[PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Ilpo Järvinen]

The commit 0238f352a63a ("PCI/bwctrl: Replace lbms_count with
PCI_LINK_LBMS_SEEN flag") remove all code related to
pcie_bwctrl_lbms_rwsem but forgot to remove the rwsem itself.
Remove it and the associated info from the comment now.

Fixes: 0238f352a63a ("PCI/bwctrl: Replace lbms_count with PCI_LINK_LBMS_SEEN flag")
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
---

Bjorn, this should be folded into the original commit I think.

 drivers/pci/pcie/bwctrl.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)

diff --git a/drivers/pci/pcie/bwctrl.c b/drivers/pci/pcie/bwctrl.c
index fdafa20e4587..f31fbbd51490 100644
--- a/drivers/pci/pcie/bwctrl.c
+++ b/drivers/pci/pcie/bwctrl.c
@@ -45,15 +45,7 @@ struct pcie_bwctrl_data {
 	struct thermal_cooling_device *cdev;
 };
 
-/*
- * Prevent port removal during LBMS count accessors and Link Speed changes.
- *
- * These have to be differentiated because pcie_bwctrl_change_speed() calls
- * pcie_retrain_link() which uses LBMS count reset accessor on success
- * (using just one rwsem triggers "possible recursive locking detected"
- * warning).
- */
-static DECLARE_RWSEM(pcie_bwctrl_lbms_rwsem);
+/* Prevent port removal during Link Speed changes. */
 static DECLARE_RWSEM(pcie_bwctrl_setspeed_rwsem);
 
 static bool pcie_valid_speed(enum pci_bus_speed speed)

base-commit: 0238f352a63a075ac1f35ea565b5bec3057ec8bd
-- 
2.39.5

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Paul E. McKenney]

On Thu, May 08, 2025 at 12:00:36PM +0300, Ilpo Järvinen wrote:
> The commit 0238f352a63a ("PCI/bwctrl: Replace lbms_count with
> PCI_LINK_LBMS_SEEN flag") remove all code related to
> pcie_bwctrl_lbms_rwsem but forgot to remove the rwsem itself.
> Remove it and the associated info from the comment now.
> 
> Fixes: 0238f352a63a ("PCI/bwctrl: Replace lbms_count with PCI_LINK_LBMS_SEEN flag")
> Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>

Tested-by: Paul E. McKenney <paulmck@kernel.org>

> ---
> 
> Bjorn, this should be folded into the original commit I think.
> 
>  drivers/pci/pcie/bwctrl.c | 10 +---------
>  1 file changed, 1 insertion(+), 9 deletions(-)
> 
> diff --git a/drivers/pci/pcie/bwctrl.c b/drivers/pci/pcie/bwctrl.c
> index fdafa20e4587..f31fbbd51490 100644
> --- a/drivers/pci/pcie/bwctrl.c
> +++ b/drivers/pci/pcie/bwctrl.c
> @@ -45,15 +45,7 @@ struct pcie_bwctrl_data {
>  	struct thermal_cooling_device *cdev;
>  };
>  
> -/*
> - * Prevent port removal during LBMS count accessors and Link Speed changes.
> - *
> - * These have to be differentiated because pcie_bwctrl_change_speed() calls
> - * pcie_retrain_link() which uses LBMS count reset accessor on success
> - * (using just one rwsem triggers "possible recursive locking detected"
> - * warning).
> - */
> -static DECLARE_RWSEM(pcie_bwctrl_lbms_rwsem);
> +/* Prevent port removal during Link Speed changes. */
>  static DECLARE_RWSEM(pcie_bwctrl_setspeed_rwsem);
>  
>  static bool pcie_valid_speed(enum pci_bus_speed speed)
> 
> base-commit: 0238f352a63a075ac1f35ea565b5bec3057ec8bd
> -- 
> 2.39.5
> 

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Krzysztof Wilczyński]

Hello,

> The commit 0238f352a63a ("PCI/bwctrl: Replace lbms_count with
> PCI_LINK_LBMS_SEEN flag") remove all code related to
> pcie_bwctrl_lbms_rwsem but forgot to remove the rwsem itself.
> Remove it and the associated info from the comment now.
> 
> 

Applied to bwctrl, thank you!

