c:2936
   [<00000000e5b46ea5>] hid_add_device+0x151/0x390 drivers/hid/hid-core.c:2437
   [<00000000c6add147>] usbhid_probe+0x412/0x560 drivers/hid/usbhid/hid-core.c:1407
   [<00000000c33acdb4>] usb_probe_interface+0x177/0x370 drivers/usb/core/driver.c:396
   [<00000000aff7c640>] really_probe+0x159/0x480 drivers/base/dd.c:554
   [<00000000778d0bc3>] driver_probe_device+0x84/0x100 drivers/base/dd.c:738
   [<000000005108dbb5>] __device_attach_driver+0xee/0x110 drivers/base/dd.c:844

https://syzkaller.appspot.com/bug?extid=5b49c9695968d7250a26

---
 drivers/hid/wacom_sys.c |   35 ++++++++++++++++++++++++++++++++---
 1 file changed, 32 insertions(+), 3 deletions(-)

@@ -1241,6 +1241,37 @@ static int wacom_devm_sysfs_create_group
 					       group);
 }
 
+static void wacom_devm_kfifo_release(struct device *dev, void *res)
+{
+	struct kfifo_rec_ptr_2 *devres = res;
+
+	kfifo_free(devres);
+}
+
+static int wacom_devm_kfifo_alloc(struct wacom *wacom)
+{
+	struct wacom_wac *wacom_wac = &wacom->wacom_wac;
+	struct kfifo_rec_ptr_2 *pen_fifo = &wacom_wac->pen_fifo;
+	int error;
+
+	pen_fifo = devres_alloc(wacom_devm_kfifo_release,
+			      sizeof(struct kfifo_rec_ptr_2),
+			      GFP_KERNEL);
+
+	if (!pen_fifo)
+		return -ENOMEM;
+
+	error = kfifo_alloc(pen_fifo, WACOM_PKGLEN_MAX, GFP_KERNEL);
+	if (error) {
+		devres_free(pen_fifo);
+		return error;
+	}
+
+	devres_add(&wacom->hdev->dev, pen_fifo);
+
+	return 0;
+}
+
 enum led_brightness wacom_leds_brightness_get(struct wacom_led *led)
 {
 	struct wacom *wacom = led->wacom;

@@ -2697,7 +2728,7 @@ static int wacom_probe(struct hid_device
 		goto fail;
 	}
 
-	error = kfifo_alloc(&wacom_wac->pen_fifo, WACOM_PKGLEN_MAX, GFP_KERNEL);
+	error = wacom_devm_kfifo_alloc(wacom);
 	if (error)
 		goto fail;
 

@@ -2764,8 +2795,6 @@ static void wacom_remove(struct hid_devi
 	if (wacom->wacom_wac.features.type != REMOTE)
 		wacom_release_resources(wacom);
 
-	kfifo_free(&wacom_wac->pen_fifo);
-
 	hid_set_drvdata(hdev, NULL);
 }
 

From: Andreas Kemnade <[email protected]>

commit ec76c2eea903947202098090bbe07a739b5246e9 upstream.

On the GTA04A5 od->_driver_status was not set to BUS_NOTIFY_BIND_DRIVER during probe of the second mmc used for wifi. Therefore omap_device_late_idle idled the device during probing causing oopses when accessing the registers.

It was not set because od->_state was set to OMAP_DEVICE_STATE_IDLE in the notifier callback. Therefore set od->_driver_status also in that case.

This came apparent after commit 21b2cec61c04 ("mmc: Set PROBE_PREFER_ASYNCHRONOUS for drivers that existed in v4.4") causing this oops:

omap_hsmmc 480b4000.mmc: omap_device_late_idle: enabled but no driver. Idling 8<--- cut here ---
Unhandled fault: external abort on non-linefetch (0x1028) at 0xfa0b402c
...
(omap_hsmmc_set_bus_width) from [<c07996bc>] (omap_hsmmc_set_ios+0x11c/0x258) (omap_hsmmc_set_ios) from [<c077b2b0>] (mmc_power_up.part.8+0x3c/0xd0) (mmc_power_up.part.8) from [<c077c14c>] (mmc_start_host+0x88/0x9c) (mmc_start_host) from [<c077d284>] (mmc_add_host+0x58/0x84) (mmc_add_host) from [<c0799190>] (omap_hsmmc_probe+0x5fc/0x8c0) (omap_hsmmc_probe) from [<c0666728>] (platform_drv_probe+0x48/0x98) (platform_drv_probe) from [<c066457c>] (really_probe+0x1dc/0x3b4)

