netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] Octeontx2-af: Fix pci_alloc_irq_vectors() return value check
@ 2025-10-14 10:14 Harshit Mogalapalli
  2025-10-14 10:20 ` Harshit Mogalapalli
  0 siblings, 1 reply; 4+ messages in thread
From: Harshit Mogalapalli @ 2025-10-14 10:14 UTC (permalink / raw)
  To: Sunil Goutham, Linu Cherian, Geetha sowjanya, Jerin Jacob,
	hariprasad, Subbaraya Sundeep, Andrew Lunn, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Nithya Mani, netdev,
	linux-kernel
  Cc: dan.carpenter, kernel-janitors, error27, harshit.m.mogalapalli

In cgx_probe() when pci_alloc_irq_vectors() fails the error value will
be negative and that check is sufficient.

	err = pci_alloc_irq_vectors(pdev, nvec, nvec, PCI_IRQ_MSIX);
        if (err < 0 || err != nvec) {
        	...
	}

Remove the check which compares err with nvec.

Fixes: 1463f382f58d ("octeontx2-af: Add support for CGX link management")
Suggested-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
---
Only compile tested.
---
 drivers/net/ethernet/marvell/octeontx2/af/cgx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
index d374a4454836..f4d5a3c05fa4 100644
--- a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
+++ b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
@@ -1993,7 +1993,7 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 
 	nvec = pci_msix_vec_count(cgx->pdev);
 	err = pci_alloc_irq_vectors(pdev, nvec, nvec, PCI_IRQ_MSIX);
-	if (err < 0 || err != nvec) {
+	if (err < 0) {
 		dev_err(dev, "Request for %d msix vectors failed, err %d\n",
 			nvec, err);
 		goto err_release_regions;
-- 
2.39.3


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH] Octeontx2-af: Fix pci_alloc_irq_vectors() return value check
  2025-10-14 10:14 [PATCH] Octeontx2-af: Fix pci_alloc_irq_vectors() return value check Harshit Mogalapalli
@ 2025-10-14 10:20 ` Harshit Mogalapalli
  2025-10-15  8:36   ` Simon Horman
  0 siblings, 1 reply; 4+ messages in thread
From: Harshit Mogalapalli @ 2025-10-14 10:20 UTC (permalink / raw)
  To: Sunil Goutham, Linu Cherian, Geetha sowjanya, Jerin Jacob,
	hariprasad, Subbaraya Sundeep, Andrew Lunn, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Nithya Mani, netdev,
	linux-kernel
  Cc: dan.carpenter, kernel-janitors, error27

Hi,

On 14/10/25 15:44, Harshit Mogalapalli wrote:
> In cgx_probe() when pci_alloc_irq_vectors() fails the error value will
> be negative and that check is sufficient.
> 
> 	err = pci_alloc_irq_vectors(pdev, nvec, nvec, PCI_IRQ_MSIX);
>          if (err < 0 || err != nvec) {
>          	...
> 	}
> 
> Remove the check which compares err with nvec.
> 
> Fixes: 1463f382f58d ("octeontx2-af: Add support for CGX link management")
> Suggested-by: Paolo Abeni <pabeni@redhat.com>
> Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
> ---
> Only compile tested.
> ---
>   drivers/net/ethernet/marvell/octeontx2/af/cgx.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
> index d374a4454836..f4d5a3c05fa4 100644
> --- a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
> +++ b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
> @@ -1993,7 +1993,7 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
>   
>   	nvec = pci_msix_vec_count(cgx->pdev);
>   	err = pci_alloc_irq_vectors(pdev, nvec, nvec, PCI_IRQ_MSIX);
> -	if (err < 0 || err != nvec) {
> +	if (err < 0) {
>   		dev_err(dev, "Request for %d msix vectors failed, err %d\n",
>   			nvec, err);


Now that I think about it more, maybe we want to error out when err != 
nvec as well ? In that case maybe the right thing to do is leave the 
check as is and set err = -EXYZ before goto ?

Thanks,
Harshit>   		goto err_release_regions;


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] Octeontx2-af: Fix pci_alloc_irq_vectors() return value check
  2025-10-14 10:20 ` Harshit Mogalapalli
@ 2025-10-15  8:36   ` Simon Horman
  2025-10-15  8:56     ` Harshit Mogalapalli
  0 siblings, 1 reply; 4+ messages in thread
From: Simon Horman @ 2025-10-15  8:36 UTC (permalink / raw)
  To: Harshit Mogalapalli
  Cc: Sunil Goutham, Linu Cherian, Geetha sowjanya, Jerin Jacob,
	hariprasad, Subbaraya Sundeep, Andrew Lunn, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Nithya Mani, netdev,
	linux-kernel, dan.carpenter, kernel-janitors, error27

