6.9.8-rc1 review

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Oliver Neukum <[email protected]>

commit fd80731e5e9d1402cb2f85022a6abf9b1982ec5f upstream.

We need to treat super speed plus as super speed, not the default, which is full speed.

---
 drivers/usb/gadget/function/f_printer.c |    1 +
 1 file changed, 1 insertion(+)

@@ -213,6 +213,7 @@ static inline struct usb_endpoint_descri
 					struct usb_endpoint_descriptor *ss)
 {
 	switch (gadget->speed) {
+	case USB_SPEED_SUPER_PLUS:
 	case USB_SPEED_SUPER:
 		return ss;
 	case USB_SPEED_HIGH:

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Oliver Neukum <[email protected]>

commit e587a7633dfee8987a999cf253f7c52a8e09276c upstream.

printer_read() and printer_write() guard against the race against disable() by checking the dev->interface flag, which in turn is guarded by a spinlock.
These functions, however, drop the lock on multiple occasions. This means that the test has to be redone after reacquiring the lock and before doing IO.

Add the tests.

This also addresses CVE-2024-25741

---
 drivers/usb/gadget/function/f_printer.c |   39 +++++++++++++++++++++++---------
 1 file changed, 29 insertions(+), 10 deletions(-)

@@ -450,11 +450,8 @@ printer_read(struct file *fd, char __use
 	mutex_lock(&dev->lock_printer_io);
 	spin_lock_irqsave(&dev->lock, flags);
 
-	if (dev->interface < 0) {
-		spin_unlock_irqrestore(&dev->lock, flags);
-		mutex_unlock(&dev->lock_printer_io);
-		return -ENODEV;
-	}
+	if (dev->interface < 0)
+		goto out_disabled;
 
 	/* We will use this flag later to check if a printer reset happened
 	 * after we turn interrupts back on.

@@ -462,6 +459,9 @@ printer_read(struct file *fd, char __use
 	dev->reset_printer = 0;
 
 	setup_rx_reqs(dev);
+	/* this dropped the lock - need to retest */
+	if (dev->interface < 0)
+		goto out_disabled;
 
 	bytes_copied = 0;
 	current_rx_req = dev->current_rx_req;

@@ -495,6 +495,8 @@ printer_read(struct file *fd, char __use
 		wait_event_interruptible(dev->rx_wait,
 				(likely(!list_empty(&dev->rx_buffers))));
 		spin_lock_irqsave(&dev->lock, flags);
+		if (dev->interface < 0)
+			goto out_disabled;
 	}
 
 	/* We have data to return then copy it to the caller's buffer.*/

@@ -538,6 +540,9 @@ printer_read(struct file *fd, char __use
 			return -EAGAIN;
 		}
 
+		if (dev->interface < 0)
+			goto out_disabled;
+
 		/* If we not returning all the data left in this RX request
 		 * buffer then adjust the amount of data left in the buffer.
 		 * Othewise if we are done with this RX request buffer then

@@ -567,6 +572,11 @@ printer_read(struct file *fd, char __use
 		return bytes_copied;
 	else
 		return -EAGAIN;
+
+out_disabled:
+	spin_unlock_irqrestore(&dev->lock, flags);
+	mutex_unlock(&dev->lock_printer_io);
+	return -ENODEV;
 }
 
 static ssize_t

@@ -587,11 +597,8 @@ printer_write(struct file *fd, const cha
 	mutex_lock(&dev->lock_printer_io);
 	spin_lock_irqsave(&dev->lock, flags);
 
-	if (dev->interface < 0) {
-		spin_unlock_irqrestore(&dev->lock, flags);
-		mutex_unlock(&dev->lock_printer_io);
-		return -ENODEV;
-	}
+	if (dev->interface < 0)
+		goto out_disabled;
 
 	/* Check if a printer reset happens while we have interrupts on */
 	dev->reset_printer = 0;

@@ -614,6 +621,8 @@ printer_write(struct file *fd, const cha
 		wait_event_interruptible(dev->tx_wait,
 				(likely(!list_empty(&dev->tx_reqs))));
 		spin_lock_irqsave(&dev->lock, flags);
+		if (dev->interface < 0)
+			goto out_disabled;
 	}
 
 	while (likely(!list_empty(&dev->tx_reqs)) && len) {

@@ -663,6 +672,9 @@ printer_write(struct file *fd, const cha
 			return -EAGAIN;
 		}
 
+		if (dev->interface < 0)
+			goto out_disabled;
+
 		list_add(&req->list, &dev->tx_reqs_active);
 
 		/* here, we unlock, and only unlock, to avoid deadlock. */

@@ -675,6 +687,8 @@ printer_write(struct file *fd, const cha
 			mutex_unlock(&dev->lock_printer_io);
 			return -EAGAIN;
 		}
+		if (dev->interface < 0)
+			goto out_disabled;
 	}
 
 	spin_unlock_irqrestore(&dev->lock, flags);

@@ -686,6 +700,11 @@ printer_write(struct file *fd, const cha
 		return bytes_copied;
 	else
 		return -EAGAIN;
+
+out_disabled:
+	spin_unlock_irqrestore(&dev->lock, flags);
+	mutex_unlock(&dev->lock_printer_io);
+	return -ENODEV;
 }
 
 static int

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Dan Carpenter <[email protected]>

commit de644a4a86be04ed8a43ef8267d0f7d021941c5e upstream.

Call usb_phy_generic_unregister() if of_platform_populate() fails.

---
 drivers/usb/musb/da8xx.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

@@ -556,7 +556,7 @@ static int da8xx_probe(struct platform_d
 	ret = of_platform_populate(pdev->dev.of_node, NULL,
 				   da8xx_auxdata_lookup, &pdev->dev);
 	if (ret)
-		return ret;
+		goto err_unregister_phy;
 
 	pinfo = da8xx_dev_info;
 	pinfo.parent = &pdev->dev;

@@ -571,9 +571,13 @@ static int da8xx_probe(struct platform_d
 	ret = PTR_ERR_OR_ZERO(glue->musb);
 	if (ret) {
 		dev_err(&pdev->dev, "failed to register musb device: %d\n", ret);
-		usb_phy_generic_unregister(glue->usb_phy);
+		goto err_unregister_phy;
 	}
 
+	return 0;
+
+err_unregister_phy:
+	usb_phy_generic_unregister(glue->usb_phy);
 	return ret;
 }
 

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Nikita Zhandarovich <[email protected]>

commit 2eabb655a968b862bc0c31629a09f0fbf3c80d51 upstream.

Syzbot is still reporting quite an old issue [1] that occurs due to incomplete checking of present usb endpoints. As such, wrong endpoints types may be used at urb sumbitting stage which in turn triggers a warning in usb_submit_urb().

Fix the issue by verifying that required endpoint types are present for both in and out endpoints, taking into account cmd endpoint type.

Unfortunately, this patch has not been tested on real hardware.

[1] Syzbot report:
usb 1-1: BOGUS urb xfer, pipe 1 != type 3

WARNING: CPU: 0 PID: 8667 at drivers/usb/core/urb.c:502 usb_submit_urb+0xed2/0x18a0 drivers/usb/core/urb.c:502
Modules linked in:
CPU: 0 PID: 8667 Comm: kworker/0:4 Not tainted 5.14.0-rc4-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Workqueue: usb_hub_wq hub_event
RIP: 0010:usb_submit_urb+0xed2/0x18a0 drivers/usb/core/urb.c:502
...
Call Trace:
 cxacru_cm+0x3c0/0x8e0 drivers/usb/atm/cxacru.c:649
 cxacru_card_status+0x22/0xd0 drivers/usb/atm/cxacru.c:760
 cxacru_bind+0x7ac/0x11a0 drivers/usb/atm/cxacru.c:1209
 usbatm_usb_probe+0x321/0x1ae0 drivers/usb/atm/usbatm.c:1055
 cxacru_usb_probe+0xdf/0x1e0 drivers/usb/atm/cxacru.c:1363
 usb_probe_interface+0x315/0x7f0 drivers/usb/core/driver.c:396
 call_driver_probe drivers/base/dd.c:517 [inline]
 really_probe+0x23c/0xcd0 drivers/base/dd.c:595
 __driver_probe_device+0x338/0x4d0 drivers/base/dd.c:747
 driver_probe_device+0x4c/0x1a0 drivers/base/dd.c:777
 __device_attach_driver+0x20b/0x2f0 drivers/base/dd.c:894
 bus_for_each_drv+0x15f/0x1e0 drivers/base/bus.c:427
 __device_attach+0x228/0x4a0 drivers/base/dd.c:965
 bus_probe_device+0x1e4/0x290 drivers/base/bus.c:487
 device_add+0xc2f/0x2180 drivers/base/core.c:3354
 usb_set_configuration+0x113a/0x1910 drivers/usb/core/message.c:2170
 usb_generic_driver_probe+0xba/0x100 drivers/usb/core/generic.c:238
 usb_probe_device+0xd9/0x2c0 drivers/usb/core/driver.c:293

---
 drivers/usb/atm/cxacru.c |   14 ++++++++++++++
 1 file changed, 14 insertions(+)

@@ -1131,6 +1131,7 @@ static int cxacru_bind(struct usbatm_dat
 	struct cxacru_data *instance;
 	struct usb_device *usb_dev = interface_to_usbdev(intf);
 	struct usb_host_endpoint *cmd_ep = usb_dev->ep_in[CXACRU_EP_CMD];
+	struct usb_endpoint_descriptor *in, *out;
 	int ret;
 
 	/* instance init */

@@ -1176,6 +1177,19 @@ static int cxacru_bind(struct usbatm_dat
 		ret = -ENODEV;
 		goto fail;
 	}
+
+	if (usb_endpoint_xfer_int(&cmd_ep->desc))
+		ret = usb_find_common_endpoints(intf->cur_altsetting,
+						NULL, NULL, &in, &out);
+	else
+		ret = usb_find_common_endpoints(intf->cur_altsetting,
+						&in, &out, NULL, NULL);
+
+	if (ret) {
+		usb_err(usbatm_instance, "cxacru_bind: interface has incorrect endpoints\n");
+		ret = -ENODEV;
+		goto fail;
+	}
 
 	if ((cmd_ep->desc.bmAttributes & USB_ENDPOINT_XFERTYPE_MASK)
 			== USB_ENDPOINT_XFER_INT) {

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Meng Li <[email protected]>

commit 7838de15bb700c2898a7d741db9b1f3cbc86c136 upstream.

When config CONFIG_USB_DWC3_DUAL_ROLE is selected, and trigger system to enter suspend status with below command:
echo mem > /sys/power/state
There will be a deadlock issue occurring. Detailed invoking path as below:
dwc3_suspend_common()

    spin_lock_irqsave(&dwc->lock, flags);              <-- 1st
    dwc3_gadget_suspend(dwc);
        dwc3_gadget_soft_disconnect(dwc);
            spin_lock_irqsave(&dwc->lock, flags);      <-- 2nd

This issue is exposed by commit c7ebd8149ee5 ("usb: dwc3: gadget: Fix NULL pointer dereference in dwc3_gadget_suspend") that removes the code of checking whether dwc->gadget_driver is NULL or not. It causes the following code is executed and deadlock occurs when trying to get the spinlock. In fact, the root cause is the commit 5265397f9442("usb: dwc3:
Remove DWC3 locking during gadget suspend/resume") that forgot to remove the lock of otg mode. So, remove the redundant lock of otg mode during gadget suspend/resume.

---
 drivers/usb/dwc3/core.c |    6 ------
 1 file changed, 6 deletions(-)

@@ -2084,7 +2084,6 @@ assert_reset:
 
 static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
 {
-	unsigned long	flags;
 	u32 reg;
 
 	switch (dwc->current_dr_role) {

@@ -2122,9 +2121,7 @@ static int dwc3_suspend_common(struct dw
 			break;
 
 		if (dwc->current_otg_role == DWC3_OTG_ROLE_DEVICE) {
-			spin_lock_irqsave(&dwc->lock, flags);
 			dwc3_gadget_suspend(dwc);
-			spin_unlock_irqrestore(&dwc->lock, flags);
 			synchronize_irq(dwc->irq_gadget);
 		}
 

@@ -2141,7 +2138,6 @@ static int dwc3_suspend_common(struct dw
 
 static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg)
 {
-	unsigned long	flags;
 	int		ret;
 	u32		reg;
 

@@ -2190,9 +2186,7 @@ static int dwc3_resume_common(struct dwc
 		if (dwc->current_otg_role == DWC3_OTG_ROLE_HOST) {
 			dwc3_otg_host_init(dwc);
 		} else if (dwc->current_otg_role == DWC3_OTG_ROLE_DEVICE) {
-			spin_lock_irqsave(&dwc->lock, flags);
 			dwc3_gadget_resume(dwc);
-			spin_unlock_irqrestore(&dwc->lock, flags);
 		}
 
 		break;

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Jeremy Kerr <[email protected]>

commit dba7567c2fbbf10a4de2471cdb0e16e5572dc007 upstream.

In the aspeed UDC setup, we configure the UDC hardware with the assigned USB device address.

However, we have an off-by-one in the bitmask, so we're only setting the lower 6 bits of the address (USB addresses being 7 bits, and the hardware bitmask being bits 0:6).

This means that device enumeration fails if the assigned address is greater than 64:

[  344.607255] usb 1-1: new high-speed USB device number 63 using ehci-platform
[  344.808459] usb 1-1: New USB device found, idVendor=cc00, idProduct=cc00, bcdDevice= 6.10
[  344.817684] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[  344.825671] usb 1-1: Product: Test device
[  344.831075] usb 1-1: Manufacturer: Test vendor
[  344.836335] usb 1-1: SerialNumber: 00
[  349.917181] usb 1-1: USB disconnect, device number 63
[  352.036775] usb 1-1: new high-speed USB device number 64 using ehci-platform
[  352.249432] usb 1-1: device descriptor read/all, error -71
[  352.696740] usb 1-1: new high-speed USB device number 65 using ehci-platform
[  352.909431] usb 1-1: device descriptor read/all, error -71

Use the correct mask of 0x7f (rather than 0x3f), and generate this through the GENMASK macro, so we have numbers that correspond exactly to the hardware register definition.

---
 drivers/usb/gadget/udc/aspeed_udc.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -66,8 +66,8 @@
 #define USB_UPSTREAM_EN			BIT(0)
 
 /* Main config reg */
-#define UDC_CFG_SET_ADDR(x)		((x) & 0x3f)
-#define UDC_CFG_ADDR_MASK		(0x3f)
+#define UDC_CFG_SET_ADDR(x)		((x) & UDC_CFG_ADDR_MASK)
+#define UDC_CFG_ADDR_MASK		GENMASK(6, 0)
 
 /* Interrupt ctrl & status reg */
 #define UDC_IRQ_EP_POOL_NAK		BIT(17)

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Javier Carrasco <[email protected]>

commit c68942624e254a4e8a65afcd3c17ed95acda5489 upstream.

The device_for_each_child_node() macro requires explicit calls to fwnode_handle_put() in all early exits of the loop if the child node is not required outside. Otherwise, the child node's refcount is not decremented and the resource is not released.

The current implementation of pmic_glink_ucsi_probe() makes use of the device_for_each_child_node(), but does not release the child node on early returns. Add the missing calls to fwnode_handle_put().

---
 drivers/usb/typec/ucsi/ucsi_glink.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

@@ -365,6 +365,7 @@ static int pmic_glink_ucsi_probe(struct
 		ret = fwnode_property_read_u32(fwnode, "reg", &port);
 		if (ret < 0) {
 			dev_err(dev, "missing reg property of %pOFn\n", fwnode);
+			fwnode_handle_put(fwnode);
 			return ret;
 		}
 

@@ -379,9 +380,11 @@ static int pmic_glink_ucsi_probe(struct
 		if (!desc)
 			continue;
 
-		if (IS_ERR(desc))
+		if (IS_ERR(desc)) {
+			fwnode_handle_put(fwnode);
 			return dev_err_probe(dev, PTR_ERR(desc),
 					     "unable to acquire orientation gpio\n");
+		}
 		ucsi->port_orientation[port] = desc;
 
 		ucsi->port_switch[port] = fwnode_typec_switch_get(fwnode);

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Ferry Toth <[email protected]>

commit 24bf27b92b1c6a322faa88977de2207aa8c26509 upstream.

This reverts commit 76c945730cdffb572c7767073cc6515fd3f646b4.

Prerequisite revert for the reverting of the original commit f49449fbc21e.

---
 drivers/usb/gadget/function/u_ether.c |    2 --
 1 file changed, 2 deletions(-)

@@ -1163,8 +1163,6 @@ struct net_device *gether_connect(struct
 		if (netif_running(dev->net))
 			eth_start(dev, GFP_ATOMIC);
 
-		netif_device_attach(dev->net);
-
 	/* on error, disable any endpoints  */
 	} else {
 		(void) usb_ep_disable(link->out_ep);

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Ferry Toth <[email protected]>

commit c50814a288dcee687285abc0cf935e9fe8928e59 upstream.

This reverts commit f49449fbc21e7e9550a5203902d69c8ae7dfd918.