---
 arch/arm/mach-omap2/omap_device.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

@@ -239,10 +239,12 @@ static int _omap_device_notifier_call(st
 		break;
 	case BUS_NOTIFY_BIND_DRIVER:
 		od = to_omap_device(pdev);
-		if (od && (od->_state == OMAP_DEVICE_STATE_ENABLED) &&
-		    pm_runtime_status_suspended(dev)) {
+		if (od) {
 			od->_driver_status = BUS_NOTIFY_BIND_DRIVER;
-			pm_runtime_set_active(dev);
+			if (od->_state == OMAP_DEVICE_STATE_ENABLED &&
+			    pm_runtime_status_suspended(dev)) {
+				pm_runtime_set_active(dev);
+			}
 		}
 		break;
 	case BUS_NOTIFY_ADD_DEVICE:

From: Chunyan Zhang <[email protected]>

commit 0b884fe71f9ee6a5df35e677154256ea2099ebb8 upstream.

If the i2c device SCL bus being pulled up due to some exception before message transfer done, the system cannot receive the completing interrupt signal any more, it would not exit waiting loop until MAX_SCHEDULE_TIMEOUT jiffies eclipse, that would make the system seemed hang up. To avoid that happen, this patch adds a specific timeout for message transfer.

---
 drivers/i2c/busses/i2c-sprd.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

@@ -71,6 +71,8 @@
 
 /* timeout (ms) for pm runtime autosuspend */
 #define SPRD_I2C_PM_TIMEOUT	1000
+/* timeout (ms) for transfer message */
+#define I2C_XFER_TIMEOUT	1000
 
 /* SPRD i2c data structure */
 struct sprd_i2c {

@@ -244,6 +246,7 @@ static int sprd_i2c_handle_msg(struct i2
 			       struct i2c_msg *msg, bool is_last_msg)
 {
 	struct sprd_i2c *i2c_dev = i2c_adap->algo_data;
+	unsigned long time_left;
 
 	i2c_dev->msg = msg;
 	i2c_dev->buf = msg->buf;

@@ -273,7 +276,10 @@ static int sprd_i2c_handle_msg(struct i2
 
 	sprd_i2c_opt_start(i2c_dev);
 
-	wait_for_completion(&i2c_dev->complete);
+	time_left = wait_for_completion_timeout(&i2c_dev->complete,
+				msecs_to_jiffies(I2C_XFER_TIMEOUT));
+	if (!time_left)
+		return -ETIMEDOUT;
 
 	return i2c_dev->err;
 }

From: Colin Ian King <[email protected]>

commit 943bdd0cecad06da8392a33093230e30e501eccc upstream.

Currently there is an unlikely case where cpufreq_cpu_get() returns a NULL policy and this will cause a NULL pointer dereference later on.

Fix this by passing the policy to transition_frequency_fidvid() from the caller and hence eliminating the need for the cpufreq_cpu_get() and cpufreq_cpu_put().

Thanks to Viresh Kumar for suggesting the fix.

Addresses-Coverity: ("Dereference null return")

---
 drivers/cpufreq/powernow-k8.c |    9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

@@ -881,9 +881,9 @@ static int get_transition_latency(struct
 
 /* Take a frequency, and issue the fid/vid transition command */
 static int transition_frequency_fidvid(struct powernow_k8_data *data,
-		unsigned int index)
+		unsigned int index,
+		struct cpufreq_policy *policy)
 {
-	struct cpufreq_policy *policy;
 	u32 fid = 0;
 	u32 vid = 0;
 	int res;

@@ -915,9 +915,6 @@ static int transition_frequency_fidvid(s
 	freqs.old = find_khz_freq_from_fid(data->currfid);
 	freqs.new = find_khz_freq_from_fid(fid);
 
-	policy = cpufreq_cpu_get(smp_processor_id());
-	cpufreq_cpu_put(policy);
-
 	cpufreq_freq_transition_begin(policy, &freqs);
 	res = transition_fid_vid(data, fid, vid);
 	cpufreq_freq_transition_end(policy, &freqs, res);

@@ -972,7 +969,7 @@ static long powernowk8_target_fn(void *a
 
 	powernow_k8_acpi_pst_values(data, newstate);
 
-	ret = transition_frequency_fidvid(data, newstate);
+	ret = transition_frequency_fidvid(data, newstate, pol);
 
 	if (ret) {
 		pr_err("transition frequency failed\n");

From: Roman Guskov <[email protected]>

commit a590370d918fc66c62df6620445791fbe840344a upstream.

if cur_bpw <= 8 and xfer_len < 4 then the value of fthlv will be 1 and SPI registers content may have been lost.