[1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem
      https://git.kernel.org/pci/pci/c/256ab8a30905

	Krzysztof

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Lukas Wunner]

On Wed, May 14, 2025 at 05:28:33PM +0000, Krzysztof Wilczy??ski wrote:
> > The commit 0238f352a63a ("PCI/bwctrl: Replace lbms_count with
> > PCI_LINK_LBMS_SEEN flag") remove all code related to
> > pcie_bwctrl_lbms_rwsem but forgot to remove the rwsem itself.
> > Remove it and the associated info from the comment now.
> > 
> > 
> 
> Applied to bwctrl, thank you!
> 
> [1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem
>       https://git.kernel.org/pci/pci/c/256ab8a30905

This is now an individual commit on the bwctrl branch, but Ilpo
requested to squash it with the other commit already on that branch...

   "Bjorn, this should be folded into the original commit I think."

...because that other commit breaks the build:

https://lore.kernel.org/r/3840f086-91cf-4fec-8004-b272a21d86cf@paulmck-laptop/

Thanks,

Lukas

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Krzysztof Wilczyński]

[...]
> > Applied to bwctrl, thank you!
> > 
> > [1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem
> >       https://git.kernel.org/pci/pci/c/256ab8a30905
> 
> This is now an individual commit on the bwctrl branch, but Ilpo
> requested to squash it with the other commit already on that branch...
> 
>    "Bjorn, this should be folded into the original commit I think."
> 
> ...because that other commit breaks the build:
> 
> https://lore.kernel.org/r/3840f086-91cf-4fec-8004-b272a21d86cf@paulmck-laptop/

Done.  Squashed with the first commit from Ilpo, see:

  https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git/commit/?h=bwctrl&id=2389d8dc38fee18176c49e9c4804f5ecc55807fa

Let me know if there is anything else needed.

Thank you!

	Krzysztof

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Lukas Wunner]

On Thu, May 15, 2025 at 05:43:46PM +0900, Krzysztof Wilczy??ski wrote:
> Done.  Squashed with the first commit from Ilpo, see:
> 
>   https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git/commit/?h=bwctrl&id=2389d8dc38fee18176c49e9c4804f5ecc55807fa

Awesome, thank you!

> Let me know if there is anything else needed.

Actually, two small things:

- That patch on the pci/bwctrl topic branch is still marked "New"
  in patchwork, even though it's been applied:
  https://patchwork.kernel.org/project/linux-pci/patch/20250422115548.1483-1-ilpo.jarvinen@linux.intel.com/

- Version 1 of the same patch is likewise marked "New", even though
  it's been superseded:
  https://patchwork.kernel.org/project/linux-pci/patch/20250417124633.11470-1-ilpo.jarvinen@linux.intel.com/

Unfortunately I can't update it myself because I'm not the submitter.
(Ilpo could do it if he has a patchwork.kernel.org account.)


Something unrelated (only if you feel like doing it):

On the pci/enumeration branch, Bjorn queued up a revert which was
waiting to be ack'ed by AMD IOMMU maintainers:
https://lore.kernel.org/r/20250425163259.GA546441@bhelgaas/

In the meantime the ack has arrived:
https://lore.kernel.org/r/aCLv7cN_s1Z4abEl@8bytes.org/

So the remaining housekeeping items are:

- Add Jörg's Acked-by to commit e86c7278eba8 on pci/enumeration
- Remove the "XXX" marker from the subject line
- Remove "Needs AMD IOMMU ack" from the commit message

Again, only if you feel like doing it.  No urgency.

Thanks,

Lukas

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 2279 bytes --]

On Thu, 15 May 2025, Lukas Wunner wrote:

> On Thu, May 15, 2025 at 05:43:46PM +0900, Krzysztof Wilczy??ski wrote:
> > Done.  Squashed with the first commit from Ilpo, see:

Thanks Krzysztof for handling this, I should have put the note about 
squashing it to the resubmission but I forgot (this time I didn't do 
the diff against the previous version before sending it which I normally 
do).

> >   https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git/commit/?h=bwctrl&id=2389d8dc38fee18176c49e9c4804f5ecc55807fa
> 
> Awesome, thank you!
> 
> > Let me know if there is anything else needed.
> 
> Actually, two small things:
> 
> - That patch on the pci/bwctrl topic branch is still marked "New"
>   in patchwork, even though it's been applied:
>   https://patchwork.kernel.org/project/linux-pci/patch/20250422115548.1483-1-ilpo.jarvinen@linux.intel.com/
> 
> - Version 1 of the same patch is likewise marked "New", even though
>   it's been superseded:
>   https://patchwork.kernel.org/project/linux-pci/patch/20250417124633.11470-1-ilpo.jarvinen@linux.intel.com/
> 
> Unfortunately I can't update it myself because I'm not the submitter.
> (Ilpo could do it if he has a patchwork.kernel.org account.)

I'm a pdx86 maintainer so I do have an account, yes. I actually had the 
patchwork page listing my PCI patches already open, but I just hadn't hit 
"update" button yet.

I've done those two changes now.

I'm a bit hesitant to mark "Accepted" state though, I did it this time 
but in general I feel I might be overstepping my authority even if I know 
some patches have been accepted.

> Something unrelated (only if you feel like doing it):
> 
> On the pci/enumeration branch, Bjorn queued up a revert which was
> waiting to be ack'ed by AMD IOMMU maintainers:
> https://lore.kernel.org/r/20250425163259.GA546441@bhelgaas/
> 
> In the meantime the ack has arrived:
> https://lore.kernel.org/r/aCLv7cN_s1Z4abEl@8bytes.org/
> 
> So the remaining housekeeping items are:
> 
> - Add Jörg's Acked-by to commit e86c7278eba8 on pci/enumeration
> - Remove the "XXX" marker from the subject line
> - Remove "Needs AMD IOMMU ack" from the commit message
> 
> Again, only if you feel like doing it.  No urgency.


-- 
 i.

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Lukas Wunner]

On Thu, May 15, 2025 at 03:21:39PM +0300, Ilpo Järvinen wrote:
> I'm a bit hesitant to mark "Accepted" state though, I did it this time 
> but in general I feel I might be overstepping my authority even if I know 
> some patches have been accepted.

Bjorn has encouraged submitters to mark their own patches as "Superseded":

   "If you're really gung-ho, you can go to Patchwork [2] and mark
    your superseded patches as "Superseded" so I don't have to do that
    myself."

    https://lore.kernel.org/r/20171026223701.GA25649@bhelgaas-glaptop.roam.corp.google.com/

... and I was assuming that also applies to marking one's own patches
as "Accepted", but I might be jumping to the wrong conclusion.

Thanks,

Lukas

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Krzysztof Wilczy??ski]

Hello,

[...]
> > Done.  Squashed with the first commit from Ilpo, see:
> > 
> >   https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git/commit/?h=bwctrl&id=2389d8dc38fee18176c49e9c4804f5ecc55807fa
> 
> Awesome, thank you!
> 
> > Let me know if there is anything else needed.
> 
> Actually, two small things:
> 
> - That patch on the pci/bwctrl topic branch is still marked "New"
>   in patchwork, even though it's been applied:
>   https://patchwork.kernel.org/project/linux-pci/patch/20250422115548.1483-1-ilpo.jarvinen@linux.intel.com/
> 
> - Version 1 of the same patch is likewise marked "New", even though
>   it's been superseded:
>   https://patchwork.kernel.org/project/linux-pci/patch/20250417124633.11470-1-ilpo.jarvinen@linux.intel.com/
> 
> Unfortunately I can't update it myself because I'm not the submitter.
> (Ilpo could do it if he has a patchwork.kernel.org account.)