This commit breaks u_ether on some setups (at least Merrifield). The fix "usb: gadget: u_ether: Re-attach netif device to mirror detachment" party restores u-ether. However the netif usb: remains up even usb is switched from device to host mode. This creates problems for user space as the interface remains in the routing table while not realy present and network managers (connman) not detecting a network change.

Various attempts to find the root cause were unsuccesful up to now. Therefore revert until a solution is found.

---
 drivers/usb/gadget/function/u_ether.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1200,7 +1200,7 @@ void gether_disconnect(struct gether *li
 
 	DBG(dev, "%s\n", __func__);
 
-	netif_device_detach(dev->net);
+	netif_stop_queue(dev->net);
 	netif_carrier_off(dev->net);
 
 	/* disable endpoints, forcing (synchronous) completion

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Fabrice Gasnier <[email protected]>

commit 8e1ec117efdfd4b2f59f57bd0ad16b4edf5b963f upstream.

Sometimes errors are seen, when doing DR swap, like:

[   24.672481] ucsi-stm32g0-i2c 0-0035: UCSI_GET_PDOS failed (-5)
[   24.720188] ucsi-stm32g0-i2c 0-0035: ucsi_handle_connector_change:
 GET_CONNECTOR_STATUS failed (-5)

There may be some race, which lead to read CCI, before the command complete flag is set, hence returning -EIO. Similar fix has been done also in ucsi_acpi [1].

In case of a spurious or otherwise delayed notification it is possible that CCI still reports the previous completion. The UCSI spec is aware of this and provides two completion bits in CCI, one for normal commands and one for acks. As acks and commands alternate the notification handler can determine if the completion bit is from the current command.

To fix this add the ACK_PENDING bit for ucsi_stm32g0 and only complete commands if the completion bit matches.

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

---
 drivers/usb/typec/ucsi/ucsi_stm32g0.c |   19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

@@ -64,6 +64,7 @@ struct ucsi_stm32g0 {
 	struct completion complete;
 	struct device *dev;
 	unsigned long flags;
+#define ACK_PENDING	2
 	const char *fw_name;
 	struct ucsi *ucsi;
 	bool suspended;

@@ -395,9 +396,13 @@ static int ucsi_stm32g0_sync_write(struc
 				   size_t len)
 {
 	struct ucsi_stm32g0 *g0 = ucsi_get_drvdata(ucsi);
+	bool ack = UCSI_COMMAND(*(u64 *)val) == UCSI_ACK_CC_CI;
 	int ret;
 
-	set_bit(COMMAND_PENDING, &g0->flags);
+	if (ack)
+		set_bit(ACK_PENDING, &g0->flags);
+	else
+		set_bit(COMMAND_PENDING, &g0->flags);
 
 	ret = ucsi_stm32g0_async_write(ucsi, offset, val, len);
 	if (ret)

@@ -405,9 +410,14 @@ static int ucsi_stm32g0_sync_write(struc
 
 	if (!wait_for_completion_timeout(&g0->complete, msecs_to_jiffies(5000)))
 		ret = -ETIMEDOUT;
+	else
+		return 0;
 
 out_clear_bit:
-	clear_bit(COMMAND_PENDING, &g0->flags);
+	if (ack)
+		clear_bit(ACK_PENDING, &g0->flags);
+	else
+		clear_bit(COMMAND_PENDING, &g0->flags);
 
 	return ret;
 }

@@ -428,8 +438,9 @@ static irqreturn_t ucsi_stm32g0_irq_hand
 	if (UCSI_CCI_CONNECTOR(cci))
 		ucsi_connector_change(g0->ucsi, UCSI_CCI_CONNECTOR(cci));
 
-	if (test_bit(COMMAND_PENDING, &g0->flags) &&
-	    cci & (UCSI_CCI_ACK_COMPLETE | UCSI_CCI_COMMAND_COMPLETE))
+	if (cci & UCSI_CCI_ACK_COMPLETE && test_and_clear_bit(ACK_PENDING, &g0->flags))
+		complete(&g0->complete);
+	if (cci & UCSI_CCI_COMMAND_COMPLETE && test_and_clear_bit(COMMAND_PENDING, &g0->flags))
 		complete(&g0->complete);
 
 	return IRQ_HANDLED;

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Jos Wang <[email protected]>

commit fc1d1a712b517bbcb383b1f1f7ef478e7d0579f2 upstream.

This is a workaround for STAR 4846132, which only affects DWC_usb31 version2.00a operating in host mode.

There is a problem in DWC_usb31 version 2.00a operating in host mode that would cause a CSR read timeout When CSR read coincides with RAM Clock Gating Entry. By disable Clock Gating, sacrificing power consumption for normal operation.

---
 drivers/usb/dwc3/core.c |   20 +++++++++++++++++++-
 1 file changed, 19 insertions(+), 1 deletion(-)

@@ -879,12 +879,16 @@ static bool dwc3_core_is_valid(struct dw
 
 static void dwc3_core_setup_global_control(struct dwc3 *dwc)
 {
+	unsigned int power_opt;
+	unsigned int hw_mode;
 	u32 reg;
 
 	reg = dwc3_readl(dwc->regs, DWC3_GCTL);
 	reg &= ~DWC3_GCTL_SCALEDOWN_MASK;
+	hw_mode = DWC3_GHWPARAMS0_MODE(dwc->hwparams.hwparams0);
+	power_opt = DWC3_GHWPARAMS1_EN_PWROPT(dwc->hwparams.hwparams1);
 
-	switch (DWC3_GHWPARAMS1_EN_PWROPT(dwc->hwparams.hwparams1)) {
+	switch (power_opt) {
 	case DWC3_GHWPARAMS1_EN_PWROPT_CLK:
 		/**
 		 * WORKAROUND: DWC3 revisions between 2.10a and 2.50a have an

@@ -917,6 +921,20 @@ static void dwc3_core_setup_global_contr
 		break;
 	}
 
+	/*
+	 * This is a workaround for STAR#4846132, which only affects
+	 * DWC_usb31 version2.00a operating in host mode.
+	 *
+	 * There is a problem in DWC_usb31 version 2.00a operating
+	 * in host mode that would cause a CSR read timeout When CSR
+	 * read coincides with RAM Clock Gating Entry. By disable
+	 * Clock Gating, sacrificing power consumption for normal
+	 * operation.
+	 */
+	if (power_opt != DWC3_GHWPARAMS1_EN_PWROPT_NO &&
+	    hw_mode != DWC3_GHWPARAMS0_MODE_GADGET && DWC3_VER_IS(DWC31, 200A))
+		reg |= DWC3_GCTL_DSBLCLKGTNG;
+
 	/* check if current dwc3 is on simulation board */
 	if (dwc->hwparams.hwparams6 & DWC3_GHWPARAMS6_EN_FPGA) {
 		dev_info(dwc->dev, "Running with FPGA optimizations\n");

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Doug Brown <[email protected]>

commit c5603e2a621dac10c5e21cc430848ebcfa6c7e01 upstream.

This reverts commit 7bfb915a597a301abb892f620fe5c283a9fdbd77.

This commit broke pxa and omap-serial, because it inhibited them from calling stop_tx() if their TX FIFOs weren't completely empty. This resulted in these two drivers hanging during transmits because the TX interrupt would stay enabled, and a new TX interrupt would never fire.

---
 include/linux/serial_core.h |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

@@ -789,8 +789,7 @@ enum UART_TX_FLAGS {
 	if (pending < WAKEUP_CHARS) {					      \
 		uart_write_wakeup(__port);				      \
 									      \
-		if (!((flags) & UART_TX_NOSTOP) && pending == 0 &&	      \
-		    __port->ops->tx_empty(__port))			      \
+		if (!((flags) & UART_TX_NOSTOP) && pending == 0)	      \
 			__port->ops->stop_tx(__port);			      \
 	}								      \
 									      \

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Crescent Hsieh <[email protected]>

commit 0ac18dac43103ab1df6d26ec9a781c0126f83ced upstream.

Normally, the number of ports is indicated by the third digit of the device ID on Moxa PCI serial boards. For example, 0x1121 indicates a device with 2 ports.

However, CP116E_A_A and CP116E_A_B are exceptions; they have 8 ports, but the third digit of the device ID is 6.

This patch introduces a function to retrieve the number of ports on Moxa PCI serial boards, addressing the issue described above.

---
 drivers/tty/serial/8250/8250_pci.c |   13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

@@ -1985,6 +1985,17 @@ enum {
 	MOXA_SUPP_RS485 = BIT(2),
 };
 
+static unsigned short moxa_get_nports(unsigned short device)
+{
+	switch (device) {
+	case PCI_DEVICE_ID_MOXA_CP116E_A_A:
+	case PCI_DEVICE_ID_MOXA_CP116E_A_B:
+		return 8;
+	}
+
+	return FIELD_GET(0x00F0, device);
+}
+
 static bool pci_moxa_is_mini_pcie(unsigned short device)
 {
 	if (device == PCI_DEVICE_ID_MOXA_CP102N	||

@@ -2038,7 +2049,7 @@ static int pci_moxa_init(struct pci_dev
 {
 	unsigned short device = dev->device;
 	resource_size_t iobar_addr = pci_resource_start(dev, 2);
-	unsigned int num_ports = (device & 0x00F0) >> 4, i;
+	unsigned int i, num_ports = moxa_get_nports(device);
 	u8 val, init_mode = MOXA_RS232;
 
 	if (!(pci_moxa_supported_rs(dev) & MOXA_SUPP_RS232)) {

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Udit Kumar <[email protected]>

commit 9d141c1e615795eeb93cd35501ad144ee997a826 upstream.

As per Errata i2310[0], Erroneous timeout can be triggered, if this Erroneous interrupt is not cleared then it may leads to storm of interrupts, therefore apply Errata i2310 solution.

[0] https://www.ti.com/lit/pdf/sprz536 page 23

---
 drivers/tty/serial/8250/8250_omap.c |   25 ++++++++++++++++++++-----
 1 file changed, 20 insertions(+), 5 deletions(-)

@@ -116,6 +116,10 @@
 /* RX FIFO occupancy indicator */
 #define UART_OMAP_RX_LVL		0x19
 
+/* Timeout low and High */
+#define UART_OMAP_TO_L                 0x26
+#define UART_OMAP_TO_H                 0x27
+
 /*
  * Copy of the genpd flags for the console.
  * Only used if console suspend is disabled

@@ -664,13 +668,24 @@ static irqreturn_t omap8250_irq(int irq,
 
 	/*
 	 * On K3 SoCs, it is observed that RX TIMEOUT is signalled after
-	 * FIFO has been drained, in which case a dummy read of RX FIFO
-	 * is required to clear RX TIMEOUT condition.
+	 * FIFO has been drained or erroneously.
+	 * So apply solution of Errata i2310 as mentioned in
+	 * https://www.ti.com/lit/pdf/sprz536
 	 */
 	if (priv->habit & UART_RX_TIMEOUT_QUIRK &&
-	    (iir & UART_IIR_RX_TIMEOUT) == UART_IIR_RX_TIMEOUT &&
-	    serial_port_in(port, UART_OMAP_RX_LVL) == 0) {
-		serial_port_in(port, UART_RX);
+		(iir & UART_IIR_RX_TIMEOUT) == UART_IIR_RX_TIMEOUT) {
+		unsigned char efr2, timeout_h, timeout_l;
+
+		efr2 = serial_in(up, UART_OMAP_EFR2);
+		timeout_h = serial_in(up, UART_OMAP_TO_H);
+		timeout_l = serial_in(up, UART_OMAP_TO_L);
+		serial_out(up, UART_OMAP_TO_H, 0xFF);
+		serial_out(up, UART_OMAP_TO_L, 0xFF);
+		serial_out(up, UART_OMAP_EFR2, UART_OMAP_EFR2_TIMEOUT_BEHAVE);
+		serial_in(up, UART_IIR);
+		serial_out(up, UART_OMAP_EFR2, efr2);
+		serial_out(up, UART_OMAP_TO_H, timeout_h);
+		serial_out(up, UART_OMAP_TO_L, timeout_l);
 	}
 
 	/* Stop processing interrupts on input overrun */

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Stefan Eichenberger <[email protected]>

commit a81dbd0463eca317eee44985a66aa6cc2ce5c101 upstream.

Set the receiver level to something > 0 before calling imx_uart_start_rx in rs485_config. This is necessary to avoid an interrupt storm that might prevent the system from booting. This was seen on an i.MX7 device when the rs485-rts-active-low property was active in the device tree.

---
 drivers/tty/serial/imx.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

@@ -1960,8 +1960,10 @@ static int imx_uart_rs485_config(struct
 
 	/* Make sure Rx is enabled in case Tx is active with Rx disabled */
 	if (!(rs485conf->flags & SER_RS485_ENABLED) ||
-	    rs485conf->flags & SER_RS485_RX_DURING_TX)
+	    rs485conf->flags & SER_RS485_RX_DURING_TX) {
+		imx_uart_setup_ufcr(sport, TXTL_DEFAULT, RXTL_DEFAULT);
 		imx_uart_start_rx(port);
+	}
 
 	return 0;
 }

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Jonas Gorski <[email protected]>

commit 9bb43b9e8d9a288a214e9b17acc9e46fda3977cf upstream.

Analogue to uart_port_tx_flags() introduced in commit 3ee07964d407 ("serial: core: introduce uart_port_tx_flags()"), add a _flags variant for uart_port_tx_limited().

---
 include/linux/serial_core.h |   18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

@@ -829,6 +829,24 @@ enum UART_TX_FLAGS {
 })
 
 /**
+ * uart_port_tx_limited_flags -- transmit helper for uart_port with count limiting with flags
+ * @port: uart port
+ * @ch: variable to store a character to be written to the HW
+ * @flags: %UART_TX_NOSTOP or similar
+ * @count: a limit of characters to send
+ * @tx_ready: can HW accept more data function
+ * @put_char: function to write a character
+ * @tx_done: function to call after the loop is done
+ *
+ * See uart_port_tx_limited() for more details.
+ */
+#define uart_port_tx_limited_flags(port, ch, flags, count, tx_ready, put_char, tx_done) ({ \
+	unsigned int __count = (count);							   \
+	__uart_port_tx(port, ch, flags, tx_ready, put_char, tx_done, __count,		   \
+			__count--);							   \
+})
+
+/**
  * uart_port_tx -- transmit helper for uart_port
  * @port: uart port
  * @ch: variable to store a character to be written to the HW

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Jonas Gorski <[email protected]>

commit ea55c65dedf40e9c1911dc1e63e26bc9a59692b9 upstream.

When bcm63xx-uart was converted to uart_port_tx_limited(), it implicitly added a call to stop_tx(). This causes garbage to be put out on the serial console. To fix this, pass UART_TX_NOSTOP in flags, and manually call stop_tx() ourselves analogue to how a similar issue was fixed in commit 7be50f2e8f20 ("serial: mxs-auart: fix tx").

---
 drivers/tty/serial/bcm63xx_uart.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

@@ -308,8 +308,8 @@ static void bcm_uart_do_tx(struct uart_p
 
 	val = bcm_uart_readl(port, UART_MCTL_REG);
 	val = (val & UART_MCTL_TXFIFOFILL_MASK) >> UART_MCTL_TXFIFOFILL_SHIFT;
-
-	pending = uart_port_tx_limited(port, ch, port->fifosize - val,
+	pending = uart_port_tx_limited_flags(port, ch, UART_TX_NOSTOP,
+		port->fifosize - val,
 		true,
 		bcm_uart_writel(port, ch, UART_FIFO_REG),
 		({}));

@@ -320,6 +320,9 @@ static void bcm_uart_do_tx(struct uart_p
 	val = bcm_uart_readl(port, UART_IR_REG);
 	val &= ~UART_TX_INT_MASK;
 	bcm_uart_writel(port, val, UART_IR_REG);
+
+	if (uart_tx_stopped(port))
+		bcm_uart_stop_tx(port);
 }
 
 /*

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Dirk Su <[email protected]>

commit 3cd59d8ef8df7d7a079f54d56502dae8f716b39b upstream.

HP EliteBook 645/665 G11 needs ALC236_FIXUP_HP_MUTE_LED_MICMUTE_VREF quirk to make mic-mute/audio-mute working.

---
 sound/pci/hda/patch_realtek.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -10187,6 +10187,9 @@ static const struct snd_pci_quirk alc269
 	SND_PCI_QUIRK(0x103c, 0x8c7c, "HP ProBook 445 G11", ALC236_FIXUP_HP_MUTE_LED_MICMUTE_VREF),
 	SND_PCI_QUIRK(0x103c, 0x8c7d, "HP ProBook 465 G11", ALC236_FIXUP_HP_MUTE_LED_MICMUTE_VREF),
 	SND_PCI_QUIRK(0x103c, 0x8c7e, "HP ProBook 465 G11", ALC236_FIXUP_HP_MUTE_LED_MICMUTE_VREF),
+	SND_PCI_QUIRK(0x103c, 0x8c7f, "HP EliteBook 645 G11", ALC236_FIXUP_HP_MUTE_LED_MICMUTE_VREF),
+	SND_PCI_QUIRK(0x103c, 0x8c80, "HP EliteBook 645 G11", ALC236_FIXUP_HP_MUTE_LED_MICMUTE_VREF),
+	SND_PCI_QUIRK(0x103c, 0x8c81, "HP EliteBook 665 G11", ALC236_FIXUP_HP_MUTE_LED_MICMUTE_VREF),
 	SND_PCI_QUIRK(0x103c, 0x8c89, "HP ProBook 460 G11", ALC236_FIXUP_HP_GPIO_LED),
 	SND_PCI_QUIRK(0x103c, 0x8c8a, "HP EliteBook 630", ALC236_FIXUP_HP_GPIO_LED),
 	SND_PCI_QUIRK(0x103c, 0x8c8c, "HP EliteBook 660", ALC236_FIXUP_HP_GPIO_LED),

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Nathan Chancellor <[email protected]>

commit 1c07c9be87dd3dd0634033bf08728b32465f08fb upstream.