* If SPI data register is accessed as a 16-bit register and DSIZE <= 8bit, better to select FTHLV = 2, 4, 6 etc

* If SPI data register is accessed as a 32-bit register and DSIZE > 8bit, better to select FTHLV = 2, 4, 6 etc, while if DSIZE <= 8bit, better to select FTHLV = 4, 8, 12 etc

---
 drivers/spi/spi-stm32.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -299,9 +299,9 @@ static u32 stm32_spi_prepare_fthlv(struc
 
 	/* align packet size with data registers access */
 	if (spi->cur_bpw > 8)
-		fthlv -= (fthlv % 2); /* multiple of 2 */
+		fthlv += (fthlv % 2) ? 1 : 0;
 	else
-		fthlv -= (fthlv % 4); /* multiple of 4 */
+		fthlv += (fthlv % 4) ? (4 - (fthlv % 4)) : 0;
 
 	return fthlv;
 }

From: Christophe JAILLET <[email protected]>

commit 33cbd54dc515cc04b5a603603414222b4bb1448d upstream.

'mtk_hsdma_hw_deinit()' should be called in the error handling path of the probe function to undo a previous 'mtk_hsdma_hw_init()' call, as already done in the remove function.

---
 drivers/dma/mediatek/mtk-hsdma.c |    1 +
 1 file changed, 1 insertion(+)

@@ -1007,6 +1007,7 @@ static int mtk_hsdma_probe(struct platfo
 	return 0;
 
 err_free:
+	mtk_hsdma_hw_deinit(hsdma);
 	of_dma_controller_free(pdev->dev.of_node);
 err_unregister:
 	dma_async_device_unregister(dd);

From: Shravya Kumbham <[email protected]>

commit 99974aedbd73523969afb09f33c6e3047cd0ddae upstream.

dma_async_device_register() can return non-zero error code. Add condition to check the return value of dma_async_device_register function and handle the error path.

Addresses-Coverity: Event check_return.

---
 drivers/dma/xilinx/xilinx_dma.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

@@ -2713,7 +2713,11 @@ static int xilinx_dma_probe(struct platf
 	}
 
 	/* Register the DMA engine with the core */
-	dma_async_device_register(&xdev->common);
+	err = dma_async_device_register(&xdev->common);
+	if (err) {
+		dev_err(xdev->dev, "failed to register the dma device\n");
+		goto error;
+	}
 
 	err = of_dma_controller_register(node, of_dma_xilinx_xlate,
 					 xdev);

From: Shravya Kumbham <[email protected]>

commit faeb0731be0a31e2246b21a85fa7dabbd750101d upstream.

In xilinx_dma_child_probe function, the nr_channels variable is passed to of_property_read_u32() which expects an u32 return value pointer. Modify the nr_channels variable type from int to u32 to fix the incompatible parameter coverity warning.

Addresses-Coverity: Event incompatible_param.

---
 drivers/dma/xilinx/xilinx_dma.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -2529,7 +2529,8 @@ static int xilinx_dma_chan_probe(struct
 static int xilinx_dma_child_probe(struct xilinx_dma_device *xdev,
 				    struct device_node *node)
 {
-	int ret, i, nr_channels = 1;
+	int ret, i;
+	u32 nr_channels = 1;
 
 	ret = of_property_read_u32(node, "dma-channels", &nr_channels);
 	if ((ret < 0) && xdev->mcdma)

From: Shravya Kumbham <[email protected]>

commit 2d5efea64472469117dc1a9a39530069e95b21e9 upstream.

Typecast the fls(width -1) with (enum dmaengine_alignment) in xilinx_dma_chan_probe function to fix the coverity warning.

Addresses-Coverity: Event mixed_enum_type.

