Patch "PCI/sysfs: Ensure devices are powered for config reads" has been added to the 6.6-stable tree

This is a note to let you know that I've just added the patch titled

PCI/sysfs: Ensure devices are powered for config reads

to the 6.6-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
pci-sysfs-ensure-devices-are-powered-for-config-reads.patch
and it can be found in the queue-6.6 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.

Wrap these access in pci_config_pm_runtime_{get,put}() like most of the rest of the similar sysfs attributes.

Notably, "max_link_speed" does not access config registers; it returns a cached value since d2bd39c0456b ("PCI: Store all PCIe Supported Link Speeds").

Fixes 56c1af4606f0 ("PCI: Add sysfs max_link_speed/width, current_link_speed/width, etc")

Signed-off-by Brian Norris <[email protected]>

Signed-off-by Brian Norris <[email protected]>

Signed-off-by Bjorn Helgaas <[email protected]>

Cc [email protected]

Link https://patch.msgid.link/20250924095711.v2.1.Ibb5b6ca1e2c059e04ec53140cd98a44f2684c668@changeid open here

Signed-off-by Greg Kroah-Hartman <[email protected]>

---
 drivers/pci/pci-sysfs.c |   20 +++++++++++++++++++-
 1 file changed, 19 insertions(+), 1 deletion(-)

drivers/pci/pci-sysfs.c +19 -1

@@ -196,8 +196,14 @@ static ssize_t max_link_width_show(struc
 				   struct device_attribute *attr, char *buf)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
+	ssize_t ret;
 
-	return sysfs_emit(buf, "%u\n", pcie_get_width_cap(pdev));
+	/* We read PCI_EXP_LNKCAP, so we need the device to be accessible. */
+	pci_config_pm_runtime_get(pdev);
+	ret = sysfs_emit(buf, "%u\n", pcie_get_width_cap(pdev));
+	pci_config_pm_runtime_put(pdev);
+
+	return ret;
 }
 static DEVICE_ATTR_RO(max_link_width);
 

@@ -209,7 +215,10 @@ static ssize_t current_link_speed_show(s
 	int err;
 	enum pci_bus_speed speed;
 
+	pci_config_pm_runtime_get(pci_dev);
 	err = pcie_capability_read_word(pci_dev, PCI_EXP_LNKSTA, &linkstat);
+	pci_config_pm_runtime_put(pci_dev);
+
 	if (err)
 		return -EINVAL;
 

@@ -226,7 +235,10 @@ static ssize_t current_link_width_show(s
 	u16 linkstat;
 	int err;
 
+	pci_config_pm_runtime_get(pci_dev);
 	err = pcie_capability_read_word(pci_dev, PCI_EXP_LNKSTA, &linkstat);
+	pci_config_pm_runtime_put(pci_dev);
+
 	if (err)
 		return -EINVAL;
 

@@ -242,7 +254,10 @@ static ssize_t secondary_bus_number_show
 	u8 sec_bus;
 	int err;
 
+	pci_config_pm_runtime_get(pci_dev);
 	err = pci_read_config_byte(pci_dev, PCI_SECONDARY_BUS, &sec_bus);
+	pci_config_pm_runtime_put(pci_dev);
+
 	if (err)
 		return -EINVAL;
 

@@ -258,7 +273,10 @@ static ssize_t subordinate_bus_number_sh
 	u8 sub_bus;
 	int err;
 
+	pci_config_pm_runtime_get(pci_dev);
 	err = pci_read_config_byte(pci_dev, PCI_SUBORDINATE_BUS, &sub_bus);
+	pci_config_pm_runtime_put(pci_dev);
+
 	if (err)
 		return -EINVAL;
 

Patches currently in stable-queue which might be from [email protected] are

queue-6.6/pci-sysfs-ensure-devices-are-powered-for-config-reads.patch

Hi,

Quoted 8 lines · Greg Kroah-Hartman

This is a note to let you know that I've just added the patch titled

PCI/sysfs: Ensure devices are powered for config reads

to the 6.6-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

Quoted 6 lines

The filename of the patch is:
pci-sysfs-ensure-devices-are-powered-for-config-reads.patch
and it can be found in the queue-6.6 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.

Adding to the stable tree is good IMO, but one note about exactly how to do so below:

Quoted 6 lines

Wrap these access in pci_config_pm_runtime_{get,put}() like most of the rest of the similar sysfs attributes.

Notably, "max_link_speed" does not access config registers; it returns a cached value since d2bd39c0456b ("PCI: Store all PCIe Supported Link Speeds").

^^ This note about commit d2bd39c0456b was specifically to provide hints about backporting. Without commit d2bd39c0456b, the solution is somewhat incomplete. We should either backport commit d2bd39c0456b as well, or we should adapt the change to add pci_config_pm_runtime_{get,put}() in max_link_speed_show() too.

Commit d2bd39c0456b was already ported to 6.12.y, but seemingly no further.

If adapting this change to pre-commit-d2bd39c0456b is better, I can submit an updated version here.

Without commit d2bd39c0456b, it just means that the 'max_link_speed' sysfs attribute is still susceptible to accessing a powered-down device/link. We're in no worse state than we were without this patch. And frankly, people are not likely to notice if they haven't already, since I'd guess most systems don't suspend devices this aggressively.

Quoted 28 lines · Brian Norris

Hi,

Quoted 8 lines · Greg Kroah-Hartman

This is a note to let you know that I've just added the patch titled

PCI/sysfs: Ensure devices are powered for config reads

to the 6.6-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

Quoted 6 lines

The filename of the patch is:
pci-sysfs-ensure-devices-are-powered-for-config-reads.patch
and it can be found in the queue-6.6 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree, please let <[email protected]> know about it.

Adding to the stable tree is good IMO, but one note about exactly how to do so below:

Quoted 6 lines

Wrap these access in pci_config_pm_runtime_{get,put}() like most of the rest of the similar sysfs attributes.

Notably, "max_link_speed" does not access config registers; it returns a cached value since d2bd39c0456b ("PCI: Store all PCIe Supported Link Speeds").

Quoted 5 lines

^^ This note about commit d2bd39c0456b was specifically to provide hints about backporting. Without commit d2bd39c0456b, the solution is somewhat incomplete. We should either backport commit d2bd39c0456b as well, or we should adapt the change to add pci_config_pm_runtime_{get,put}() in max_link_speed_show() too.

I missed that "hint", you need to make it bindingly obvious as I churn through the giant "-rc1 merge dump" very quickly as obviously those are changes that were not serious enough to make it into -final :)

Quoted 6 lines

Commit d2bd39c0456b was already ported to 6.12.y, but seemingly no further.

If adapting this change to pre-commit-d2bd39c0456b is better, I can submit an updated version here.

Quoted 5 lines

Without commit d2bd39c0456b, it just means that the 'max_link_speed' sysfs attribute is still susceptible to accessing a powered-down device/link. We're in no worse state than we were without this patch. And frankly, people are not likely to notice if they haven't already, since I'd guess most systems don't suspend devices this aggressively.

I'll gladly accept a fixed up patch for this, thanks.

Quoted 10 lines · Greg Kroah-Hartman

Quoted 8 lines · Brian Norris

If adapting this change to pre-commit-d2bd39c0456b is better, I can submit an updated version here.

Without commit d2bd39c0456b, it just means that the 'max_link_speed' sysfs attribute is still susceptible to accessing a powered-down device/link. We're in no worse state than we were without this patch. And frankly, people are not likely to notice if they haven't already, since I'd guess most systems don't suspend devices this aggressively.

Quoted 1 line

I'll gladly accept a fixed up patch for this, thanks.

I'll try to get that out today.