Work for __counted_by on generic pointers in structures (not just flexible array members) has started landing in Clang 19 (current tip of tree). During the development of this feature, a restriction was added to __counted_by to prevent the flexible array member's element type from including a flexible array member itself such as:

  struct foo {
    int count;
    char buf[];
  };

  struct bar {
    int count;
    struct foo data[] __counted_by(count);
  };

because the size of data cannot be calculated with the standard array size formula:

sizeof(struct foo) * count

This restriction was downgraded to a warning but due to CONFIG_WERROR, it can still break the build. The application of __counted_by on the ports member of 'struct mxser_board' triggers this restriction, resulting in:

drivers/tty/mxser.c:291:2: error: 'counted_by' should not be applied to an array with element of unknown size because 'struct mxser_port' is a struct type with a flexible array member. This will be an error in a future compiler version [-Werror,-Wbounds-safety-counted-by-elt-type-unknown-size]

    291 |         struct mxser_port ports[] __counted_by(nports);
        |         ^~~~~~~~~~~~~~~~~~~~~~~~~

1 error generated.

Remove this use of __counted_by to fix the warning/error. However, rather than remove it altogether, leave it commented, as it may be possible to support this in future compiler releases.

---
 drivers/tty/mxser.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -288,7 +288,7 @@ struct mxser_board {
 	enum mxser_must_hwid must_hwid;
 	speed_t max_baud;
 
-	struct mxser_port ports[] __counted_by(nports);
+	struct mxser_port ports[] /* __counted_by(nports) */;
 };
 
 static DECLARE_BITMAP(mxser_boards, MXSER_BOARDS);

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Jean-Michel Hautbois <[email protected]>

commit 7c92a8bd53f24d50c8cf4aba53bb75505b382fed upstream.

Most of the colfires have up to 5 UARTs but MCF54418 has up-to 10 ! Change the maximum value authorized.

---
 drivers/tty/serial/mcf.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -462,7 +462,7 @@ static const struct uart_ops mcf_uart_op
 	.verify_port	= mcf_verify_port,
 };
 
-static struct mcf_uart mcf_ports[4];
+static struct mcf_uart mcf_ports[10];
 
 #define	MCF_MAXPORTS	ARRAY_SIZE(mcf_ports)
 

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Shigeru Yoshida <[email protected]>

commit b7cdf1dd5d2a2d8200efd98d1893684db48fe134 upstream.

syzbot reported kernel-infoleak in raw_recvmsg() [1]. j1939_send_one() creates full frame including unused data, but it doesn't initialize it. This causes the kernel-infoleak issue. Fix this by initializing unused data.

[1]

BUG: KMSAN: kernel-infoleak in instrument_copy_to_user include/linux/instrumented.h:114 [inline]
BUG: KMSAN: kernel-infoleak in copy_to_user_iter lib/iov_iter.c:24 [inline]
BUG: KMSAN: kernel-infoleak in iterate_ubuf include/linux/iov_iter.h:29 [inline]
BUG: KMSAN: kernel-infoleak in iterate_and_advance2 include/linux/iov_iter.h:245 [inline]
BUG: KMSAN: kernel-infoleak in iterate_and_advance include/linux/iov_iter.h:271 [inline]
BUG: KMSAN: kernel-infoleak in _copy_to_iter+0x366/0x2520 lib/iov_iter.c:185
 instrument_copy_to_user include/linux/instrumented.h:114 [inline]
 copy_to_user_iter lib/iov_iter.c:24 [inline]
 iterate_ubuf include/linux/iov_iter.h:29 [inline]
 iterate_and_advance2 include/linux/iov_iter.h:245 [inline]
 iterate_and_advance include/linux/iov_iter.h:271 [inline]
 _copy_to_iter+0x366/0x2520 lib/iov_iter.c:185
 copy_to_iter include/linux/uio.h:196 [inline]
 memcpy_to_msg include/linux/skbuff.h:4113 [inline]
 raw_recvmsg+0x2b8/0x9e0 net/can/raw.c:1008
 sock_recvmsg_nosec net/socket.c:1046 [inline]
 sock_recvmsg+0x2c4/0x340 net/socket.c:1068
 ____sys_recvmsg+0x18a/0x620 net/socket.c:2803
 ___sys_recvmsg+0x223/0x840 net/socket.c:2845
 do_recvmmsg+0x4fc/0xfd0 net/socket.c:2939
 __sys_recvmmsg net/socket.c:3018 [inline]
 __do_sys_recvmmsg net/socket.c:3041 [inline]
 __se_sys_recvmmsg net/socket.c:3034 [inline]
 __x64_sys_recvmmsg+0x397/0x490 net/socket.c:3034
 x64_sys_call+0xf6c/0x3b50 arch/x86/include/generated/asm/syscalls_64.h:300
 do_syscall_x64 arch/x86/entry/common.c:52 [inline]
 do_syscall_64+0xcf/0x1e0 arch/x86/entry/common.c:83
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

Uninit was created at:

 slab_post_alloc_hook mm/slub.c:3804 [inline]
 slab_alloc_node mm/slub.c:3845 [inline]
 kmem_cache_alloc_node+0x613/0xc50 mm/slub.c:3888
 kmalloc_reserve+0x13d/0x4a0 net/core/skbuff.c:577
 __alloc_skb+0x35b/0x7a0 net/core/skbuff.c:668
 alloc_skb include/linux/skbuff.h:1313 [inline]
 alloc_skb_with_frags+0xc8/0xbf0 net/core/skbuff.c:6504
 sock_alloc_send_pskb+0xa81/0xbf0 net/core/sock.c:2795
 sock_alloc_send_skb include/net/sock.h:1842 [inline]
 j1939_sk_alloc_skb net/can/j1939/socket.c:878 [inline]
 j1939_sk_send_loop net/can/j1939/socket.c:1142 [inline]
 j1939_sk_sendmsg+0xc0a/0x2730 net/can/j1939/socket.c:1277
 sock_sendmsg_nosec net/socket.c:730 [inline]
 __sock_sendmsg+0x30f/0x380 net/socket.c:745
 ____sys_sendmsg+0x877/0xb60 net/socket.c:2584
 ___sys_sendmsg+0x28d/0x3c0 net/socket.c:2638
 __sys_sendmsg net/socket.c:2667 [inline]
 __do_sys_sendmsg net/socket.c:2676 [inline]
 __se_sys_sendmsg net/socket.c:2674 [inline]
 __x64_sys_sendmsg+0x307/0x4a0 net/socket.c:2674
 x64_sys_call+0xc4b/0x3b50 arch/x86/include/generated/asm/syscalls_64.h:47
 do_syscall_x64 arch/x86/entry/common.c:52 [inline]
 do_syscall_64+0xcf/0x1e0 arch/x86/entry/common.c:83
 entry_SYSCALL_64_after_hwframe+0x77/0x7f

Bytes 12-15 of 16 are uninitialized
Memory access of size 16 starts at ffff888120969690 Data copied to user address 00000000200017c0

CPU: 1 PID: 5050 Comm: syz-executor198 Not tainted 6.9.0-rc5-syzkaller-00031-g71b1543c83d6 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 03/27/2024

---
 net/can/j1939/main.c |    6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

@@ -30,10 +30,6 @@ MODULE_ALIAS("can-proto-" __stringify(CA
 /* CAN_HDR: #bytes before can_frame data part */
 #define J1939_CAN_HDR (offsetof(struct can_frame, data))
 
-/* CAN_FTR: #bytes beyond data part */
-#define J1939_CAN_FTR (sizeof(struct can_frame) - J1939_CAN_HDR - \
-		 sizeof(((struct can_frame *)0)->data))
-
 /* lowest layer */
 static void j1939_can_recv(struct sk_buff *iskb, void *data)
 {

@@ -342,7 +338,7 @@ int j1939_send_one(struct j1939_priv *pr
 	memset(cf, 0, J1939_CAN_HDR);
 
 	/* make it a full can frame again */
-	skb_put(skb, J1939_CAN_FTR + (8 - dlc));
+	skb_put_zero(skb, 8 - dlc);
 
 	canid = CAN_EFF_FLAG |
 		(skcb->priority << 26) |

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Oleksij Rempel <[email protected]>

commit 9ad1da14ab3bf23087ae45fe399d84a109ddb81a upstream.

Addresses an issue where a CAN bus error during a BAM transmission could stall the socket queue, preventing further transmissions even after the bus error is resolved. The fix activates the next queued session after the error recovery, allowing communication to continue.

---
 net/can/j1939/transport.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -1681,6 +1681,8 @@ static int j1939_xtp_rx_rts_session_acti
 
 		j1939_session_timers_cancel(session);
 		j1939_session_cancel(session, J1939_XTP_ABORT_BUSY);
+		if (session->transmission)
+			j1939_session_deactivate_activate_next(session);
 
 		return -EBUSY;
 	}

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Oleksij Rempel <[email protected]>

commit d3e2904f71ea0fe7eaff1d68a2b0363c888ea0fb upstream.

This patch enhances error handling in scenarios with RTS (Request to Send) messages arriving closely. It replaces the less informative WARN_ON_ONCE backtraces with a new error handling method. This provides clearer error messages and allows for the early termination of problematic sessions. Previously, sessions were only released at the end of j1939_xtp_rx_rts().

Potentially this could be reproduced with something like:
testj1939 -r vcan0:0x80 &
while true; do

	# send first RTS
	cansend vcan0 18EC8090#1014000303002301;
	# send second RTS
	cansend vcan0 18EC8090#1014000303002301;
	# send abort
	cansend vcan0 18EC8090#ff00000000002301;
done

---
 net/can/j1939/transport.c |   19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

@@ -1593,8 +1593,8 @@ j1939_session *j1939_xtp_rx_rts_session_
 	struct j1939_sk_buff_cb skcb = *j1939_skb_to_cb(skb);
 	struct j1939_session *session;
 	const u8 *dat;
+	int len, ret;
 	pgn_t pgn;
-	int len;
 
 	netdev_dbg(priv->ndev, "%s\n", __func__);
 

@@ -1653,7 +1653,22 @@ j1939_session *j1939_xtp_rx_rts_session_
 	session->tskey = priv->rx_tskey++;
 	j1939_sk_errqueue(session, J1939_ERRQUEUE_RX_RTS);
 
-	WARN_ON_ONCE(j1939_session_activate(session));
+	ret = j1939_session_activate(session);
+	if (ret) {
+		/* Entering this scope indicates an issue with the J1939 bus.
+		 * Possible scenarios include:
+		 * - A time lapse occurred, and a new session was initiated
+		 *   due to another packet being sent correctly. This could
+		 *   have been caused by too long interrupt, debugger, or being
+		 *   out-scheduled by another task.
+		 * - The bus is receiving numerous erroneous packets, either
+		 *   from a malfunctioning device or during a test scenario.
+		 */
+		netdev_alert(priv->ndev, "%s: 0x%p: concurrent session with same addr (%02x %02x) is already active.\n",
+			     __func__, session, skcb.addr.sa, skcb.addr.da);
+		j1939_session_put(session);
+		return NULL;
+	}
 
 	return session;
 }

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Mostafa Saleh <[email protected]>

commit 9eee5330656bf92f51cb1f09b2dc9f8cf975b3d1 upstream.

KFENCE reports the following UAF:

 BUG: KFENCE: use-after-free read in __pci_enable_msi_range+0x2c0/0x488

Use-after-free read at 0x0000000024629571 (in kfence-#12):

  __pci_enable_msi_range+0x2c0/0x488
  pci_alloc_irq_vectors_affinity+0xec/0x14c
  pci_alloc_irq_vectors+0x18/0x28

kfence-#12: 0x0000000008614900-0x00000000e06c228d, size=104, cache=kmalloc-128

allocated by task 81 on cpu 7 at 10.808142s:

  __kmem_cache_alloc_node+0x1f0/0x2bc
  kmalloc_trace+0x44/0x138
  msi_alloc_desc+0x3c/0x9c
  msi_domain_insert_msi_desc+0x30/0x78
  msi_setup_msi_desc+0x13c/0x184
  __pci_enable_msi_range+0x258/0x488
  pci_alloc_irq_vectors_affinity+0xec/0x14c
  pci_alloc_irq_vectors+0x18/0x28

freed by task 81 on cpu 7 at 10.811436s:

  msi_domain_free_descs+0xd4/0x10c
  msi_domain_free_locked.part.0+0xc0/0x1d8
  msi_domain_alloc_irqs_all_locked+0xb4/0xbc
  pci_msi_setup_msi_irqs+0x30/0x4c
  __pci_enable_msi_range+0x2a8/0x488
  pci_alloc_irq_vectors_affinity+0xec/0x14c
  pci_alloc_irq_vectors+0x18/0x28

Descriptor allocation done in:
__pci_enable_msi_range

    msi_capability_init
        msi_setup_msi_desc
            msi_insert_msi_desc
                msi_domain_insert_msi_desc
                    msi_alloc_desc
                        ...

Freed in case of failure in __msi_domain_alloc_locked() __pci_enable_msi_range
msi_capability_init
pci_msi_setup_msi_irqs
msi_domain_alloc_irqs_all_locked
msi_domain_alloc_locked
__msi_domain_alloc_locked => fails msi_domain_free_locked
...

That failure propagates back to pci_msi_setup_msi_irqs() in msi_capability_init() which accesses the descriptor for unmasking in the error exit path.

Cure it by copying the descriptor and using the copy for the error exit path unmask operation.

[ tglx: Massaged change log ]

---
 drivers/pci/msi/msi.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

@@ -349,7 +349,7 @@ static int msi_capability_init(struct pc
 			       struct irq_affinity *affd)
 {
 	struct irq_affinity_desc *masks = NULL;
-	struct msi_desc *entry;
+	struct msi_desc *entry, desc;
 	int ret;
 
 	/* Reject multi-MSI early on irq domain enabled architectures */

@@ -374,6 +374,12 @@ static int msi_capability_init(struct pc
 	/* All MSIs are unmasked by default; mask them all */
 	entry = msi_first_desc(&dev->dev, MSI_DESC_ALL);
 	pci_msi_mask(entry, msi_multi_mask(entry));
+	/*
+	 * Copy the MSI descriptor for the error path because
+	 * pci_msi_setup_msi_irqs() will free it for the hierarchical
+	 * interrupt domain case.
+	 */
+	memcpy(&desc, entry, sizeof(desc));
 
 	/* Configure MSI capability structure */
 	ret = pci_msi_setup_msi_irqs(dev, nvec, PCI_CAP_ID_MSI);

@@ -393,7 +399,7 @@ static int msi_capability_init(struct pc
 	goto unlock;
 
 err:
-	pci_msi_unmask(entry, msi_multi_mask(entry));
+	pci_msi_unmask(&desc, msi_multi_mask(&desc));
 	pci_free_msi_irqs(dev);
 fail:
 	dev->msi_enabled = 0;

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Nathan Chancellor <[email protected]>

commit 440e2051c577896275c0e0513ec26964e04c7810 upstream.

Work for __counted_by on generic pointers in structures (not just flexible array members) has started landing in Clang 19 (current tip of tree). During the development of this feature, a restriction was added to __counted_by to prevent the flexible array member's element type from including a flexible array member itself such as:

  struct foo {
    int count;
    char buf[];
  };

  struct bar {
    int count;
    struct foo data[] __counted_by(count);
  };

because the size of data cannot be calculated with the standard array size formula:

sizeof(struct foo) * count

This restriction was downgraded to a warning but due to CONFIG_WERROR, it can still break the build. The application of __counted_by on the fod member of 'struct nvmet_fc_tgt_queue' triggers this restriction, resulting in:

drivers/nvme/target/fc.c:151:2: error: 'counted_by' should not be applied to an array with element of unknown size because 'struct nvmet_fc_fcp_iod' is a struct type with a flexible array member. This will be an error in a future compiler version [-Werror,-Wbounds-safety-counted-by-elt-type-unknown-size]

    151 |         struct nvmet_fc_fcp_iod         fod[] __counted_by(sqsize);
        |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

1 error generated.

Remove this use of __counted_by to fix the warning/error. However, rather than remove it altogether, leave it commented, as it may be possible to support this in future compiler releases.