---
 drivers/dma/xilinx/xilinx_dma.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -2426,7 +2426,7 @@ static int xilinx_dma_chan_probe(struct
 		has_dre = false;
 
 	if (!has_dre)
-		xdev->common.copy_align = fls(width - 1);
+		xdev->common.copy_align = (enum dmaengine_alignment)fls(width - 1);
 
 	if (of_device_is_compatible(node, "xlnx,axi-vdma-mm2s-channel") ||
 	    of_device_is_compatible(node, "xlnx,axi-dma-mm2s-channel") ||

From: Arnd Bergmann <[email protected]>

commit e186620d7bf11b274b985b839c38266d7918cc05 upstream.

Without crc32, the driver fails to link:

arm-linux-gnueabi-ld: drivers/net/wireless/ath/wil6210/fw.o: in function `wil_fw_verify':
fw.c:(.text+0x74c): undefined reference to `crc32_le'
arm-linux-gnueabi-ld: drivers/net/wireless/ath/wil6210/fw.o:fw.c:(.text+0x758): more undefined references to `crc32_le' follow

---
 drivers/net/wireless/ath/wil6210/Kconfig |    1 +
 1 file changed, 1 insertion(+)

@@ -1,6 +1,7 @@
 config WIL6210
 	tristate "Wilocity 60g WiFi card wil6210 support"
 	select WANT_DEV_COREDUMP
+	select CRC32
 	depends on CFG80211
 	depends on PCI
 	default n

From: Arnd Bergmann <[email protected]>

commit 36a106a4c1c100d55ba3d32a21ef748cfcd4fa99 upstream.

Without crc32, the driver fails to link:

arm-linux-gnueabi-ld: drivers/block/rsxx/config.o: in function `rsxx_load_config':
config.c:(.text+0x124): undefined reference to `crc32_le'

---
 drivers/block/Kconfig |    1 +
 1 file changed, 1 insertion(+)

@@ -474,6 +474,7 @@ config BLK_DEV_RBD
 config BLK_DEV_RSXX
 	tristate "IBM Flash Adapter 900GB Full Height PCIe Device Driver"
 	depends on PCI
+	select CRC32
 	help
 	  Device driver for IBM's high speed PCIe SSD
 	  storage device: Flash Adapter 900GB Full Height.

From: Arnd Bergmann <[email protected]>

commit 19cd3403cb0d522dd5e10188eef85817de29e26e upstream.

Without CRC32 support, this fails to link:

arm-linux-gnueabi-ld: drivers/lightnvm/pblk-init.o: in function `pblk_init':
pblk-init.c:(.text+0x2654): undefined reference to `crc32_le'
arm-linux-gnueabi-ld: drivers/lightnvm/pblk-init.o: in function `pblk_exit':
pblk-init.c:(.text+0x2a7c): undefined reference to `crc32_le'

---
 drivers/lightnvm/Kconfig |    1 +
 1 file changed, 1 insertion(+)

@@ -19,6 +19,7 @@ if NVM
 
 config NVM_PBLK
 	tristate "Physical Block Device Open-Channel SSD target"
+	select CRC32
 	help
 	  Allows an open-channel SSD to be exposed as a block device to the
 	  host. The target assumes the device exposes raw flash and must be

From: Dinghao Liu <[email protected]>

commit ff2b46d7cff80d27d82f7f3252711f4ca1666129 upstream.

When irq_domain_get_irq_data() or irqd_cfg() fails at i == 0, data allocated by kzalloc() has not been freed before returning, which leads to memleak.

---
 drivers/iommu/intel_irq_remapping.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -1373,6 +1373,8 @@ static int intel_irq_remapping_alloc(str
 		irq_data = irq_domain_get_irq_data(domain, virq + i);
 		irq_cfg = irqd_cfg(irq_data);
 		if (!irq_data || !irq_cfg) {
+			if (!i)
+				kfree(data);
 			ret = -EINVAL;
 			goto out_free_data;
 		}

From: Dinghao Liu <[email protected]>

commit 5b0bb12c58ac7d22e05b5bfdaa30a116c8c32e32 upstream.

When mlx5_create_flow_group() fails, ft->g should be freed just like when kvzalloc() fails. The caller of mlx5e_create_l2_table_groups() does not catch this issue on failure, which leads to memleak.

