Add a pr_err() to print the add device error in find_device_iter()
Signed-off-by: Abhinav Jain <[email protected]>
---
drivers/pci/pcie/aer.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
index ac6293c24976..0e1ad2998116 100644
--- a/drivers/pci/pcie/aer.c
+++ b/drivers/pci/pcie/aer.c
@@ -885,7 +885,8 @@ static int find_device_iter(struct pci_dev *dev, void *data)
/* List this device */
if (add_error_device(e_info, dev)) {
/* We cannot handle more... Stop iteration */
- /* TODO: Should print error message here? */
+ pr_err("find_device_iter: Cannot handle more devices.
+ Stopping iteration");
return 1;
}
--
2.34.1
On Mon, 2024-04-15 at 16:10 +0000, Abhinav Jain wrote:
> Add a pr_err() to print the add device error in find_device_iter()
[]
> diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
[]
> @@ -885,7 +885,8 @@ static int find_device_iter(struct pci_dev *dev, void *data)
> /* List this device */
> if (add_error_device(e_info, dev)) {
> /* We cannot handle more... Stop iteration */
> - /* TODO: Should print error message here? */
> + pr_err("find_device_iter: Cannot handle more devices.
> + Stopping iteration");
You are adding unnecessary whitespace after the period.
String concatenation keeps _all_ the whitespace.
The format is fine on a single line too.
Something like:
pr_notice("%s: Cannot handle more devices - iteration stopped\n",
__func__);
Print the add device error in find_device_iter()
Signed-off-by: Abhinav Jain <[email protected]>
PATCH v1 link : https://lore.kernel.org/all/[email protected]/
Changes since v1:
- Replaced pr_err() with pr_notice()
- Removed unncessary whitespaces
---
drivers/pci/pcie/aer.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
index 0e1ad2998116..8b820a74dd6b 100644
--- a/drivers/pci/pcie/aer.c
+++ b/drivers/pci/pcie/aer.c
@@ -885,8 +885,8 @@ static int find_device_iter(struct pci_dev *dev, void *data)
/* List this device */
if (add_error_device(e_info, dev)) {
/* We cannot handle more... Stop iteration */
- pr_err("find_device_iter: Cannot handle more devices.
- Stopping iteration");
+ pr_notice("%s: Cannot handle more devices - iteration stopped\n",
+ __func__);
return 1;
}
--
2.34.1
On Wed, Jun 05, 2024 at 09:23:44PM +0000, Abhinav Jain wrote:
> Print the add device error in find_device_iter()
>
> Signed-off-by: Abhinav Jain <[email protected]>
>
> PATCH v1 link : https://lore.kernel.org/all/[email protected]/
>
> Changes since v1:
> - Replaced pr_err() with pr_notice()
> - Removed unncessary whitespaces
> ---
Thanks for looking at this.
- It doesn't apply to -rc1 (the TODO message is missing). In PCI,
we normally apply patches on topic branches based on -rc1.
- The subject should be more specific so it makes sense all by
itself, e.g., "Log note if we find too many devices with errors"
- Add period at end of sentence in commit log.
- Move historical notes (v1 URL, changes since v1) below the "---"
line so they don't get included in the commit log.
- __func__ is not relevant here -- that's generally a debugging
thing. We can find the function by searching for the message
text. In cases like this, I'd rather have something that helps
identify a *device* that's related to the message, e.g., the
pci_dev in this case. So I'd suggest pci_err(dev, "...") here.
- I'd keep pci_err() instead of switching to pr_notice(). If we get
this message, we should re-think the way we collect this
information, so I want to hear about it.
> drivers/pci/pcie/aer.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c
> index 0e1ad2998116..8b820a74dd6b 100644
> --- a/drivers/pci/pcie/aer.c
> +++ b/drivers/pci/pcie/aer.c
> @@ -885,8 +885,8 @@ static int find_device_iter(struct pci_dev *dev, void *data)
> /* List this device */
> if (add_error_device(e_info, dev)) {
> /* We cannot handle more... Stop iteration */
> - pr_err("find_device_iter: Cannot handle more devices.
> - Stopping iteration");
> + pr_notice("%s: Cannot handle more devices - iteration stopped\n",
> + __func__);
> return 1;
> }
>
> --
> 2.34.1
>
On Wed, 5 Jun 2024 16:58:48 -0500, Bjorn Helgaas wrote:
> - It doesn't apply to -rc1 (the TODO message is missing). In PCI,
> we normally apply patches on topic branches based on -rc1.
Thank you for the detailed feedback. I was looking at mainline only.
> - The subject should be more specific so it makes sense all by
> itself, e.g., "Log note if we find too many devices with errors"
> - Add period at end of sentence in commit log.
> - Move historical notes (v1 URL, changes since v1) below the "---"
> line so they don't get included in the commit log.
I have included these changes to the v3. Please find it here:
https://lore.kernel.org/all/[email protected]/
> - __func__ is not relevant here -- that's generally a debugging
> thing. We can find the function by searching for the message
> text. In cases like this, I'd rather have something that helps
> identify a *device* that's related to the message, e.g., the
> pci_dev in this case. So I'd suggest pci_err(dev, "...") here.
> - I'd keep pci_err() instead of switching to pr_notice(). If we get
> this message, we should re-think the way we collect this
> information, so I want to hear about it.
I understand, this helped me get a clear picture of what needs to be
done. I have accordingly added two pci_err for the same. Please review.