---
 drivers/nvme/target/fc.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -148,7 +148,7 @@ struct nvmet_fc_tgt_queue {
 	struct workqueue_struct		*work_q;
 	struct kref			ref;
 	/* array of fcp_iods */
-	struct nvmet_fc_fcp_iod		fod[] __counted_by(sqsize);
+	struct nvmet_fc_fcp_iod		fod[] /* __counted_by(sqsize) */;
 } __aligned(sizeof(unsigned long long));
 
 struct nvmet_fc_hostport {

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Rafael J. Wysocki <[email protected]>

commit a1ff59784b277795a613beaa5d3dd9c5595c69a7 upstream.

It is reported that single-thread performance on some hybrid systems dropped significantly after commit 7feec7430edd ("ACPI: CPPC: Only probe for _CPC if CPPC v2 is acked") which prevented _CPC from being used if the support for it had not been confirmed by the platform firmware.

The problem is that if the platform firmware does not confirm CPPC v2 support, cppc_get_perf_caps() returns an error which prevents the intel_pstate driver from enabling ITMT. Consequently, the scheduler does not get any hints on CPU performance differences, so in a hybrid system some tasks may run on CPUs with lower capacity even though they should be running on high-capacity CPUs.

To address this, modify intel_pstate to use the information from MSR_HWP_CAPABILITIES to enable ITMT if CPPC is not available (which is done already if the highest performance number coming from CPPC is not realistic).

---
 drivers/cpufreq/intel_pstate.c |   13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

@@ -357,15 +357,14 @@ static void intel_pstate_set_itmt_prio(i
 	int ret;
 
 	ret = cppc_get_perf_caps(cpu, &cppc_perf);
-	if (ret)
-		return;
-
 	/*
-	 * On some systems with overclocking enabled, CPPC.highest_perf is hardcoded to 0xff.
-	 * In this case we can't use CPPC.highest_perf to enable ITMT.
-	 * In this case we can look at MSR_HWP_CAPABILITIES bits [8:0] to decide.
+	 * If CPPC is not available, fall back to MSR_HWP_CAPABILITIES bits [8:0].
+	 *
+	 * Also, on some systems with overclocking enabled, CPPC.highest_perf is
+	 * hardcoded to 0xff, so CPPC.highest_perf cannot be used to enable ITMT.
+	 * Fall back to MSR_HWP_CAPABILITIES then too.
 	 */
-	if (cppc_perf.highest_perf == CPPC_MAX_PERF)
+	if (ret || cppc_perf.highest_perf == CPPC_MAX_PERF)
 		cppc_perf.highest_perf = HWP_HIGHEST_PERF(READ_ONCE(all_cpu_data[cpu]->hwp_cap_cached));
 
 	/*

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Huacai Chen <[email protected]>

commit 2d64eaeeeda5659d52da1af79d237269ba3c2d2c upstream.

Multi-bridge machines required that all eiointc controllers in the system are initialized, otherwise the system does not boot.

The initialization happens on the boot CPU during early boot and relies on cpu_to_node() for identifying the individual nodes.

That works when the number of possible CPUs is large enough, but with a command line limit, e.g. "nr_cpus=$N" for kdump, but fails when the CPUs of the secondary nodes are not covered.

During early ACPI enumeration all CPU to node mappings are recorded up to CONFIG_NR_CPUS. These are accessible via early_cpu_to_node() even in the case that "nr_cpus=N" truncates the number of possible CPUs and only provides the possible CPUs via cpu_to_node() translation.

Change the node lookup in the driver to use early_cpu_to_node() so that even with a limitation on the number of possible CPUs all eointc instances are initialized.

This can't obviously cure the case where CONFIG_NR_CPUS is too small.

[ tglx: Massaged changelog ]

---
 drivers/irqchip/irq-loongson-eiointc.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

@@ -15,6 +15,7 @@
 #include <linux/irqchip/chained_irq.h>
 #include <linux/kernel.h>
 #include <linux/syscore_ops.h>
+#include <asm/numa.h>
 
 #define EIOINTC_REG_NODEMAP	0x14a0
 #define EIOINTC_REG_IPMAP	0x14c0

@@ -339,7 +340,7 @@ static int __init pch_msi_parse_madt(uni
 	int node;
 
 	if (cpu_has_flatmode)
-		node = cpu_to_node(eiointc_priv[nr_pics - 1]->node * CORES_PER_EIO_NODE);
+		node = early_cpu_to_node(eiointc_priv[nr_pics - 1]->node * CORES_PER_EIO_NODE);
 	else
 		node = eiointc_priv[nr_pics - 1]->node;
 

@@ -431,7 +432,7 @@ int __init eiointc_acpi_init(struct irq_
 		goto out_free_handle;
 
 	if (cpu_has_flatmode)
-		node = cpu_to_node(acpi_eiointc->node * CORES_PER_EIO_NODE);
+		node = early_cpu_to_node(acpi_eiointc->node * CORES_PER_EIO_NODE);
 	else
 		node = acpi_eiointc->node;
 	acpi_set_vec_parent(node, priv->eiointc_domain, pch_group);

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Huacai Chen <[email protected]>

commit 6ef8eb5125722c241fd60d7b0c872d5c2e5dd4ca upstream.

After the rework of "Parallel CPU bringup", the cmdline "nosmp" and "maxcpus=0" parameters are not working anymore. These parameters set setup_max_cpus to zero and that's handed to bringup_nonboot_cpus().

The code there does a decrement before checking for zero, which brings it into the negative space and brings up all CPUs.

Add a zero check at the beginning of the function to prevent this.

[ tglx: Massaged change log ]

---
 kernel/cpu.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -1859,6 +1859,9 @@ static inline bool cpuhp_bringup_cpus_pa
 
 void __init bringup_nonboot_cpus(unsigned int max_cpus)
 {
+	if (!max_cpus)
+		return;
+
 	/* Try parallel bringup optimization if enabled */
 	if (cpuhp_bringup_cpus_parallel(max_cpus))
 		return;

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Yuntao Wang <[email protected]>

commit 932d8476399f622aa0767a4a0a9e78e5341dc0e1 upstream.

Commit 4205e4786d0b ("cpu/hotplug: Provide dynamic range for prepare stage") added a dynamic range for the prepare states, but did not handle the assignment of the dynstate variable in __cpuhp_setup_state_cpuslocked().

This causes the corresponding startup callback not to be invoked when calling __cpuhp_setup_state_cpuslocked() with the CPUHP_BP_PREPARE_DYN parameter, even though it should be.

Currently, the users of __cpuhp_setup_state_cpuslocked(), for one reason or another, have not triggered this bug.

---
 kernel/cpu.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

@@ -2449,7 +2449,7 @@ EXPORT_SYMBOL_GPL(__cpuhp_state_add_inst
  * The caller needs to hold cpus read locked while calling this function.
  * Return:
  *   On success:
- *      Positive state number if @state is CPUHP_AP_ONLINE_DYN;
+ *      Positive state number if @state is CPUHP_AP_ONLINE_DYN or CPUHP_BP_PREPARE_DYN;
  *      0 for all other states
  *   On failure: proper (negative) error code
  */

@@ -2472,7 +2472,7 @@ int __cpuhp_setup_state_cpuslocked(enum
 	ret = cpuhp_store_callbacks(state, name, startup, teardown,
 				    multi_instance);
 
-	dynstate = state == CPUHP_AP_ONLINE_DYN;
+	dynstate = state == CPUHP_AP_ONLINE_DYN || state == CPUHP_BP_PREPARE_DYN;
 	if (ret > 0 && dynstate) {
 		state = ret;
 		ret = 0;

@@ -2503,8 +2503,8 @@ int __cpuhp_setup_state_cpuslocked(enum
 out:
 	mutex_unlock(&cpuhp_state_mutex);
 	/*
-	 * If the requested state is CPUHP_AP_ONLINE_DYN, return the
-	 * dynamically allocated state in case of success.
+	 * If the requested state is CPUHP_AP_ONLINE_DYN or CPUHP_BP_PREPARE_DYN,
+	 * return the dynamically allocated state in case of success.
 	 */
 	if (!ret && dynstate)
 		return state;

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Huacai Chen <[email protected]>

commit a9c3ee5d0fdb069b54902300df6ac822027f3b0a upstream.

The liointc hardware provides separate Interrupt Status Registers (ISR) for each core. The current code uses always the ISR of core #0, which works during boot because by default all interrupts are routed to core #0.

When the interrupt routing changes in the firmware configuration then this causes interrupts to be lost because they are not configured in the corresponding core.

Use the core index to access the correct ISR instead of a hardcoded 0.

[ tglx: Massaged changelog ]

---
 drivers/irqchip/irq-loongson-liointc.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -28,7 +28,7 @@
 
 #define LIOINTC_INTC_CHIP_START	0x20
 
-#define LIOINTC_REG_INTC_STATUS	(LIOINTC_INTC_CHIP_START + 0x20)
+#define LIOINTC_REG_INTC_STATUS(core)	(LIOINTC_INTC_CHIP_START + 0x20 + (core) * 8)
 #define LIOINTC_REG_INTC_EN_STATUS	(LIOINTC_INTC_CHIP_START + 0x04)
 #define LIOINTC_REG_INTC_ENABLE	(LIOINTC_INTC_CHIP_START + 0x08)
 #define LIOINTC_REG_INTC_DISABLE	(LIOINTC_INTC_CHIP_START + 0x0c)

@@ -217,7 +217,7 @@ static int liointc_init(phys_addr_t addr
 		goto out_free_priv;
 
 	for (i = 0; i < LIOINTC_NUM_CORES; i++)
-		priv->core_isr[i] = base + LIOINTC_REG_INTC_STATUS;
+		priv->core_isr[i] = base + LIOINTC_REG_INTC_STATUS(i);
 
 	for (i = 0; i < LIOINTC_NUM_PARENT; i++)
 		priv->handler[i].parent_int_map = parent_int_map[i];

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Dragan Simic <[email protected]>

commit 9cc5f3bf63aa98bd7cc7ce8a8599077fde13283e upstream.

The compiled dtb files aren't executable, so install them with 0644 as their permission mode, instead of defaulting to 0755 for the permission mode and installing them with the executable bits set.

Some Linux distributions, including Debian, [1][2][3] already include fixes in their kernel package build recipes to change the dtb file permissions to 0644 in their kernel packages. These changes, when additionally propagated into the long-term kernel versions, will allow such distributions to remove their downstream fixes.

[1] https://salsa.debian.org/kernel-team/linux/-/merge_requests/642
[2] https://salsa.debian.org/kernel-team/linux/-/merge_requests/749
[3] https://salsa.debian.org/kernel-team/linux/-/blob/debian/6.8.12-1/debian/rules.real#L193

---
 scripts/Makefile.dtbinst |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -17,7 +17,7 @@ include $(srctree)/scripts/Kbuild.includ
 dst := $(INSTALL_DTBS_PATH)
 
 quiet_cmd_dtb_install = INSTALL $@
-      cmd_dtb_install = install -D $< $@
+      cmd_dtb_install = install -D -m 0644 $< $@
 
 $(dst)/%: $(obj)/%
 	$(call cmd,dtb_install)

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Arnd Bergmann <[email protected]>

commit 30766f1105d6d2459c3b9fe34a3e52b637a72950 upstream.

The unusual function calling conventions on SuperH ended up causing sync_file_range to have the wrong argument order, with the 'flags' argument getting sorted before 'nbytes' by the compiler.

In userspace, I found that musl, glibc, uclibc and strace all expect the normal calling conventions with 'nbytes' last, so changing the kernel to match them should make all of those work.

In order to be able to also fix libc implementations to work with existing kernels, they need to be able to tell which ABI is used. An easy way to do this is to add yet another system call using the sync_file_range2 ABI that works the same on all architectures.

Old user binaries can now work on new kernels, and new binaries can try the new sync_file_range2() to work with new kernels or fall back to the old sync_file_range() version if that doesn't exist.

---
 arch/sh/kernel/sys_sh32.c           |   11 +++++++++++
 arch/sh/kernel/syscalls/syscall.tbl |    3 ++-
 2 files changed, 13 insertions(+), 1 deletion(-)

@@ -59,3 +59,14 @@ asmlinkage int sys_fadvise64_64_wrapper(
 				 (u64)len0 << 32 | len1, advice);
 #endif
 }
+
+/*
+ * swap the arguments the way that libc wants them instead of
+ * moving flags ahead of the 64-bit nbytes argument
+ */
+SYSCALL_DEFINE6(sh_sync_file_range6, int, fd, SC_ARG64(offset),
+                SC_ARG64(nbytes), unsigned int, flags)
+{
+        return ksys_sync_file_range(fd, SC_VAL64(loff_t, offset),
+                                    SC_VAL64(loff_t, nbytes), flags);
+}

@@ -321,7 +321,7 @@
 311	common	set_robust_list			sys_set_robust_list
 312	common	get_robust_list			sys_get_robust_list
 313	common	splice				sys_splice
-314	common	sync_file_range			sys_sync_file_range
+314	common	sync_file_range			sys_sh_sync_file_range6
 315	common	tee				sys_tee
 316	common	vmsplice			sys_vmsplice
 317	common	move_pages			sys_move_pages

@@ -395,6 +395,7 @@
 385	common	pkey_alloc			sys_pkey_alloc
 386	common	pkey_free			sys_pkey_free
 387	common	rseq				sys_rseq
+388	common	sync_file_range2		sys_sync_file_range2
 # room for arch specific syscalls
 393	common	semget				sys_semget
 394	common	semctl				sys_semctl

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Naohiro Aota <[email protected]>

commit b9fd2affe4aa99a4ca14ee87e1f38fea22ece52a upstream.

When creating a new block group, it calls btrfs_add_new_free_space() to add the entire block group range into the free space accounting. __btrfs_add_free_space_zoned() checks if size == block_group->length to detect the initial free space adding, and proceed that case properly.

However, if the zone_capacity == zone_size and the over-write speed is fast enough, the entire zone can be over-written within one transaction. That confuses __btrfs_add_free_space_zoned() to handle it as an initial free space accounting. As a result, that block group becomes a strange state: 0 used bytes, 0 zone_unusable bytes, but alloc_offset == zone_capacity (no allocation anymore).

The initial free space accounting can properly be checked by checking alloc_offset too.

---
 fs/btrfs/free-space-cache.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -2697,7 +2697,7 @@ static int __btrfs_add_free_space_zoned(
 	u64 offset = bytenr - block_group->start;
 	u64 to_free, to_unusable;
 	int bg_reclaim_threshold = 0;
-	bool initial = (size == block_group->length);
+	bool initial = ((size == block_group->length) && (block_group->alloc_offset == 0));
 	u64 reclaimable_unusable;
 
 	WARN_ON(!initial && offset + size > block_group->zone_capacity);

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Arnd Bergmann <[email protected]>

commit 3339b99ef6fe38dac43b534cba3a8a0e29fb2eff upstream.

Both of these architectures require u64 function arguments to be passed in even/odd pairs of registers or stack slots, which in case of sync_file_range would result in a seven-argument system call that is not currently possible. The system call is therefore incompatible with all existing binaries.

While it would be possible to implement support for seven arguments like on mips, it seems better to use a six-argument version, either with the normal argument order but misaligned as on most architectures or with the reordered sync_file_range2() calling conventions as on arm and powerpc.

---
 arch/csky/include/uapi/asm/unistd.h    |    1 +
 arch/hexagon/include/uapi/asm/unistd.h |    1 +
 2 files changed, 2 insertions(+)

@@ -6,6 +6,7 @@
 #define __ARCH_WANT_SYS_CLONE3
 #define __ARCH_WANT_SET_GET_RLIMIT
 #define __ARCH_WANT_TIME32_SYSCALLS
+#define __ARCH_WANT_SYNC_FILE_RANGE2
 #include <asm-generic/unistd.h>
 
 #define __NR_set_thread_area	(__NR_arch_specific_syscall + 0)

@@ -36,5 +36,6 @@
 #define __ARCH_WANT_SYS_VFORK
 #define __ARCH_WANT_SYS_FORK
 #define __ARCH_WANT_TIME32_SYSCALLS
+#define __ARCH_WANT_SYNC_FILE_RANGE2
 
 #include <asm-generic/unistd.h>

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Arnd Bergmann <[email protected]>

commit 896842284c6ccba25ec9d78b7b6e62cdd507c083 upstream.

fadvise64_64() has two 64-bit arguments at the wrong alignment for hexagon, which turns them into a 7-argument syscall that is not supported by Linux.

The downstream musl port for hexagon actually asks for a 6-argument version the same way we do it on arm, csky, powerpc, so make the kernel do it the same way to avoid having to change both.