---
 drivers/net/ethernet/mellanox/mlx5/core/en_fs.c |    1 +
 1 file changed, 1 insertion(+)

@@ -1312,6 +1312,7 @@ err_destroy_groups:
 	ft->g[ft->num_groups] = NULL;
 	mlx5e_destroy_groups(ft);
 	kvfree(in);
+	kfree(ft->g);
 
 	return err;
 }

From: Dinghao Liu <[email protected]>

commit 7a6eb072a9548492ead086f3e820e9aac71c7138 upstream.

mlx5e_create_ttc_table_groups() frees ft->g on failure of kvzalloc(), but such failure will be caught by its caller in mlx5e_create_ttc_table() and ft->g will be freed again in mlx5e_destroy_flow_table(). The same issue also occurs in mlx5e_create_ttc_table_groups(). Set ft->g to NULL after kfree() to avoid double free.

---
 drivers/net/ethernet/mellanox/mlx5/core/en_fs.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -893,6 +893,7 @@ static int mlx5e_create_ttc_table_groups
 	in = kvzalloc(inlen, GFP_KERNEL);
 	if (!in) {
 		kfree(ft->g);
+		ft->g = NULL;
 		return -ENOMEM;
 	}
 

@@ -1033,6 +1034,7 @@ static int mlx5e_create_inner_ttc_table_
 	in = kvzalloc(inlen, GFP_KERNEL);
 	if (!in) {
 		kfree(ft->g);
+		ft->g = NULL;
 		return -ENOMEM;
 	}
 

From: Xiaolei Wang <[email protected]>

commit cffa4b2122f5f3e53cf3d529bbc74651f95856d5 upstream.

After initializing the regmap through syscon_regmap_lookup_by_compatible, then regmap_attach_dev to the device, because the debugfs_name has been allocated, there is no need to redistribute it again

unreferenced object 0xd8399b80 (size 64):
comm "swapper/0", pid 1, jiffies 4294937641 (age 278.590s) hex dump (first 32 bytes):

	64 75 6d 6d 79 2d 69 6f 6d 75 78 63 2d 67 70 72
dummy-iomuxc-gpr
	40 32 30 65 34 30 30 30 00 7f 52 5b d8 7e 42 69
@20e4000..R[.~Bi
  backtrace:
    [<ca384d6f>] kasprintf+0x2c/0x54
    [<6ad3bbc2>] regmap_debugfs_init+0xdc/0x2fc
    [<bc4181da>] __regmap_init+0xc38/0xd88
    [<1f7e0609>] of_syscon_register+0x168/0x294
    [<735e8766>] device_node_get_regmap+0x6c/0x98
    [<d96c8982>] imx6ul_init_machine+0x20/0x88
    [<0456565b>] customize_machine+0x1c/0x30
    [<d07393d8>] do_one_initcall+0x80/0x3ac
    [<7e584867>] kernel_init_freeable+0x170/0x1f0
    [<80074741>] kernel_init+0x8/0x120
    [<285d6f28>] ret_from_fork+0x14/0x20
    [<00000000>] 0x0

---
 drivers/base/regmap/regmap-debugfs.c |   11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

@@ -579,18 +579,25 @@ void regmap_debugfs_init(struct regmap *
 		devname = dev_name(map->dev);
 
 	if (name) {
-		map->debugfs_name = kasprintf(GFP_KERNEL, "%s-%s",
+		if (!map->debugfs_name) {
+			map->debugfs_name = kasprintf(GFP_KERNEL, "%s-%s",
 					      devname, name);
+			if (!map->debugfs_name)
+				return;
+		}
 		name = map->debugfs_name;
 	} else {
 		name = devname;
 	}
 
 	if (!strcmp(name, "dummy")) {
-		kfree(map->debugfs_name);
+		if (!map->debugfs_name)
+			kfree(map->debugfs_name);
 
 		map->debugfs_name = kasprintf(GFP_KERNEL, "dummy%d",
 						dummy_index);
+		if (!map->debugfs_name)
+				return;
 		name = map->debugfs_name;
 		dummy_index++;
 	}

From: Arnd Bergmann <[email protected]>

commit 69931e11288520c250152180ecf9b6ac5e6e40ed upstream.