On Tue, Oct 14, 2025 at 03:50:47PM +0530, Harshit Mogalapalli wrote:
> Hi,
> 
> On 14/10/25 15:44, Harshit Mogalapalli wrote:
> > In cgx_probe() when pci_alloc_irq_vectors() fails the error value will
> > be negative and that check is sufficient.
> > 
> > 	err = pci_alloc_irq_vectors(pdev, nvec, nvec, PCI_IRQ_MSIX);
> >          if (err < 0 || err != nvec) {
> >          	...
> > 	}
> > 
> > Remove the check which compares err with nvec.
> > 
> > Fixes: 1463f382f58d ("octeontx2-af: Add support for CGX link management")
> > Suggested-by: Paolo Abeni <pabeni@redhat.com>
> > Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
> > ---
> > Only compile tested.
> > ---
> >   drivers/net/ethernet/marvell/octeontx2/af/cgx.c | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
> > index d374a4454836..f4d5a3c05fa4 100644
> > --- a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
> > +++ b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
> > @@ -1993,7 +1993,7 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
> >   	nvec = pci_msix_vec_count(cgx->pdev);
> >   	err = pci_alloc_irq_vectors(pdev, nvec, nvec, PCI_IRQ_MSIX);
> > -	if (err < 0 || err != nvec) {
> > +	if (err < 0) {
> >   		dev_err(dev, "Request for %d msix vectors failed, err %d\n",
> >   			nvec, err);
> 
> 
> Now that I think about it more, maybe we want to error out when err != nvec
> as well ? In that case maybe the right thing to do is leave the check as is
> and set err = -EXYZ before goto ?
> 
> Thanks,
> Harshit>   		goto err_release_regions;

Hi Harshit,

My reading of the documentation of pci_alloc_irq_vectors() is that
Because nvec is passed as both the min and max desired vectors,
either nvecs will be allocated, or -ENOSPC will be returned.

Maybe it is worth adding a comment about that to patch description
or the code. But I think the code in this patch is correct.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] Octeontx2-af: Fix pci_alloc_irq_vectors() return value check
  2025-10-15  8:36   ` Simon Horman
@ 2025-10-15  8:56     ` Harshit Mogalapalli
  0 siblings, 0 replies; 4+ messages in thread
From: Harshit Mogalapalli @ 2025-10-15  8:56 UTC (permalink / raw)
  To: Simon Horman
  Cc: Sunil Goutham, Linu Cherian, Geetha sowjanya, Jerin Jacob,
	hariprasad, Subbaraya Sundeep, Andrew Lunn, David S. Miller,
	Eric Dumazet, Jakub Kicinski, Paolo Abeni, Nithya Mani, netdev,
	linux-kernel, dan.carpenter, kernel-janitors, error27

Hi Simon,
On 15/10/25 14:06, Simon Horman wrote:
> On Tue, Oct 14, 2025 at 03:50:47PM +0530, Harshit Mogalapalli wrote:
>> Hi,
>>
>> On 14/10/25 15:44, Harshit Mogalapalli wrote:
>>> In cgx_probe() when pci_alloc_irq_vectors() fails the error value will
>>> be negative and that check is sufficient.
>>>
>>> 	err = pci_alloc_irq_vectors(pdev, nvec, nvec, PCI_IRQ_MSIX);
>>>           if (err < 0 || err != nvec) {
>>>           	...
>>> 	}
>>>
>>> Remove the check which compares err with nvec.
>>>
>>> Fixes: 1463f382f58d ("octeontx2-af: Add support for CGX link management")
>>> Suggested-by: Paolo Abeni <pabeni@redhat.com>
>>> Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
>>> ---
>>> Only compile tested.
>>> ---
>>>    drivers/net/ethernet/marvell/octeontx2/af/cgx.c | 2 +-
>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
>>> index d374a4454836..f4d5a3c05fa4 100644
>>> --- a/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
>>> +++ b/drivers/net/ethernet/marvell/octeontx2/af/cgx.c
>>> @@ -1993,7 +1993,7 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
>>>    	nvec = pci_msix_vec_count(cgx->pdev);
>>>    	err = pci_alloc_irq_vectors(pdev, nvec, nvec, PCI_IRQ_MSIX);
>>> -	if (err < 0 || err != nvec) {
>>> +	if (err < 0) {
>>>    		dev_err(dev, "Request for %d msix vectors failed, err %d\n",
>>>    			nvec, err);
>>
>>
>> Now that I think about it more, maybe we want to error out when err != nvec
>> as well ? In that case maybe the right thing to do is leave the check as is
>> and set err = -EXYZ before goto ?
>>
>> Thanks,
>> Harshit>   		goto err_release_regions;
> 
> Hi Harshit,
> 
> My reading of the documentation of pci_alloc_irq_vectors() is that
> Because nvec is passed as both the min and max desired vectors,
> either nvecs will be allocated, or -ENOSPC will be returned.
> 

Ah thats a nice summary of the correctness of this patch.> Maybe it is 
worth adding a comment about that to patch description
> or the code. But I think the code in this patch is correct.

Definitely will add this and send v2.

Thanks,
Harshit

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2025-10-15  8:57 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-10-14 10:14 [PATCH] Octeontx2-af: Fix pci_alloc_irq_vectors() return value check Harshit Mogalapalli
2025-10-14 10:20 ` Harshit Mogalapalli
2025-10-15  8:36   ` Simon Horman
2025-10-15  8:56     ` Harshit Mogalapalli

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).