---
 arch/hexagon/include/asm/syscalls.h |    6 ++++++
 arch/hexagon/kernel/syscalltab.c    |    7 +++++++
 2 files changed, 13 insertions(+)
 create mode 100644 arch/hexagon/include/asm/syscalls.h

@@ -0,0 +1,6 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#include <asm-generic/syscalls.h>
+
+asmlinkage long sys_hexagon_fadvise64_64(int fd, int advice,
+	                                  u32 a2, u32 a3, u32 a4, u32 a5);

@@ -14,6 +14,13 @@
 #undef __SYSCALL
 #define __SYSCALL(nr, call) [nr] = (call),
 
+SYSCALL_DEFINE6(hexagon_fadvise64_64, int, fd, int, advice,
+		SC_ARG64(offset), SC_ARG64(len))
+{
+	return ksys_fadvise64_64(fd, SC_VAL64(loff_t, offset), SC_VAL64(loff_t, len), advice);
+}
+#define sys_fadvise64_64 sys_hexagon_fadvise64_64
+
 void *sys_call_table[__NR_syscalls] = {
 #include <asm/unistd.h>
 };

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Jann Horn <[email protected]>

commit 4f2a129b33a2054e62273edd5a051c34c08d96e9 upstream.

<[email protected]>, Maxime Ripard <[email protected]>, Thomas Zimmermann <[email protected]>

filp->pid is supposed to be a refcounted pointer; however, before this patch, drm_file_update_pid() only increments the refcount of a struct pid after storing a pointer to it in filp->pid and dropping the dev->filelist_mutex, making the following race possible:

process A process B
========= =========

                        begin drm_file_update_pid
                        mutex_lock(&dev->filelist_mutex)
                        rcu_replace_pointer(filp->pid, <pid B>, 1)
                        mutex_unlock(&dev->filelist_mutex)

begin drm_file_update_pid
mutex_lock(&dev->filelist_mutex)
rcu_replace_pointer(filp->pid, <pid A>, 1)
mutex_unlock(&dev->filelist_mutex)
get_pid(<pid A>)
synchronize_rcu()
put_pid(<pid B>) *** pid B reaches refcount 0 and is freed here ***

                        get_pid(<pid B>)   *** UAF ***
                        synchronize_rcu()
                        put_pid(<pid A>)

As far as I know, this race can only occur with CONFIG_PREEMPT_RCU=y because it requires RCU to detect a quiescent state in code that is not explicitly calling into the scheduler.

This race leads to use-after-free of a "struct pid".
It is probably somewhat hard to hit because process A has to pass through a synchronize_rcu() operation while process B is between mutex_unlock() and get_pid().

Fix it by ensuring that by the time a pointer to the current task's pid is stored in the file, an extra reference to the pid has been taken.

This fix also removes the condition for synchronize_rcu(); I think that optimization is unnecessary complexity, since in that case we would usually have bailed out on the lockless check above.

---
 drivers/gpu/drm/drm_file.c |    8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

@@ -469,14 +469,12 @@ void drm_file_update_pid(struct drm_file
 
 	dev = filp->minor->dev;
 	mutex_lock(&dev->filelist_mutex);
+	get_pid(pid);
 	old = rcu_replace_pointer(filp->pid, pid, 1);
 	mutex_unlock(&dev->filelist_mutex);
 
-	if (pid != old) {
-		get_pid(pid);
-		synchronize_rcu();
-		put_pid(old);
-	}
+	synchronize_rcu();
+	put_pid(old);
 }
 
 /**

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Ma Ke <[email protected]>

commit 66edf3fb331b6c55439b10f9862987b0916b3726 upstream.

In nv17_tv_get_ld_modes(), the return value of drm_mode_duplicate() is assigned to mode, which will lead to a possible NULL pointer dereference on failure of drm_mode_duplicate(). Add a check to avoid npd.

---
 drivers/gpu/drm/nouveau/dispnv04/tvnv17.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -209,6 +209,8 @@ static int nv17_tv_get_ld_modes(struct d
 		struct drm_display_mode *mode;
 
 		mode = drm_mode_duplicate(encoder->dev, tv_mode);
+		if (!mode)
+			continue;
 
 		mode->clock = tv_norm->tv_enc_mode.vrefresh *
 			mode->htotal / 1000 *

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Thomas Zimmermann <[email protected]>

commit d92a7580392ad4681b1d4f9275d00b95375ebe01 upstream.

Only export struct fb_info.fix.smem_start if that is required by the user and the memory does not come from vmalloc().

Setting struct fb_info.fix.smem_start breaks systems where DMA memory is backed by vmalloc address space. An example error is shown below.

[    3.536043] ------------[ cut here ]------------
[    3.540716] virt_to_phys used for non-linear address: 000000007fc4f540 (0xffff800086001000)
[    3.552628] WARNING: CPU: 4 PID: 61 at arch/arm64/mm/physaddr.c:12 __virt_to_phys+0x68/0x98
[    3.565455] Modules linked in:
[    3.568525] CPU: 4 PID: 61 Comm: kworker/u12:5 Not tainted 6.6.23-06226-g4986cc3e1b75-dirty #250
[    3.577310] Hardware name: NXP i.MX95 19X19 board (DT)
[    3.582452] Workqueue: events_unbound deferred_probe_work_func
[    3.588291] pstate: 60400009 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[    3.595233] pc : __virt_to_phys+0x68/0x98
[    3.599246] lr : __virt_to_phys+0x68/0x98
[    3.603276] sp : ffff800083603990
[    3.677939] Call trace:
[    3.680393]  __virt_to_phys+0x68/0x98
[    3.684067]  drm_fbdev_dma_helper_fb_probe+0x138/0x238
[    3.689214]  __drm_fb_helper_initial_config_and_unlock+0x2b0/0x4c0
[    3.695385]  drm_fb_helper_initial_config+0x4c/0x68
[    3.700264]  drm_fbdev_dma_client_hotplug+0x8c/0xe0
[    3.705161]  drm_client_register+0x60/0xb0
[    3.709269]  drm_fbdev_dma_setup+0x94/0x148

Additionally, DMA memory is assumed to by contiguous in physical address space, which is not guaranteed by vmalloc().

Resolve this by checking the module flag drm_leak_fbdev_smem when DRM allocated the instance of struct fb_info. Fbdev-dma then only sets smem_start only if required (via FBINFO_HIDE_SMEM_START). Also guarantee that the framebuffer is not located in vmalloc address space.

---
 drivers/gpu/drm/drm_fb_helper.c |    6 +++---
 drivers/gpu/drm/drm_fbdev_dma.c |    5 ++++-
 2 files changed, 7 insertions(+), 4 deletions(-)

@@ -524,6 +524,9 @@ struct fb_info *drm_fb_helper_alloc_info
 	if (!info)
 		return ERR_PTR(-ENOMEM);
 
+	if (!drm_leak_fbdev_smem)
+		info->flags |= FBINFO_HIDE_SMEM_START;
+
 	ret = fb_alloc_cmap(&info->cmap, 256, 0);
 	if (ret)
 		goto err_release;

@@ -1860,9 +1863,6 @@ __drm_fb_helper_initial_config_and_unloc
 	info = fb_helper->info;
 	info->var.pixclock = 0;
 
-	if (!drm_leak_fbdev_smem)
-		info->flags |= FBINFO_HIDE_SMEM_START;
-
 	/* Need to drop locks to avoid recursive deadlock in
 	 * register_framebuffer. This is ok because the only thing left to do is
 	 * register the fbdev emulation instance in kernel_fb_helper_list. */

@@ -130,7 +130,10 @@ static int drm_fbdev_dma_helper_fb_probe
 		info->flags |= FBINFO_READS_FAST; /* signal caching */
 	info->screen_size = sizes->surface_height * fb->pitches[0];
 	info->screen_buffer = map.vaddr;
-	info->fix.smem_start = page_to_phys(virt_to_page(info->screen_buffer));
+	if (!(info->flags & FBINFO_HIDE_SMEM_START)) {
+		if (!drm_WARN_ON(dev, is_vmalloc_addr(info->screen_buffer)))
+			info->fix.smem_start = page_to_phys(virt_to_page(info->screen_buffer));
+	}
 	info->fix.smem_len = info->screen_size;
 
 	return 0;

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Julia Zhang <[email protected]>

commit bcfa48ff785bd121316592b131ff6531e3e696bb upstream.

Instead of using state->fb->obj[0] directly, get object from framebuffer by calling drm_gem_fb_get_obj() and return error code when object is null to avoid using null object of framebuffer.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c |   18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)

@@ -3,6 +3,7 @@
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_edid.h>
 #include <drm/drm_simple_kms_helper.h>
+#include <drm/drm_gem_framebuffer_helper.h>
 #include <drm/drm_vblank.h>
 
 #include "amdgpu.h"

@@ -314,7 +315,13 @@ static int amdgpu_vkms_prepare_fb(struct
 		return 0;
 	}
 	afb = to_amdgpu_framebuffer(new_state->fb);
-	obj = new_state->fb->obj[0];
+
+	obj = drm_gem_fb_get_obj(new_state->fb, 0);
+	if (!obj) {
+		DRM_ERROR("Failed to get obj from framebuffer\n");
+		return -EINVAL;
+	}
+
 	rbo = gem_to_amdgpu_bo(obj);
 	adev = amdgpu_ttm_adev(rbo->tbo.bdev);
 

@@ -368,12 +375,19 @@ static void amdgpu_vkms_cleanup_fb(struc
 				   struct drm_plane_state *old_state)
 {
 	struct amdgpu_bo *rbo;
+	struct drm_gem_object *obj;
 	int r;
 
 	if (!old_state->fb)
 		return;
 
-	rbo = gem_to_amdgpu_bo(old_state->fb->obj[0]);
+	obj = drm_gem_fb_get_obj(old_state->fb, 0);
+	if (!obj) {
+		DRM_ERROR("Failed to get obj from framebuffer\n");
+		return;
+	}
+
+	rbo = gem_to_amdgpu_bo(obj);
 	r = amdgpu_bo_reserve(rbo, false);
 	if (unlikely(r)) {
 		DRM_ERROR("failed to reserve rbo before unpin\n");

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Janusz Krzysztofik <[email protected]>

commit 996c3412a06578e9d779a16b9e79ace18125ab50 upstream.

CI has been sporadically reporting the following issue triggered by igt@i915_selftest@live@hangcheck on ADL-P and similar machines:

<6> [414.049203] i915: Running intel_hangcheck_live_selftests/igt_reset_evict_fence
...
<6> [414.068804] i915 0000:00:02.0: [drm] GT0: GUC: submission enabled
<6> [414.068812] i915 0000:00:02.0: [drm] GT0: GUC: SLPC enabled
<3> [414.070354] Unable to pin Y-tiled fence; err:-4
<3> [414.071282] i915_vma_revoke_fence:301 GEM_BUG_ON(!i915_active_is_idle(&fence->active))
...

<4>[  609.603992] ------------[ cut here ]------------
<2>[  609.603995] kernel BUG at drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c:301!
<4>[  609.604003] invalid opcode: 0000 [#1] PREEMPT SMP NOPTI
<4>[  609.604006] CPU: 0 PID: 268 Comm: kworker/u64:3 Tainted: G     U  W          6.9.0-CI_DRM_14785-g1ba62f8cea9c+ #1
<4>[  609.604008] Hardware name: Intel Corporation Alder Lake Client Platform/AlderLake-P DDR4 RVP, BIOS RPLPFWI1.R00.4035.A00.2301200723 01/20/2023
<4>[  609.604010] Workqueue: i915 __i915_gem_free_work [i915]
<4>[  609.604149] RIP: 0010:i915_vma_revoke_fence+0x187/0x1f0 [i915]
...
<4>[  609.604271] Call Trace:
<4>[  609.604273]  <TASK>
...
<4>[  609.604716]  __i915_vma_evict+0x2e9/0x550 [i915]
<4>[  609.604852]  __i915_vma_unbind+0x7c/0x160 [i915]
<4>[  609.604977]  force_unbind+0x24/0xa0 [i915]
<4>[  609.605098]  i915_vma_destroy+0x2f/0xa0 [i915]
<4>[  609.605210]  __i915_gem_object_pages_fini+0x51/0x2f0 [i915]
<4>[  609.605330]  __i915_gem_free_objects.isra.0+0x6a/0xc0 [i915]
<4>[  609.605440]  process_scheduled_works+0x351/0x690

...

In the past, there were similar failures reported by CI from other IGT tests, observed on other platforms.

Before commit 63baf4f3d587 ("drm/i915/gt: Only wait for GPU activity before unbinding a GGTT fence"), i915_vma_revoke_fence() was waiting for
idleness of vma->active via fence_update(). That commit introduced
vma->fence->active in order for the fence_update() to be able to wait selectively on that one instead of vma->active since only idleness of fence registers was needed. But then, another commit 0d86ee35097a ("drm/i915/gt: Make fence revocation unequivocal") replaced the call to fence_update() in i915_vma_revoke_fence() with only fence_write(), and also added that GEM_BUG_ON(!i915_active_is_idle(&fence->active)) in front. No justification was provided on why we might then expect idleness of vma->fence->active without first waiting on it.

The issue can be potentially caused by a race among revocation of fence registers on one side and sequential execution of signal callbacks invoked on completion of a request that was using them on the other, still processed in parallel to revocation of those fence registers. Fix it by waiting for idleness of vma->fence->active in i915_vma_revoke_fence().

---
 drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c |    1 +
 1 file changed, 1 insertion(+)

@@ -298,6 +298,7 @@ void i915_vma_revoke_fence(struct i915_v
 		return;
 
 	GEM_BUG_ON(fence->vma != vma);
+	i915_active_wait(&fence->active);
 	GEM_BUG_ON(!i915_active_is_idle(&fence->active));
 	GEM_BUG_ON(atomic_read(&fence->pin_count));
 

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Ma Ke <[email protected]>

commit 6d411c8ccc0137a612e0044489030a194ff5c843 upstream.

In nv17_tv_get_hd_modes(), the return value of drm_mode_duplicate() is assigned to mode, which will lead to a possible NULL pointer dereference on failure of drm_mode_duplicate(). The same applies to drm_cvt_mode(). Add a check to avoid null pointer dereference.

---
 drivers/gpu/drm/nouveau/dispnv04/tvnv17.c |    4 ++++
 1 file changed, 4 insertions(+)

@@ -260,6 +260,8 @@ static int nv17_tv_get_hd_modes(struct d
 		if (modes[i].hdisplay == output_mode->hdisplay &&
 		    modes[i].vdisplay == output_mode->vdisplay) {
 			mode = drm_mode_duplicate(encoder->dev, output_mode);
+			if (!mode)
+				continue;
 			mode->type |= DRM_MODE_TYPE_PREFERRED;
 
 		} else {

@@ -267,6 +269,8 @@ static int nv17_tv_get_hd_modes(struct d
 					    modes[i].vdisplay, 60, false,
 					    (output_mode->flags &
 					     DRM_MODE_FLAG_INTERLACE), false);
+			if (!mode)
+				continue;
 		}
 
 		/* CVT modes are sometimes unsuitable... */

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Michael Strauss <[email protected]>

commit 2ec6c7f802332d1eff16f03e7c757f1543ee1183 upstream.

[WHY]
New register field added in DP2.1 SCR, needed for auxless ALPM

[HOW]
Echo value read from 0xF0007 back to sink

---
 drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c |   10 +++++++++-
 drivers/gpu/drm/amd/display/include/dpcd_defs.h                    |    5 +++++
 2 files changed, 14 insertions(+), 1 deletion(-)

@@ -1590,9 +1590,17 @@ static bool retrieve_link_cap(struct dc_
 			return false;
 	}
 
-	if (dp_is_lttpr_present(link))
+	if (dp_is_lttpr_present(link)) {
 		configure_lttpr_mode_transparent(link);
 
+		// Echo TOTAL_LTTPR_CNT back downstream
+		core_link_write_dpcd(
+				link,
+				DP_TOTAL_LTTPR_CNT,
+				&link->dpcd_caps.lttpr_caps.phy_repeater_cnt,
+				sizeof(link->dpcd_caps.lttpr_caps.phy_repeater_cnt));
+	}
+
 	/* Read DP tunneling information. */
 	status = dpcd_get_tunneling_device_data(link);
 

@@ -177,4 +177,9 @@ enum dpcd_psr_sink_states {
 #define DP_SINK_PR_PIXEL_DEVIATION_PER_LINE     0x379
 #define DP_SINK_PR_MAX_NUMBER_OF_DEVIATION_LINE 0x37A
 
+/* Remove once drm_dp_helper.h is updated upstream */
+#ifndef DP_TOTAL_LTTPR_CNT
+#define DP_TOTAL_LTTPR_CNT                                  0xF000A /* 2.1 */
+#endif
+
 #endif /* __DAL_DPCD_DEFS_H__ */

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Alex Deucher <[email protected]>

commit f6f49dda49db72e7a0b4ca32c77391d5ff5ce232 upstream.

v3.x changed the how vram width was encoded. The previous implementation actually worked correctly for most boards. Fix the implementation to work correctly everywhere.

This fixes the vram width reported in the kernel log on some boards.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -399,7 +399,7 @@ amdgpu_atomfirmware_get_vram_info(struct
 					mem_channel_number = vram_info->v30.channel_num;
 					mem_channel_width = vram_info->v30.channel_width;
 					if (vram_width)
-						*vram_width = mem_channel_number * (1 << mem_channel_width);
+						*vram_width = mem_channel_number * 16;
 					break;
 				default:
 					return -EINVAL;

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Jens Axboe <[email protected]>

commit dbcabac138fdfc730ba458ed2199ff1f29a271fc upstream.