Without this, the driver runs into a link failure

arm-linux-gnueabi-ld: drivers/net/wan/slic_ds26522.o: in function `slic_ds26522_probe':
slic_ds26522.c:(.text+0x100c): undefined reference to `byte_rev_table'
arm-linux-gnueabi-ld: slic_ds26522.c:(.text+0x1cdc): undefined reference to `byte_rev_table'
arm-linux-gnueabi-ld: drivers/net/wan/slic_ds26522.o: in function `slic_write':
slic_ds26522.c:(.text+0x1e4c): undefined reference to `byte_rev_table'

---
 drivers/net/wan/Kconfig |    1 +
 1 file changed, 1 insertion(+)

@@ -295,6 +295,7 @@ config SLIC_DS26522
 	tristate "Slic Maxim ds26522 card support"
 	depends on SPI
 	depends on FSL_SOC || ARCH_MXC || ARCH_LAYERSCAPE || COMPILE_TEST
+	select BITREVERSE
 	help
 	  This module initializes and configures the slic maxim card
 	  in T1 or E1 mode.

From: Marc Zyngier <[email protected]>

commit 2a5f1b67ec577fb1544b563086e0377f095f88e2 upstream.

We reset the guest's view of PMCR_EL0 unconditionally, based on the host's view of this register. It is however legal for an implementation not to provide any PMU, resulting in an UNDEF.

The obvious fix is to skip the reset of this shadow register when no PMU is available, sidestepping the issue entirely. If no PMU is available, the guest is not able to request a virtual PMU anyway, so not doing nothing is the right thing to do!

It is unlikely that this bug can hit any HW implementation though, as they all provide a PMU. It has been found using nested virt with the host KVM not implementing the PMU itself.

---
 arch/arm64/kvm/sys_regs.c |    4 ++++
 1 file changed, 4 insertions(+)

@@ -619,6 +619,10 @@ static void reset_pmcr(struct kvm_vcpu *
 {
 	u64 pmcr, val;
 
+	/* No PMU available, PMCR_EL0 may UNDEF... */
+	if (!kvm_arm_support_pmu_v3())
+		return;
+
 	pmcr = read_sysreg(pmcr_el0);
 	/*
 	 * Writable bits of PMCR_EL0 (ARMV8_PMU_PMCR_MASK) are reset to UNKNOWN

From: Ming Lei <[email protected]>

commit aebf5db917055b38f4945ed6d621d9f07a44ff30 upstream.

Make sure that bdgrab() is done on the 'block_device' instance before referring to it for avoiding use-after-free.

---
 block/genhd.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

@@ -208,14 +208,17 @@ struct hd_struct *disk_part_iter_next(st
 		part = rcu_dereference(ptbl->part[piter->idx]);
 		if (!part)
 			continue;
+		get_device(part_to_dev(part));
+		piter->part = part;
 		if (!part_nr_sects_read(part) &&
 		    !(piter->flags & DISK_PITER_INCL_EMPTY) &&
 		    !(piter->flags & DISK_PITER_INCL_EMPTY_PART0 &&
-		      piter->idx == 0))
+		      piter->idx == 0)) {
+			put_device(part_to_dev(part));
+			piter->part = NULL;
 			continue;
+		}
 
-		get_device(part_to_dev(part));
-		piter->part = part;
 		piter->idx += inc;
 		break;
 	}

From: Vasily Averin <[email protected]>

commit 54970a2fbb673f090b7f02d7f57b10b2e0707155 upstream.

syzbot reproduces BUG_ON in skb_checksum_help():
tun creates (bogus) skb with huge partial-checksummed area and small ip packet inside. Then ip_rcv trims the skb based on size of internal ip packet, after that csum offset points beyond of trimmed skb. Then checksum_tg() called via netfilter hook triggers BUG_ON:

        offset = skb_checksum_start_offset(skb);
        BUG_ON(offset >= skb_headlen(skb));

To work around the problem this patch forces pskb_trim_rcsum_slow() to return -EINVAL in described scenario. It allows its callers to drop such kind of packets.

---
 net/core/skbuff.c |    6 ++++++
 1 file changed, 6 insertions(+)

@@ -1853,6 +1853,12 @@ int pskb_trim_rcsum_slow(struct sk_buff
 		skb->csum = csum_block_sub(skb->csum,
 					   skb_checksum(skb, len, delta, 0),
 					   len);
+	} else if (skb->ip_summed == CHECKSUM_PARTIAL) {
+		int hdlen = (len > skb_headlen(skb)) ? skb_headlen(skb) : len;
+		int offset = skb_checksum_start_offset(skb) + skb->csum_offset;
+
+		if (offset + sizeof(__sum16) > hdlen)
+			return -EINVAL;
 	}
 	return __pskb_trim(skb, len);
 }

From: Dan Carpenter <[email protected]>

commit f6bcb4c7f366905b66ce8ffca7190118244bb642 upstream.