I had a look and I think Ilpo beat me to it.  Thank you Ilpo!

Also, thank you for chasing after this.  Much appreciated.  Good to know
people use Patchwork to track what is up with our tree. :)

> Something unrelated (only if you feel like doing it):
> 
> On the pci/enumeration branch, Bjorn queued up a revert which was
> waiting to be ack'ed by AMD IOMMU maintainers:
> https://lore.kernel.org/r/20250425163259.GA546441@bhelgaas/
> 
> In the meantime the ack has arrived:
> https://lore.kernel.org/r/aCLv7cN_s1Z4abEl@8bytes.org/
> 
> So the remaining housekeeping items are:
> 
> - Add Jörg's Acked-by to commit e86c7278eba8 on pci/enumeration
> - Remove the "XXX" marker from the subject line
> - Remove "Needs AMD IOMMU ack" from the commit message

Done.  Have a look to make sure that things look correct now:

  https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git/commit/?h=enumeration&id=3be5fa236649da6404f1bca1491bf02d4b0d5cce

> Again, only if you feel like doing it.  No urgency.

No problem!  Any time. :)

Thank you!

	Krzysztof

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Krzysztof Wilczy??ski]

Hello,

[...]
> > > Done.  Squashed with the first commit from Ilpo, see:
> 
> Thanks Krzysztof for handling this, I should have put the note about 
> squashing it to the resubmission but I forgot (this time I didn't do 
> the diff against the previous version before sending it which I normally 
> do).

No problem!  I am sorry for missing the ask in the first place.

[...]
> > > Let me know if there is anything else needed.
> > 
> > Actually, two small things:
> > 
> > - That patch on the pci/bwctrl topic branch is still marked "New"
> >   in patchwork, even though it's been applied:
> >   https://patchwork.kernel.org/project/linux-pci/patch/20250422115548.1483-1-ilpo.jarvinen@linux.intel.com/
> > 
> > - Version 1 of the same patch is likewise marked "New", even though
> >   it's been superseded:
> >   https://patchwork.kernel.org/project/linux-pci/patch/20250417124633.11470-1-ilpo.jarvinen@linux.intel.com/
> > 
> > Unfortunately I can't update it myself because I'm not the submitter.
> > (Ilpo could do it if he has a patchwork.kernel.org account.)
> 
> I'm a pdx86 maintainer so I do have an account, yes. I actually had the 
> patchwork page listing my PCI patches already open, but I just hadn't hit 
> "update" button yet.
> 
> I've done those two changes now.

Thank you!

	Krzysztof

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Krzysztof Wilczy??ski]

Hello,

> > I'm a bit hesitant to mark "Accepted" state though, I did it this time 
> > but in general I feel I might be overstepping my authority even if I know 
> > some patches have been accepted.
> 
> Bjorn has encouraged submitters to mark their own patches as "Superseded":
> 
>    "If you're really gung-ho, you can go to Patchwork [2] and mark
>     your superseded patches as "Superseded" so I don't have to do that
>     myself."
> 
>     https://lore.kernel.org/r/20171026223701.GA25649@bhelgaas-glaptop.roam.corp.google.com/
> 
> ... and I was assuming that also applies to marking one's own patches
> as "Accepted", but I might be jumping to the wrong conclusion.

Between Bjorn, Lorenzo, Mani and I, we are getting better as leveraging
Patchwork for a little bit of tracking and "project management", so to
speak.

But, any help with the housekeeping always appreciated. :)

Thank you!

	Krzysztof

Re: [PATCH 1/1] PCI/bwctrl: Remove also pcie_bwctrl_lbms_rwsem

[Krzysztof Wilczy??ski]

Hello,

[...]
> But, any help with the housekeeping always appreciated. :)

Speaking of housekeeping...

Lukas, do you prefer US ASCII when sending e-mails?

I've noticed that the letter in my surname got broken here. It's also
interesting that my "ń" broke, but Ilpo's "ä" didn't.  I assumed both
would be encoded as UTF-8.

Anyway, just curious. :)

	Krzysztof