Before SQPOLL was transitioned to managing its own task_work, the core used TWA_SIGNAL_NO_IPI to ensure that task_work was processed. If not, we can't be sure that all task_work is processed at SQPOLL thread exit time.

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

@@ -1374,8 +1374,8 @@ static void io_req_normal_work_add(struc
 	if (ctx->flags & IORING_SETUP_SQPOLL) {
 		struct io_sq_data *sqd = ctx->sq_data;
 
-		if (wq_has_sleeper(&sqd->wait))
-			wake_up(&sqd->wait);
+		if (sqd->thread)
+			__set_notify_signal(sqd->thread);
 		return;
 	}
 

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Sven Eckelmann <[email protected]>

commit 537a350d14321c8cca5efbf0a33a404fec3a9f9e upstream.

The internal handling of VLAN IDs in batman-adv is only specified for following encodings:

* VLAN is used
- bit 15 is 1
- bit 11 - bit 0 is the VLAN ID (0-4095)
- remaining bits are 0
* No VLAN is used
- bit 15 is 0
- remaining bits are 0

batman-adv was only preparing new translation table entries (based on its soft interface information) using this encoding format. But the receive path was never checking if entries in the roam or TT TVLVs were also following this encoding.

It was therefore possible to create more than the expected maximum of 4096
+ 1 entries in the originator VLAN list. Simply by setting the "remaining bits" to "random" values in corresponding TVLV.

---
 net/batman-adv/originator.c |   27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

@@ -12,6 +12,7 @@
 #include <linux/errno.h>
 #include <linux/etherdevice.h>
 #include <linux/gfp.h>
+#include <linux/if_vlan.h>
 #include <linux/jiffies.h>
 #include <linux/kref.h>
 #include <linux/list.h>

@@ -132,6 +133,29 @@ batadv_orig_node_vlan_get(struct batadv_
 }
 
 /**
+ * batadv_vlan_id_valid() - check if vlan id is in valid batman-adv encoding
+ * @vid: the VLAN identifier
+ *
+ * Return: true when either no vlan is set or if VLAN is in correct range,
+ *  false otherwise
+ */
+static bool batadv_vlan_id_valid(unsigned short vid)
+{
+	unsigned short non_vlan = vid & ~(BATADV_VLAN_HAS_TAG | VLAN_VID_MASK);
+
+	if (vid == 0)
+		return true;
+
+	if (!(vid & BATADV_VLAN_HAS_TAG))
+		return false;
+
+	if (non_vlan)
+		return false;
+
+	return true;
+}
+
+/**
  * batadv_orig_node_vlan_new() - search and possibly create an orig_node_vlan
  *  object
  * @orig_node: the originator serving the VLAN

@@ -149,6 +173,9 @@ batadv_orig_node_vlan_new(struct batadv_
 {
 	struct batadv_orig_node_vlan *vlan;
 
+	if (!batadv_vlan_id_valid(vid))
+		return NULL;
+
 	spin_lock_bh(&orig_node->vlan_list_lock);
 
 	/* first look if an object for this vid already exists */

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Vitor Soares <[email protected]>

commit d8fb63e46c884c898a38f061c2330f7729e75510 upstream.

When the mcp251xfd_start_xmit() function fails, the driver stops processing messages, and the interrupt routine does not return, running indefinitely even after killing the running application.

Error messages:

[  441.298819] mcp251xfd spi2.0 can0: ERROR in mcp251xfd_start_xmit: -16
[  441.306498] mcp251xfd spi2.0 can0: Transmit Event FIFO buffer not empty. (seq=0x000017c7, tef_tail=0x000017cf, tef_head=0x000017d0, tx_head=0x000017d3).

... and repeat forever.

The issue can be triggered when multiple devices share the same SPI interface. And there is concurrent access to the bus.

The problem occurs because tx_ring->head increments even if mcp251xfd_start_xmit() fails. Consequently, the driver skips one TX package while still expecting a response in mcp251xfd_handle_tefif_one().

Resolve the issue by starting a workqueue to write the tx obj synchronously if err = -EBUSY. In case of another error, decrement tx_ring->head, remove skb from the echo stack, and drop the message.

---
 drivers/net/can/spi/mcp251xfd/mcp251xfd-core.c |   14 +++++-
 drivers/net/can/spi/mcp251xfd/mcp251xfd-tx.c   |   55 +++++++++++++++++++++----
 drivers/net/can/spi/mcp251xfd/mcp251xfd.h      |    5 ++
 3 files changed, 65 insertions(+), 9 deletions(-)

@@ -1618,11 +1618,20 @@ static int mcp251xfd_open(struct net_dev
 	clear_bit(MCP251XFD_FLAGS_DOWN, priv->flags);
 	can_rx_offload_enable(&priv->offload);
 
+	priv->wq = alloc_ordered_workqueue("%s-mcp251xfd_wq",
+					   WQ_FREEZABLE | WQ_MEM_RECLAIM,
+					   dev_name(&spi->dev));
+	if (!priv->wq) {
+		err = -ENOMEM;
+		goto out_can_rx_offload_disable;
+	}
+	INIT_WORK(&priv->tx_work, mcp251xfd_tx_obj_write_sync);
+
 	err = request_threaded_irq(spi->irq, NULL, mcp251xfd_irq,
 				   IRQF_SHARED | IRQF_ONESHOT,
 				   dev_name(&spi->dev), priv);
 	if (err)
-		goto out_can_rx_offload_disable;
+		goto out_destroy_workqueue;
 
 	err = mcp251xfd_chip_interrupts_enable(priv);
 	if (err)

@@ -1634,6 +1643,8 @@ static int mcp251xfd_open(struct net_dev
 
  out_free_irq:
 	free_irq(spi->irq, priv);
+ out_destroy_workqueue:
+	destroy_workqueue(priv->wq);
  out_can_rx_offload_disable:
 	can_rx_offload_disable(&priv->offload);
 	set_bit(MCP251XFD_FLAGS_DOWN, priv->flags);

@@ -1661,6 +1672,7 @@ static int mcp251xfd_stop(struct net_dev
 	hrtimer_cancel(&priv->tx_irq_timer);
 	mcp251xfd_chip_interrupts_disable(priv);
 	free_irq(ndev->irq, priv);
+	destroy_workqueue(priv->wq);
 	can_rx_offload_disable(&priv->offload);
 	mcp251xfd_timestamp_stop(priv);
 	mcp251xfd_chip_stop(priv, CAN_STATE_STOPPED);

@@ -633,6 +633,10 @@ struct mcp251xfd_priv {
 	struct mcp251xfd_rx_ring *rx[MCP251XFD_FIFO_RX_NUM];
 	struct mcp251xfd_tx_ring tx[MCP251XFD_FIFO_TX_NUM];
 
+	struct workqueue_struct *wq;
+	struct work_struct tx_work;
+	struct mcp251xfd_tx_obj *tx_work_obj;
+
 	DECLARE_BITMAP(flags, __MCP251XFD_FLAGS_SIZE__);
 
 	u8 rx_ring_num;

@@ -952,6 +956,7 @@ void mcp251xfd_skb_set_timestamp(const s
 void mcp251xfd_timestamp_init(struct mcp251xfd_priv *priv);
 void mcp251xfd_timestamp_stop(struct mcp251xfd_priv *priv);
 
+void mcp251xfd_tx_obj_write_sync(struct work_struct *work);
 netdev_tx_t mcp251xfd_start_xmit(struct sk_buff *skb,
 				 struct net_device *ndev);
 

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Niklas Cassel <[email protected]>

commit eeb25a09c5e0805d92e4ebd12c4b0ad0df1b0295 upstream.

.probe() (ahci_init_one()) calls sysfs_add_file_to_group(), however, if probe() fails after this call, we currently never call sysfs_remove_file_from_group().

(The sysfs_remove_file_from_group() call in .remove() (ahci_remove_one()) does not help, as .remove() is not called on .probe() error.)

Thus, if probe() fails after the sysfs_add_file_to_group() call, the next time we insmod the module we will get:

sysfs: cannot create duplicate filename '/devices/pci0000:00/0000:00:04.0/remapped_nvme'

CPU: 11 PID: 954 Comm: modprobe Not tainted 6.10.0-rc5 #43
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-2.fc40 04/01/2014
Call Trace:
 <TASK>
 dump_stack_lvl+0x5d/0x80
 sysfs_warn_dup.cold+0x17/0x23
 sysfs_add_file_mode_ns+0x11a/0x130
 sysfs_add_file_to_group+0x7e/0xc0
 ahci_init_one+0x31f/0xd40 [ahci]

---
 drivers/ata/ahci.c |   17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

@@ -1975,8 +1975,10 @@ static int ahci_init_one(struct pci_dev
 	n_ports = max(ahci_nr_ports(hpriv->cap), fls(hpriv->port_map));
 
 	host = ata_host_alloc_pinfo(&pdev->dev, ppi, n_ports);
-	if (!host)
-		return -ENOMEM;
+	if (!host) {
+		rc = -ENOMEM;
+		goto err_rm_sysfs_file;
+	}
 	host->private_data = hpriv;
 
 	if (ahci_init_msi(pdev, n_ports, hpriv) < 0) {

@@ -2031,11 +2033,11 @@ static int ahci_init_one(struct pci_dev
 	/* initialize adapter */
 	rc = ahci_configure_dma_masks(pdev, hpriv);
 	if (rc)
-		return rc;
+		goto err_rm_sysfs_file;
 
 	rc = ahci_pci_reset_controller(host);
 	if (rc)
-		return rc;
+		goto err_rm_sysfs_file;
 
 	ahci_pci_init_controller(host);
 	ahci_pci_print_info(host);

@@ -2044,10 +2046,15 @@ static int ahci_init_one(struct pci_dev
 
 	rc = ahci_host_activate(host, &ahci_sht);
 	if (rc)
-		return rc;
+		goto err_rm_sysfs_file;
 
 	pm_runtime_put_noidle(&pdev->dev);
 	return 0;
+
+err_rm_sysfs_file:
+	sysfs_remove_file_from_group(&pdev->dev.kobj,
+				     &dev_attr_remapped_nvme.attr, NULL);
+	return rc;
 }
 
 static void ahci_shutdown_one(struct pci_dev *pdev)

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Niklas Cassel <[email protected]>

commit 1066fe825987da007669d7c25306b4dbb50bd7dd upstream.

We got another report that CT1000BX500SSD1 does not work with LPM.

If you look in libata-core.c, we have six different Crucial devices that are marked with ATA_HORKAGE_NOLPM. This model would have been the seventh. (This quirk is used on Crucial models starting with both CT* and Crucial_CT*)

It is obvious that this vendor does not have a great history of supporting LPM properly, therefore, add the ATA_HORKAGE_NOLPM quirk for all Crucial BX SSD1 models.

---
 drivers/ata/libata-core.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

@@ -4181,8 +4181,7 @@ static const struct ata_blacklist_entry
 	{ "PIONEER BD-RW   BDR-205",	NULL,	ATA_HORKAGE_NOLPM },
 
 	/* Crucial devices with broken LPM support */
-	{ "CT500BX100SSD1",		NULL,	ATA_HORKAGE_NOLPM },
-	{ "CT240BX500SSD1",		NULL,	ATA_HORKAGE_NOLPM },
+	{ "CT*0BX*00SSD1",		NULL,	ATA_HORKAGE_NOLPM },
 
 	/* 512GB MX100 with MU01 firmware has both queued TRIM and LPM issues */
 	{ "Crucial_CT512MX100*",	"MU01",	ATA_HORKAGE_NO_NCQ_TRIM |

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Niklas Cassel <[email protected]>

commit ab9e0c529eb7cafebdd31fe1644524e80a48b05d upstream.

If e.g. the ata_port_alloc() call in ata_host_alloc() fails, we will jump to the err_out label, which will call devres_release_group(). devres_release_group() will trigger a call to ata_host_release(). ata_host_release() calls kfree(host), so executing the kfree(host) in ata_host_alloc() will lead to a double free:

kernel BUG at mm/slub.c:553!

Oops: invalid opcode: 0000 [#1] PREEMPT SMP NOPTI
CPU: 11 PID: 599 Comm: (udev-worker) Not tainted 6.10.0-rc5 #47
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-2.fc40 04/01/2014
RIP: 0010:kfree+0x2cf/0x2f0
Code: 5d 41 5e 41 5f 5d e9 80 d6 ff ff 4d 89 f1 41 b8 01 00 00 00 48 89 d9 48 89 da
RSP: 0018:ffffc90000f377f0 EFLAGS: 00010246
RAX: ffff888112b1f2c0 RBX: ffff888112b1f2c0 RCX: ffff888112b1f320
RDX: 000000000000400b RSI: ffffffffc02c9de5 RDI: ffff888112b1f2c0
RBP: ffffc90000f37830 R08: 0000000000000000 R09: 0000000000000000
R10: ffffc90000f37610 R11: 617461203a736b6e R12: ffffea00044ac780
R13: ffff888100046400 R14: ffffffffc02c9de5 R15: 0000000000000006
FS:  00007f2f1cabe980(0000) GS:ffff88813b380000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f2f1c3acf75 CR3: 0000000111724000 CR4: 0000000000750ef0
PKRU: 55555554
Call Trace:
 <TASK>
 ? __die_body.cold+0x19/0x27
 ? die+0x2e/0x50
 ? do_trap+0xca/0x110
 ? do_error_trap+0x6a/0x90
 ? kfree+0x2cf/0x2f0
 ? exc_invalid_op+0x50/0x70
 ? kfree+0x2cf/0x2f0
 ? asm_exc_invalid_op+0x1a/0x20
 ? ata_host_alloc+0xf5/0x120 [libata]
 ? ata_host_alloc+0xf5/0x120 [libata]
 ? kfree+0x2cf/0x2f0
 ata_host_alloc+0xf5/0x120 [libata]
 ata_host_alloc_pinfo+0x14/0xa0 [libata]
 ahci_init_one+0x6c9/0xd20 [ahci]

Ensure that we will not call kfree(host) twice, by performing the kfree() only if the devres_open_group() call failed.

---
 drivers/ata/libata-core.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

@@ -5621,8 +5621,10 @@ struct ata_host *ata_host_alloc(struct d
 	if (!host)
 		return NULL;
 
-	if (!devres_open_group(dev, NULL, GFP_KERNEL))
-		goto err_free;
+	if (!devres_open_group(dev, NULL, GFP_KERNEL)) {
+		kfree(host);
+		return NULL;
+	}
 
 	dr = devres_alloc(ata_devres_release, 0, GFP_KERNEL);
 	if (!dr)

@@ -5654,8 +5656,6 @@ struct ata_host *ata_host_alloc(struct d
 
  err_out:
 	devres_release_group(dev, NULL);
- err_free:
-	kfree(host);
 	return NULL;
 }
 EXPORT_SYMBOL_GPL(ata_host_alloc);

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Arnd Bergmann <[email protected]>

commit 4b8e88e563b5f666446d002ad0dc1e6e8e7102b0 upstream.

The old ftruncate() syscall, using the 32-bit off_t misses a sign extension when called in compat mode on 64-bit architectures. As a result, passing a negative length accidentally succeeds in truncating to file size between 2GiB and 4GiB.

Changing the type of the compat syscall to the signed compat_off_t changes the behavior so it instead returns -EINVAL.

The native entry point, the truncate() syscall and the corresponding loff_t based variants are all correct already and do not suffer from this mistake.

---
 fs/open.c                |    4 ++--
 include/linux/compat.h   |    2 +-
 include/linux/syscalls.h |    2 +-
 3 files changed, 4 insertions(+), 4 deletions(-)

@@ -202,13 +202,13 @@ long do_sys_ftruncate(unsigned int fd, l
 	return error;
 }
 
-SYSCALL_DEFINE2(ftruncate, unsigned int, fd, unsigned long, length)
+SYSCALL_DEFINE2(ftruncate, unsigned int, fd, off_t, length)
 {
 	return do_sys_ftruncate(fd, length, 1);
 }
 
 #ifdef CONFIG_COMPAT
-COMPAT_SYSCALL_DEFINE2(ftruncate, unsigned int, fd, compat_ulong_t, length)
+COMPAT_SYSCALL_DEFINE2(ftruncate, unsigned int, fd, compat_off_t, length)
 {
 	return do_sys_ftruncate(fd, length, 1);
 }

@@ -608,7 +608,7 @@ asmlinkage long compat_sys_fstatfs(unsig
 asmlinkage long compat_sys_fstatfs64(unsigned int fd, compat_size_t sz,
 				     struct compat_statfs64 __user *buf);
 asmlinkage long compat_sys_truncate(const char __user *, compat_off_t);
-asmlinkage long compat_sys_ftruncate(unsigned int, compat_ulong_t);
+asmlinkage long compat_sys_ftruncate(unsigned int, compat_off_t);
 /* No generic prototype for truncate64, ftruncate64, fallocate */
 asmlinkage long compat_sys_openat(int dfd, const char __user *filename,
 				  int flags, umode_t mode);

@@ -418,7 +418,7 @@ asmlinkage long sys_listmount(const stru
 			      u64 __user *mnt_ids, size_t nr_mnt_ids,
 			      unsigned int flags);
 asmlinkage long sys_truncate(const char __user *path, long length);
-asmlinkage long sys_ftruncate(unsigned int fd, unsigned long length);
+asmlinkage long sys_ftruncate(unsigned int fd, off_t length);
 #if BITS_PER_LONG == 32
 asmlinkage long sys_truncate64(const char __user *path, loff_t length);
 asmlinkage long sys_ftruncate64(unsigned int fd, loff_t length);

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Arnd Bergmann <[email protected]>

commit d3882564a77c21eb746ba5364f3fa89b88de3d61 upstream.