This code will leak "map->debugfs_name" because the if statement is reversed so it only frees NULL pointers instead of non-NULL. In fact the if statement is not required and should just be removed because kfree() accepts NULL pointers.

---
 drivers/base/regmap/regmap-debugfs.c |    4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

@@ -591,9 +591,7 @@ void regmap_debugfs_init(struct regmap *
 	}
 
 	if (!strcmp(name, "dummy")) {
-		if (!map->debugfs_name)
-			kfree(map->debugfs_name);
-
+		kfree(map->debugfs_name);
 		map->debugfs_name = kasprintf(GFP_KERNEL, "dummy%d",
 						dummy_index);
 		if (!map->debugfs_name)

All tests passing for Tegra ...

Test results for stable-v4.19:

    12 builds:	12 pass, 0 fail
    22 boots:	22 pass, 0 fail
    38 tests:	38 pass, 0 fail

Linux version: 4.19.168-rc1-g710affe26b43
Boards tested: tegra124-jetson-tk1, tegra186-p2771-0000,

                tegra194-p2972-0000, tegra20-ventana,
                tegra210-p2371-2180, tegra30-cardhu-a04

Jon

Compiled and booted on my test system. No dmesg regressions.

thanks,
-- Shuah

Build results:

	total: 155 pass: 155 fail: 0
Qemu test results:
	total: 418 pass: 418 fail: 0

Guenter

Results from Linaro’s test farm.
No regressions on arm64, arm, x86_64, and i386.

Summary
------------------------------------------------------------------------

kernel: 4.19.168-rc1
git repo: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git
git branch: linux-4.19.y
git commit: 710affe26b4363782b54ed8cc7bf77d6fdfbe151
git describe: v4.19.167-44-g710affe26b43
Test details: https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-4.19.y/build/v4.19.167-44-g710affe26b43

No regressions (compared to build v4.19.167)

No fixes (compared to build v4.19.167)

Ran 48282 total tests in the following environments and test suites.

Environments
--------------
- arm
- arm64
- dragonboard-410c - arm64
- hi6220-hikey - arm64
- i386
- juno-r2 - arm64
- juno-r2-compat
- juno-r2-kasan
- mips
- nxp-ls2088
- qemu-arm64-clang
- qemu-arm64-kasan
- qemu-x86_64-clang
- qemu-x86_64-kasan
- qemu_arm
- qemu_arm64
- qemu_arm64-compat
- qemu_i386
- qemu_x86_64
- qemu_x86_64-compat
- s390
- sparc
- x15 - arm
- x86_64
- x86-kasan
- x86_64

Test Suites
-----------
* build
* linux-log-parser
* install-android-platform-tools-r2600
* kselftest
* kvm-unit-tests
* ltp-cap_bounds-tests
* ltp-commands-tests
* ltp-containers-tests
* ltp-controllers-tests
* ltp-cpuhotplug-tests
* ltp-crypto-tests
* ltp-cve-tests
* ltp-math-tests
* ltp-nptl-tests
* ltp-pty-tests
* ltp-sched-tests
* ltp-securebits-tests
* ltp-syscalls-tests
* perf
* fwts
* libhugetlbfs
* ltp-fcntl-locktests-tests
* ltp-filecaps-tests
* ltp-fs-tests
* ltp-fs_bind-tests
* ltp-fs_perms_simple-tests
* ltp-fsx-tests
* ltp-hugetlb-tests
* ltp-io-tests
* ltp-ipc-tests
* ltp-mm-tests
* network-basic-tests
* v4l2-compliance
* ltp-dio-tests
* ltp-open-posix-tests
* ltp-tracing-tests
* rcutorture
* kselftest-vsyscall-mode-native
* kselftest-vsyscall-mode-none

Linaro LKFT
https://lkft.linaro.org

Hi!

CIP testing did not find any problems here:

https://gitlab.com/cip-project/cip-testing/linux-stable-rc-ci/-/tree/linux-4.19.y

Best regards,
Pavel

DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany