Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1621828 > unrolled thread
| Started by | Christoph Hellwig <hch@lst.de> |
|---|---|
| First post | 2017-04-12 08:10 +0200 |
| Last post | 2017-04-13 11:00 +0200 |
| Articles | 6 — 3 participants |
Back to article view | Back to linux.kernel
please use pci_alloc_irq_vectors in thunderx_edac. Christoph Hellwig <hch@lst.de> - 2017-04-12 08:10 +0200
Re: please use pci_alloc_irq_vectors in thunderx_edac. Borislav Petkov <bp@alien8.de> - 2017-04-12 11:30 +0200
Re: please use pci_alloc_irq_vectors in thunderx_edac. Christoph Hellwig <hch@lst.de> - 2017-04-12 19:20 +0200
Re: please use pci_alloc_irq_vectors in thunderx_edac. Borislav Petkov <bp@alien8.de> - 2017-04-12 19:30 +0200
Re: please use pci_alloc_irq_vectors in thunderx_edac. Christoph Hellwig <hch@lst.de> - 2017-04-12 19:40 +0200
Re: please use pci_alloc_irq_vectors in thunderx_edac. Johannes Thumshirn <jthumshirn@suse.de> - 2017-04-13 11:00 +0200
| From | Christoph Hellwig <hch@lst.de> |
|---|---|
| Date | 2017-04-12 08:10 +0200 |
| Subject | please use pci_alloc_irq_vectors in thunderx_edac. |
| Message-ID | <tvjDB-Xu-25@gated-at.bofh.it> |
Hi Sergey, you're new thunderx_edac driver uses pci_enable_msix_exact for no good reason. Can you please switch it to use pci_alloc_irq_vectors before it gets merged to mainline? Thanks, Christoph
[toc] | [next] | [standalone]
| From | Borislav Petkov <bp@alien8.de> |
|---|---|
| Date | 2017-04-12 11:30 +0200 |
| Message-ID | <tvmL8-2Jw-11@gated-at.bofh.it> |
| In reply to | #1621828 |
On Wed, Apr 12, 2017 at 08:06:00AM +0200, Christoph Hellwig wrote:
> Hi Sergey,
>
> you're new thunderx_edac driver uses pci_enable_msix_exact for no
> good reason. Can you please switch it to use pci_alloc_irq_vectors
> before it gets merged to mainline?
Do we need to mark pci_enable_msix_exact() deprecated or so, or maybe a
checkpatch rule?
It says so in some text file in Documentation:
4.3 Legacy APIs
The following old APIs to enable and disable MSI or MSI-X interrupts should
not be used in new code:
pci_enable_msi() /* deprecated */
pci_disable_msi() /* deprecated */
pci_enable_msix_range() /* deprecated */
pci_enable_msix_exact() /* deprecated */
pci_disable_msix() /* deprecated */
but maybe we should enforce it?
--
Regards/Gruss,
Boris.
Good mailing practices for 400: avoid top-posting and trim the reply.
[toc] | [prev] | [next] | [standalone]
| From | Christoph Hellwig <hch@lst.de> |
|---|---|
| Date | 2017-04-12 19:20 +0200 |
| Message-ID | <tvu5Y-7gS-9@gated-at.bofh.it> |
| In reply to | #1621956 |
On Wed, Apr 12, 2017 at 11:21:01AM +0200, Borislav Petkov wrote: > Do we need to mark pci_enable_msix_exact() deprecated or so, or maybe a > checkpatch rule? Did we ever do this for deprecated APIs before? For now I'm just watching Linux-next and sending gentle reminders :)
[toc] | [prev] | [next] | [standalone]
| From | Borislav Petkov <bp@alien8.de> |
|---|---|
| Date | 2017-04-12 19:30 +0200 |
| Message-ID | <tvufE-7kr-27@gated-at.bofh.it> |
| In reply to | #1622387 |
On Wed, Apr 12, 2017 at 07:16:22PM +0200, Christoph Hellwig wrote:
> Did we ever do this for deprecated APIs before? For now I'm just watching
> Linux-next and sending gentle reminders :)
$ git grep __deprecated
shows some have been done. I wonder whether marking
pci_enable_msix_exact() would cause a lot of warnings though. An
allyesconfig build would tell...
--
Regards/Gruss,
Boris.
Good mailing practices for 400: avoid top-posting and trim the reply.
[toc] | [prev] | [next] | [standalone]
| From | Christoph Hellwig <hch@lst.de> |
|---|---|
| Date | 2017-04-12 19:40 +0200 |
| Message-ID | <tvupk-7nO-5@gated-at.bofh.it> |
| In reply to | #1622402 |
On Wed, Apr 12, 2017 at 07:28:59PM +0200, Borislav Petkov wrote: > On Wed, Apr 12, 2017 at 07:16:22PM +0200, Christoph Hellwig wrote: > > Did we ever do this for deprecated APIs before? For now I'm just watching > > Linux-next and sending gentle reminders :) > > $ git grep __deprecated > > shows some have been done. I wonder whether marking > pci_enable_msix_exact() would cause a lot of warnings though. An > allyesconfig build would tell... It would cause lots of warnings. The problem with __deprecated is that it causes warnings for existing users, not just new ones.
[toc] | [prev] | [next] | [standalone]
| From | Johannes Thumshirn <jthumshirn@suse.de> |
|---|---|
| Date | 2017-04-13 11:00 +0200 |
| Message-ID | <tvILD-sK-3@gated-at.bofh.it> |
| In reply to | #1622411 |
On Wed, Apr 12, 2017 at 07:30:22PM +0200, Christoph Hellwig wrote: > > shows some have been done. I wonder whether marking > > pci_enable_msix_exact() would cause a lot of warnings though. An > > allyesconfig build would tell... > > It would cause lots of warnings. The problem with __deprecated is that > it causes warnings for existing users, not just new ones. So a checkpatch rule is the way to go, I guess. -- Johannes Thumshirn Storage jthumshirn@suse.de +49 911 74053 689 SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg GF: Felix Imendörffer, Jane Smithard, Graham Norton HRB 21284 (AG Nürnberg) Key fingerprint = EC38 9CAB C2C4 F25D 8600 D0D0 0393 969D 2D76 0850
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web