Using sys_io_pgetevents() as the entry point for compat mode tasks works almost correctly, but misses the sign extension for the min_nr and nr arguments.

This was addressed on parisc by switching to compat_sys_io_pgetevents_time64() in commit 6431e92fc827 ("parisc:
io_pgetevents_time64() needs compat syscall in 32-bit compat mode"), as well as by using more sophisticated system call wrappers on x86 and s390. However, arm64, mips, powerpc, sparc and riscv still have the same bug.

Change all of them over to use compat_sys_io_pgetevents_time64() like parisc already does. This was clearly the intention when the function was originally added, but it got hooked up incorrectly in the tables.

---
 arch/arm64/include/asm/unistd32.h         |    2 +-
 arch/mips/kernel/syscalls/syscall_n32.tbl |    2 +-
 arch/mips/kernel/syscalls/syscall_o32.tbl |    2 +-
 arch/powerpc/kernel/syscalls/syscall.tbl  |    2 +-
 arch/s390/kernel/syscalls/syscall.tbl     |    2 +-
 arch/sparc/kernel/syscalls/syscall.tbl    |    2 +-
 arch/x86/entry/syscalls/syscall_32.tbl    |    2 +-
 include/uapi/asm-generic/unistd.h         |    2 +-
 kernel/sys_ni.c                           |    2 +-
 9 files changed, 9 insertions(+), 9 deletions(-)

@@ -840,7 +840,7 @@ __SYSCALL(__NR_pselect6_time64, compat_s
 #define __NR_ppoll_time64 414
 __SYSCALL(__NR_ppoll_time64, compat_sys_ppoll_time64)
 #define __NR_io_pgetevents_time64 416
-__SYSCALL(__NR_io_pgetevents_time64, sys_io_pgetevents)
+__SYSCALL(__NR_io_pgetevents_time64, compat_sys_io_pgetevents_time64)
 #define __NR_recvmmsg_time64 417
 __SYSCALL(__NR_recvmmsg_time64, compat_sys_recvmmsg_time64)
 #define __NR_mq_timedsend_time64 418

@@ -354,7 +354,7 @@
 412	n32	utimensat_time64		sys_utimensat
 413	n32	pselect6_time64			compat_sys_pselect6_time64
 414	n32	ppoll_time64			compat_sys_ppoll_time64
-416	n32	io_pgetevents_time64		sys_io_pgetevents
+416	n32	io_pgetevents_time64		compat_sys_io_pgetevents_time64
 417	n32	recvmmsg_time64			compat_sys_recvmmsg_time64
 418	n32	mq_timedsend_time64		sys_mq_timedsend
 419	n32	mq_timedreceive_time64		sys_mq_timedreceive

@@ -403,7 +403,7 @@
 412	o32	utimensat_time64		sys_utimensat			sys_utimensat
 413	o32	pselect6_time64			sys_pselect6			compat_sys_pselect6_time64
 414	o32	ppoll_time64			sys_ppoll			compat_sys_ppoll_time64
-416	o32	io_pgetevents_time64		sys_io_pgetevents		sys_io_pgetevents
+416	o32	io_pgetevents_time64		sys_io_pgetevents		compat_sys_io_pgetevents_time64
 417	o32	recvmmsg_time64			sys_recvmmsg			compat_sys_recvmmsg_time64
 418	o32	mq_timedsend_time64		sys_mq_timedsend		sys_mq_timedsend
 419	o32	mq_timedreceive_time64		sys_mq_timedreceive		sys_mq_timedreceive

@@ -506,7 +506,7 @@
 412	32	utimensat_time64		sys_utimensat			sys_utimensat
 413	32	pselect6_time64			sys_pselect6			compat_sys_pselect6_time64
 414	32	ppoll_time64			sys_ppoll			compat_sys_ppoll_time64
-416	32	io_pgetevents_time64		sys_io_pgetevents		sys_io_pgetevents
+416	32	io_pgetevents_time64		sys_io_pgetevents		compat_sys_io_pgetevents_time64
 417	32	recvmmsg_time64			sys_recvmmsg			compat_sys_recvmmsg_time64
 418	32	mq_timedsend_time64		sys_mq_timedsend		sys_mq_timedsend
 419	32	mq_timedreceive_time64		sys_mq_timedreceive		sys_mq_timedreceive

@@ -418,7 +418,7 @@
 412	32	utimensat_time64	-				sys_utimensat
 413	32	pselect6_time64		-				compat_sys_pselect6_time64
 414	32	ppoll_time64		-				compat_sys_ppoll_time64
-416	32	io_pgetevents_time64	-				sys_io_pgetevents
+416	32	io_pgetevents_time64	-				compat_sys_io_pgetevents_time64
 417	32	recvmmsg_time64		-				compat_sys_recvmmsg_time64
 418	32	mq_timedsend_time64	-				sys_mq_timedsend
 419	32	mq_timedreceive_time64	-				sys_mq_timedreceive

@@ -461,7 +461,7 @@
 412	32	utimensat_time64		sys_utimensat			sys_utimensat
 413	32	pselect6_time64			sys_pselect6			compat_sys_pselect6_time64
 414	32	ppoll_time64			sys_ppoll			compat_sys_ppoll_time64
-416	32	io_pgetevents_time64		sys_io_pgetevents		sys_io_pgetevents
+416	32	io_pgetevents_time64		sys_io_pgetevents		compat_sys_io_pgetevents_time64
 417	32	recvmmsg_time64			sys_recvmmsg			compat_sys_recvmmsg_time64
 418	32	mq_timedsend_time64		sys_mq_timedsend		sys_mq_timedsend
 419	32	mq_timedreceive_time64		sys_mq_timedreceive		sys_mq_timedreceive

@@ -420,7 +420,7 @@
 412	i386	utimensat_time64	sys_utimensat
 413	i386	pselect6_time64		sys_pselect6			compat_sys_pselect6_time64
 414	i386	ppoll_time64		sys_ppoll			compat_sys_ppoll_time64
-416	i386	io_pgetevents_time64	sys_io_pgetevents
+416	i386	io_pgetevents_time64	sys_io_pgetevents		compat_sys_io_pgetevents_time64
 417	i386	recvmmsg_time64		sys_recvmmsg			compat_sys_recvmmsg_time64
 418	i386	mq_timedsend_time64	sys_mq_timedsend
 419	i386	mq_timedreceive_time64	sys_mq_timedreceive

@@ -737,7 +737,7 @@ __SC_COMP(__NR_pselect6_time64, sys_psel
 #define __NR_ppoll_time64 414
 __SC_COMP(__NR_ppoll_time64, sys_ppoll, compat_sys_ppoll_time64)
 #define __NR_io_pgetevents_time64 416
-__SYSCALL(__NR_io_pgetevents_time64, sys_io_pgetevents)
+__SC_COMP(__NR_io_pgetevents_time64, sys_io_pgetevents, compat_sys_io_pgetevents_time64)
 #define __NR_recvmmsg_time64 417
 __SC_COMP(__NR_recvmmsg_time64, sys_recvmmsg, compat_sys_recvmmsg_time64)
 #define __NR_mq_timedsend_time64 418

@@ -46,8 +46,8 @@ COND_SYSCALL(io_getevents_time32);
 COND_SYSCALL(io_getevents);
 COND_SYSCALL(io_pgetevents_time32);
 COND_SYSCALL(io_pgetevents);
-COND_SYSCALL_COMPAT(io_pgetevents_time32);
 COND_SYSCALL_COMPAT(io_pgetevents);
+COND_SYSCALL_COMPAT(io_pgetevents_time64);
 COND_SYSCALL(io_uring_setup);
 COND_SYSCALL(io_uring_enter);
 COND_SYSCALL(io_uring_register);

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Arnd Bergmann <[email protected]>

[ Upstream commit 63e2f40c9e3187641afacde4153f54b3ee4dbc8c ]

My earlier fix missed an incorrect function prototype that shows up on native 32-bit builds:

In file included from fs/notify/fanotify/fanotify_user.c:14:
include/linux/syscalls.h:248:25: error: conflicting types for 'sys_fanotify_mark'; have 'long int(int, unsigned int, u32, u32, int, const char *)' {aka 'long int(int, unsigned int, unsigned int, unsigned int, int, const char *)'}
1924 | SYSCALL32_DEFINE6(fanotify_mark,

      | ^~~~~~~~~~~~~~~~~

include/linux/syscalls.h:862:17: note: previous declaration of 'sys_fanotify_mark' with type 'long int(int, unsigned int, u64, int, const char *)' {aka 'long int(int, unsigned int, long long unsigned int, int, const char *)'}

On x86 and powerpc, the prototype is also wrong but hidden in an #ifdef, so it never caused problems.

Add another alternative declaration that matches the conditional function definition.

---
 include/linux/syscalls.h | 6 ++++++
 1 file changed, 6 insertions(+)

index 1f3d710afe4d7..ac3db59a74e23 100644
@@ -858,9 +858,15 @@ asmlinkage long sys_prlimit64(pid_t pid, unsigned int resource,
 				const struct rlimit64 __user *new_rlim,
 				struct rlimit64 __user *old_rlim);
 asmlinkage long sys_fanotify_init(unsigned int flags, unsigned int event_f_flags);
+#if defined(CONFIG_ARCH_SPLIT_ARG64)
+asmlinkage long sys_fanotify_mark(int fanotify_fd, unsigned int flags,
+                                unsigned int mask_1, unsigned int mask_2,
+				int dfd, const char  __user * pathname);
+#else
 asmlinkage long sys_fanotify_mark(int fanotify_fd, unsigned int flags,
 				  u64 mask, int fd,
 				  const char  __user *pathname);
+#endif
 asmlinkage long sys_name_to_handle_at(int dfd, const char __user *name,
 				      struct file_handle __user *handle,
 				      int __user *mnt_id, int flag);

2.43.0

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Kent Overstreet <[email protected]>

commit 692aa7a54b2b28d59f24b3bf8250837805484b99 upstream.

- bch2_sb_downgrade_validate() wasn't checking for a downgrade entry extending past the end of the superblock section

- for_each_downgrade_entry() is used in to_text() and needs to work on malformed input; it also was missing a check for a field extending past the end of the section

---
 fs/bcachefs/sb-downgrade.c |    7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

@@ -134,7 +134,8 @@ downgrade_entry_next_c(const struct bch_
 #define for_each_downgrade_entry(_d, _i)						\
 	for (const struct bch_sb_field_downgrade_entry *_i = (_d)->entries;		\
 	     (void *) _i	< vstruct_end(&(_d)->field) &&				\
-	     (void *) &_i->errors[0] < vstruct_end(&(_d)->field);			\
+	     (void *) &_i->errors[0] <= vstruct_end(&(_d)->field) &&			\
+	     (void *) downgrade_entry_next_c(_i) <= vstruct_end(&(_d)->field);		\
 	     _i = downgrade_entry_next_c(_i))
 
 static int bch2_sb_downgrade_validate(struct bch_sb *sb, struct bch_sb_field *f,

@@ -142,7 +143,9 @@ static int bch2_sb_downgrade_validate(st
 {
 	struct bch_sb_field_downgrade *e = field_to_type(f, downgrade);
 
-	for_each_downgrade_entry(e, i) {
+	for (const struct bch_sb_field_downgrade_entry *i = e->entries;
+	     (void *) i	< vstruct_end(&e->field);
+	     i = downgrade_entry_next_c(i)) {
 		if (BCH_VERSION_MAJOR(le16_to_cpu(i->version)) !=
 		    BCH_VERSION_MAJOR(le16_to_cpu(sb->version))) {
 			prt_printf(err, "downgrade entry with mismatched major version (%u != %u)",

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Kent Overstreet <[email protected]>

commit 9242a34b760648b722f4958749ad83ef7d0f7525 upstream.

Superblock downgrade entries are only two byte aligned, but section sizes are 8 byte aligned, which means we have to be careful about overrun checks; an entry that crosses the end of the section is allowed (and ignored) as long as it has zero errors.

---
 fs/bcachefs/sb-downgrade.c |    8 ++++++++
 1 file changed, 8 insertions(+)

@@ -146,6 +146,14 @@ static int bch2_sb_downgrade_validate(st
 	for (const struct bch_sb_field_downgrade_entry *i = e->entries;
 	     (void *) i	< vstruct_end(&e->field);
 	     i = downgrade_entry_next_c(i)) {
+		/*
+		 * Careful: sb_field_downgrade_entry is only 2 byte aligned, but
+		 * section sizes are 8 byte aligned - an empty entry spanning
+		 * the end of the section is allowed (and ignored):
+		 */
+		if ((void *) &i->errors[0] > vstruct_end(&e->field))
+			break;
+
 		if (BCH_VERSION_MAJOR(le16_to_cpu(i->version)) !=
 		    BCH_VERSION_MAJOR(le16_to_cpu(sb->version))) {
 			prt_printf(err, "downgrade entry with mismatched major version (%u != %u)",

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Kent Overstreet <[email protected]>

commit ddd118ab45e848b1956ef8c8ef84963a554b5b58 upstream.

Missing enum conversion

---
 fs/bcachefs/sb-downgrade.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -225,7 +225,7 @@ int bch2_sb_downgrade_update(struct bch_
 
 		dst = (void *) &darray_top(table);
 		dst->version = cpu_to_le16(src->version);
-		dst->recovery_passes[0]	= cpu_to_le64(src->recovery_passes);
+		dst->recovery_passes[0]	= cpu_to_le64(bch2_recovery_passes_to_stable(src->recovery_passes));
 		dst->recovery_passes[1]	= 0;
 		dst->nr_errors		= cpu_to_le16(src->nr_errors);
 		for (unsigned i = 0; i < src->nr_errors; i++)

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Kent Overstreet <[email protected]>

commit 247c056bde2ebc9fad2fc62332dc7cc99b58d720 upstream.

bch2_check_version_downgrade() was setting c->sb.version, which bch2_sb_set_downgrade() expects to be at the previous version; and it shouldn't even have been set directly because c->sb.version is updated by write_super().

---
 fs/bcachefs/super-io.c |   12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

@@ -1123,18 +1123,12 @@ bool bch2_check_version_downgrade(struct
 	 * c->sb will be checked before we write the superblock, so update it as
 	 * well:
 	 */
-	if (BCH_SB_VERSION_UPGRADE_COMPLETE(c->disk_sb.sb) > bcachefs_metadata_version_current) {
+	if (BCH_SB_VERSION_UPGRADE_COMPLETE(c->disk_sb.sb) > bcachefs_metadata_version_current)
 		SET_BCH_SB_VERSION_UPGRADE_COMPLETE(c->disk_sb.sb, bcachefs_metadata_version_current);
-		c->sb.version_upgrade_complete = bcachefs_metadata_version_current;
-	}
-	if (c->sb.version > bcachefs_metadata_version_current) {
+	if (c->sb.version > bcachefs_metadata_version_current)
 		c->disk_sb.sb->version = cpu_to_le16(bcachefs_metadata_version_current);
-		c->sb.version = bcachefs_metadata_version_current;
-	}
-	if (c->sb.version_min > bcachefs_metadata_version_current) {
+	if (c->sb.version_min > bcachefs_metadata_version_current)
 		c->disk_sb.sb->version_min = cpu_to_le16(bcachefs_metadata_version_current);
-		c->sb.version_min = bcachefs_metadata_version_current;
-	}
 	c->disk_sb.sb->compat[0] &= cpu_to_le64((1ULL << BCH_COMPAT_NR) - 1);
 	return ret;
 }

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Kent Overstreet <[email protected]>

commit 088d0de81220a74d7d553febb81656927f10bb16 upstream.

Compatibility fix - we no longer have a separate table for which order gc walks btrees in, and special case the stripes btree directly.

---
 fs/bcachefs/bcachefs.h       |   44 ---------------------------------------
 fs/bcachefs/btree_gc.c       |   15 ++++++-------
 fs/bcachefs/btree_gc.h       |   48 +++++++++++++++++++------------------------
 fs/bcachefs/btree_gc_types.h |   29 +++++++++++++++++++++++++
 fs/bcachefs/ec.c             |    2 -
 5 files changed, 60 insertions(+), 78 deletions(-)
 create mode 100644 fs/bcachefs/btree_gc_types.h

@@ -455,6 +455,7 @@ enum bch_time_stats {
 };
 
 #include "alloc_types.h"
+#include "btree_gc_types.h"
 #include "btree_types.h"
 #include "btree_node_scan_types.h"
 #include "btree_write_buffer_types.h"

@@ -485,49 +486,6 @@ enum bch_time_stats {
 
 struct btree;
 
-enum gc_phase {
-	GC_PHASE_NOT_RUNNING,
-	GC_PHASE_START,
-	GC_PHASE_SB,
-
-	GC_PHASE_BTREE_stripes,
-	GC_PHASE_BTREE_extents,
-	GC_PHASE_BTREE_inodes,
-	GC_PHASE_BTREE_dirents,
-	GC_PHASE_BTREE_xattrs,
-	GC_PHASE_BTREE_alloc,
-	GC_PHASE_BTREE_quotas,
-	GC_PHASE_BTREE_reflink,
-	GC_PHASE_BTREE_subvolumes,
-	GC_PHASE_BTREE_snapshots,
-	GC_PHASE_BTREE_lru,
-	GC_PHASE_BTREE_freespace,
-	GC_PHASE_BTREE_need_discard,
-	GC_PHASE_BTREE_backpointers,
-	GC_PHASE_BTREE_bucket_gens,
-	GC_PHASE_BTREE_snapshot_trees,
-	GC_PHASE_BTREE_deleted_inodes,
-	GC_PHASE_BTREE_logged_ops,
-	GC_PHASE_BTREE_rebalance_work,
-	GC_PHASE_BTREE_subvolume_children,
-
-	GC_PHASE_PENDING_DELETE,
-};
-
-struct gc_pos {
-	enum gc_phase		phase;
-	struct bpos		pos;
-	unsigned		level;
-};
-
-struct reflink_gc {
-	u64		offset;
-	u32		size;
-	u32		refcount;
-};
-
-typedef GENRADIX(struct reflink_gc) reflink_gc_table;
-
 struct io_count {
 	u64			sectors[2][BCH_DATA_NR];
 };

@@ -1080,8 +1080,7 @@ fsck_err:
 
 static inline int btree_id_gc_phase_cmp(enum btree_id l, enum btree_id r)
 {
-	return  (int) btree_id_to_gc_phase(l) -
-		(int) btree_id_to_gc_phase(r);
+	return cmp_int(gc_btree_order(l), gc_btree_order(r));
 }
 
 static int bch2_gc_btrees(struct bch_fs *c, bool initial, bool metadata_only)

@@ -1126,7 +1125,7 @@ static void mark_metadata_sectors(struct
 			min_t(u64, bucket_to_sector(ca, b + 1), end) - start;
 
 		bch2_mark_metadata_bucket(c, ca, b, type, sectors,
-					  gc_phase(GC_PHASE_SB), flags);
+					  gc_phase(GC_PHASE_sb), flags);
 		b++;
 		start += sectors;
 	} while (start < end);

@@ -1155,14 +1154,14 @@ static void bch2_mark_dev_superblock(str
 		b = ca->journal.buckets[i];
 		bch2_mark_metadata_bucket(c, ca, b, BCH_DATA_journal,
 					  ca->mi.bucket_size,
-					  gc_phase(GC_PHASE_SB), flags);
+					  gc_phase(GC_PHASE_sb), flags);
 	}
 }
 
 static void bch2_mark_superblocks(struct bch_fs *c)
 {
 	mutex_lock(&c->sb_lock);
-	gc_pos_set(c, gc_phase(GC_PHASE_SB));
+	gc_pos_set(c, gc_phase(GC_PHASE_sb));
 
 	for_each_online_member(c, ca)
 		bch2_mark_dev_superblock(c, ca, BTREE_TRIGGER_GC);

@@ -1773,7 +1772,7 @@ int bch2_gc(struct bch_fs *c, bool initi
 	if (ret)
 		goto out;
 again:
-	gc_pos_set(c, gc_phase(GC_PHASE_START));
+	gc_pos_set(c, gc_phase(GC_PHASE_start));
 
 	bch2_mark_superblocks(c);
 

@@ -1800,7 +1799,7 @@ again:
 		 */
 		bch_info(c, "Second GC pass needed, restarting:");
 		clear_bit(BCH_FS_need_another_gc, &c->flags);
-		__gc_pos_set(c, gc_phase(GC_PHASE_NOT_RUNNING));
+		__gc_pos_set(c, gc_phase(GC_PHASE_not_running));
 
 		bch2_gc_stripes_reset(c, metadata_only);
 		bch2_gc_alloc_reset(c, metadata_only);

@@ -1827,7 +1826,7 @@ out:
 
 	percpu_down_write(&c->mark_lock);
 	/* Indicates that gc is no longer in progress: */
-	__gc_pos_set(c, gc_phase(GC_PHASE_NOT_RUNNING));
+	__gc_pos_set(c, gc_phase(GC_PHASE_not_running));
 
 	bch2_gc_free(c);
 	percpu_up_write(&c->mark_lock);

@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _BCACHEFS_BTREE_GC_TYPES_H
+#define _BCACHEFS_BTREE_GC_TYPES_H
+
+#include <linux/generic-radix-tree.h>
+
+enum gc_phase {
+	GC_PHASE_not_running,
+	GC_PHASE_start,
+	GC_PHASE_sb,
+	GC_PHASE_btree,
+};
+
+struct gc_pos {
+	enum gc_phase		phase:8;
+	enum btree_id		btree:8;
+	u16			level;
+	struct bpos		pos;
+};
+
+struct reflink_gc {
+	u64		offset;
+	u32		size;
+	u32		refcount;
+};
+
+typedef GENRADIX(struct reflink_gc) reflink_gc_table;
+
+#endif /* _BCACHEFS_BTREE_GC_TYPES_H */

@@ -880,7 +880,7 @@ static int __ec_stripe_mem_alloc(struct
 	if (!genradix_ptr_alloc(&c->stripes, idx, gfp))
 		return -BCH_ERR_ENOMEM_ec_stripe_mem_alloc;
 
-	if (c->gc_pos.phase != GC_PHASE_NOT_RUNNING &&
+	if (c->gc_pos.phase != GC_PHASE_not_running &&
 	    !genradix_ptr_alloc(&c->gc_stripes, idx, gfp))
 		return -BCH_ERR_ENOMEM_ec_stripe_mem_alloc;
 

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Greg Kroah-Hartman <[email protected]>

This reverts commit b3dcad0bfd62fcc9367d8092a71918db53804f53 which is commit cd4a32e60061789676f7f018a94fcc9ec56732a0 upstream.

Turned out that this should not have been applied to the stable tree.

---
 drivers/net/phy/sfp.c |   18 ++++++------------
 1 file changed, 6 insertions(+), 12 deletions(-)

@@ -385,23 +385,18 @@ static void sfp_fixup_rollball(struct sf
 	sfp->phy_t_retry = msecs_to_jiffies(1000);
 }
 
-static void sfp_fixup_fs_2_5gt(struct sfp *sfp)
+static void sfp_fixup_fs_10gt(struct sfp *sfp)
 {
+	sfp_fixup_10gbaset_30m(sfp);
 	sfp_fixup_rollball(sfp);
 
-	/* The RollBall fixup is not enough for FS modules, the PHY chip inside
+	/* The RollBall fixup is not enough for FS modules, the AQR chip inside
 	 * them does not return 0xffff for PHY ID registers in all MMDs for the
 	 * while initializing. They need a 4 second wait before accessing PHY.
 	 */
 	sfp->module_t_wait = msecs_to_jiffies(4000);
 }
 
-static void sfp_fixup_fs_10gt(struct sfp *sfp)
-{
-	sfp_fixup_10gbaset_30m(sfp);
-	sfp_fixup_fs_2_5gt(sfp);
-}
-
 static void sfp_fixup_halny_gsfp(struct sfp *sfp)
 {
 	/* Ignore the TX_FAULT and LOS signals on this module.

@@ -477,10 +472,6 @@ static const struct sfp_quirk sfp_quirks
 	// Rollball protocol to talk to the PHY.
 	SFP_QUIRK_F("FS", "SFP-10G-T", sfp_fixup_fs_10gt),
 
-	// Fiberstore SFP-2.5G-T uses Rollball protocol to talk to the PHY and
-	// needs 4 sec wait before probing the PHY.
-	SFP_QUIRK_F("FS", "SFP-2.5G-T", sfp_fixup_fs_2_5gt),
-
 	// Fiberstore GPON-ONU-34-20BI can operate at 2500base-X, but report 1.2GBd
 	// NRZ in their EEPROM
 	SFP_QUIRK("FS", "GPON-ONU-34-20BI", sfp_quirk_2500basex,

@@ -497,6 +488,9 @@ static const struct sfp_quirk sfp_quirks
 	SFP_QUIRK("HUAWEI", "MA5671A", sfp_quirk_2500basex,
 		  sfp_fixup_ignore_tx_fault),
 
+	// FS 2.5G Base-T
+	SFP_QUIRK_M("FS", "SFP-2.5G-T", sfp_quirk_oem_2_5g),
+
 	// Lantech 8330-262D-E can operate at 2500base-X, but incorrectly report
 	// 2500MBd NRZ in their EEPROM
 	SFP_QUIRK_M("Lantech", "8330-262D-E", sfp_quirk_2500basex),

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: yangge <[email protected]>

commit bf14ed81f571f8dba31cd72ab2e50fbcc877cc31 upstream.

Since commit 5d0a661d808f ("mm/page_alloc: use only one PCP list for THP-sized allocations") no longer differentiates the migration type of pages in THP-sized PCP list, it's possible that non-movable allocation requests may get a CMA page from the list, in some cases, it's not acceptable.

If a large number of CMA memory are configured in system (for example, the CMA memory accounts for 50% of the system memory), starting a virtual machine with device passthrough will get stuck. During starting the virtual machine, it will call pin_user_pages_remote(..., FOLL_LONGTERM, ...) to pin memory. Normally if a page is present and in CMA area, pin_user_pages_remote() will migrate the page from CMA area to non-CMA area because of FOLL_LONGTERM flag. But if non-movable allocation requests return CMA memory, migrate_longterm_unpinnable_pages() will migrate a CMA page to another CMA page, which will fail to pass the check in check_and_migrate_movable_pages() and cause migration endless.

Call trace:

pin_user_pages_remote
--__gup_longterm_locked // endless loops in this function ----_get_user_pages_locked ----check_and_migrate_movable_pages ------migrate_longterm_unpinnable_pages --------alloc_migration_target

This problem will also have a negative impact on CMA itself. For example, when CMA is borrowed by THP, and we need to reclaim it through cma_alloc() or dma_alloc_coherent(), we must move those pages out to ensure CMA's users can retrieve that contigous memory. Currently, CMA's memory is occupied by non-movable pages, meaning we can't relocate them. As a result, cma_alloc() is more likely to fail.

To fix the problem above, we add one PCP list for THP, which will not introduce a new cacheline for struct per_cpu_pages. THP will have 2 PCP lists, one PCP list is used by MOVABLE allocation, and the other PCP list is used by UNMOVABLE allocation. MOVABLE allocation contains GPF_MOVABLE, and UNMOVABLE allocation contains GFP_UNMOVABLE and GFP_RECLAIMABLE.

---
 include/linux/mmzone.h |    9 ++++-----
 mm/page_alloc.c        |    9 +++++++--
 2 files changed, 11 insertions(+), 7 deletions(-)

@@ -651,13 +651,12 @@ enum zone_watermarks {
 };
 
 /*
- * One per migratetype for each PAGE_ALLOC_COSTLY_ORDER. One additional list
- * for THP which will usually be GFP_MOVABLE. Even if it is another type,
- * it should not contribute to serious fragmentation causing THP allocation
- * failures.
+ * One per migratetype for each PAGE_ALLOC_COSTLY_ORDER. Two additional lists
+ * are added for THP. One PCP list is used by GPF_MOVABLE, and the other PCP list
+ * is used by GFP_UNMOVABLE and GFP_RECLAIMABLE.
  */
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
-#define NR_PCP_THP 1
+#define NR_PCP_THP 2
 #else
 #define NR_PCP_THP 0
 #endif

@@ -521,10 +521,15 @@ out:
 
 static inline unsigned int order_to_pindex(int migratetype, int order)
 {
+	bool __maybe_unused movable;
+
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 	if (order > PAGE_ALLOC_COSTLY_ORDER) {
 		VM_BUG_ON(order != pageblock_order);
-		return NR_LOWORDER_PCP_LISTS;
+
+		movable = migratetype == MIGRATE_MOVABLE;
+
+		return NR_LOWORDER_PCP_LISTS + movable;
 	}
 #else
 	VM_BUG_ON(order > PAGE_ALLOC_COSTLY_ORDER);

@@ -538,7 +543,7 @@ static inline int pindex_to_order(unsign
 	int order = pindex / MIGRATE_PCPTYPES;
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
-	if (pindex == NR_LOWORDER_PCP_LISTS)
+	if (pindex >= NR_LOWORDER_PCP_LISTS)
 		order = pageblock_order;
 #else
 	VM_BUG_ON(order > PAGE_ALLOC_COSTLY_ORDER);

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Uwe Kleine-König <[email protected]>

commit dab8f9f0fe3aada61c0eb013dcf7d3ff75a2c336 upstream.

A small prescaler is beneficial, as this improves the resolution of the duty_cycle configuration. However if the prescaler is too small, the maximal possible period becomes considerably smaller than the requested value.

One situation where this goes wrong is the following: With a parent clock rate of 208877930 Hz and max_arr = 0xffff = 65535, a request for period = 941243 ns currently results in PSC = 1. The value for ARR is then calculated to

	ARR = 941243 * 208877930 / (1000000000 * 2) - 1 = 98301

This value is bigger than 65535 however and so doesn't fit into the respective register field. In this particular case the PWM was configured for a period of 313733.4806027616 ns (with ARR = 98301 & 0xffff). Even if ARR was configured to its maximal value, only period = 627495.6861167669 ns would be achievable.

Fix the calculation accordingly and adapt the comment to match the new algorithm.

With the calculation fixed the above case results in PSC = 2 and so an actual period of 941229.1667195285 ns.

---
 drivers/pwm/pwm-stm32.c |   18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)

@@ -321,17 +321,23 @@ static int stm32_pwm_config(struct stm32
 	 * First we need to find the minimal value for prescaler such that
 	 *
 	 *        period_ns * clkrate
-	 *   ------------------------------
+	 *   ------------------------------ < max_arr + 1
 	 *   NSEC_PER_SEC * (prescaler + 1)
 	 *
-	 * isn't bigger than max_arr.
+	 * This equation is equivalent to
+	 *
+	 *        period_ns * clkrate
+	 *   ---------------------------- < prescaler + 1
+	 *   NSEC_PER_SEC * (max_arr + 1)
+	 *
+	 * Using integer division and knowing that the right hand side is
+	 * integer, this is further equivalent to
+	 *
+	 *   (period_ns * clkrate) // (NSEC_PER_SEC * (max_arr + 1)) ≤ prescaler
 	 */
 
 	prescaler = mul_u64_u64_div_u64(period_ns, clk_get_rate(priv->clk),
-					(u64)NSEC_PER_SEC * priv->max_arr);
-	if (prescaler > 0)
-		prescaler -= 1;
-
+					(u64)NSEC_PER_SEC * ((u64)priv->max_arr + 1));
 	if (prescaler > MAX_TIM_PSC)
 		return -EINVAL;
 

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Uwe Kleine-König <[email protected]>

commit f01af3022d4a46362c5dda3d35dea939f3246d10 upstream.

"Failed to lock the clock" is an appropriate error message for clk_rate_exclusive_get() failing, but not for the clock running too fast for the driver's calculations.

Adapt the error message accordingly.

---
 drivers/pwm/pwm-stm32.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -679,7 +679,8 @@ static int stm32_pwm_probe(struct platfo
 	 * .apply() won't overflow.
 	 */
 	if (clk_get_rate(priv->clk) > 1000000000)
-		return dev_err_probe(dev, -EINVAL, "Failed to lock clock\n");
+		return dev_err_probe(dev, -EINVAL, "Clock freq too high (%lu)\n",
+				     clk_get_rate(priv->clk));
 
 	chip->ops = &stm32pwm_ops;
 

6.9-stable review patch. If anyone has any objections, please let me know.

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

From: Jonas Karlman <[email protected]>

[ Upstream commit 1fb98c855ccd7bc7f50c7a9626fbb8440454760b ]

Radxa ROCK Pi S have optional onboard SD NAND on board revision v1.1, v1.2 and v1.3, revision v1.5 changed to use optional onboard eMMC.

The optional SD NAND typically fails to initialize:

mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc0: error -110 whilst initialising SD card
mmc_host mmc0: Bus speed (slot 0) = 300000Hz (slot req 300000Hz, actual 300000HZ div = 0)
mmc0: error -110 whilst initialising SD card
mmc_host mmc0: Bus speed (slot 0) = 200000Hz (slot req 200000Hz, actual 200000HZ div = 0)
mmc0: error -110 whilst initialising SD card
mmc_host mmc0: Bus speed (slot 0) = 100000Hz (slot req 100000Hz, actual 100000HZ div = 0)
mmc0: error -110 whilst initialising SD card

Add pinctrl and cap-sd-highspeed to fix SD NAND initialization. Also drop bus-width and mmc-hs200-1_8v to fix eMMC initialization on the new v1.5 board revision, only 3v3 signal voltage is used.