When it rains it pours

When it rains it pours

[Limonciello, Mario]

Hi Bjorn,

For the _REG change that went into Linus' tree I was recently made aware 
of another system that it helps.

This system was appearing to hang during bootup which evaluating the 
USB4 _OSC.
This hang happened on both the 6.1 LTS kernel and 6.4 final kernel.

In looking at the BIOS debug log shared by the reporter I noticed that
the kernel isn't hung it's just that the BIOS was waiting to be given 
the ability to access the config space.

Backporting just that _REG patch onto 6.1 LTS kernel fixes the issue.

I'm encouraging the BIOS team to try to come up with a cleaner failure 
path for the lack of _REG being called.  However there is always the 
possibility they can't or choose not to and people try to boot older 
kernels and fail.

Given how severe this boot issue is compared to the original suspend 
issue that prompted the patch I wanted to gauge how you feel about the 
risk of taking this change back to stable.

Thanks!

Re: When it rains it pours

[Bjorn Helgaas]

On Fri, Jun 30, 2023 at 02:30:56PM -0500, Limonciello, Mario wrote:
> Hi Bjorn,
> 
> For the _REG change that went into Linus' tree I was recently made aware of
> another system that it helps.
> 
> This system was appearing to hang during bootup which evaluating the USB4
> _OSC.
> This hang happened on both the 6.1 LTS kernel and 6.4 final kernel.
> 
> In looking at the BIOS debug log shared by the reporter I noticed that
> the kernel isn't hung it's just that the BIOS was waiting to be given the
> ability to access the config space.
> 
> Backporting just that _REG patch onto 6.1 LTS kernel fixes the issue.
> 
> I'm encouraging the BIOS team to try to come up with a cleaner failure path
> for the lack of _REG being called.  However there is always the possibility
> they can't or choose not to and people try to boot older kernels and fail.
> 
> Given how severe this boot issue is compared to the original suspend issue
> that prompted the patch I wanted to gauge how you feel about the risk of
> taking this change back to stable.

I think we can do that.  But the patch is already in the pull request
for v6.5, so we'll have to wait until Linus pulls it and then ask the
stable folks to pick it up.  I don't think it should be a big deal; we
just need a mainline SHA1 for it.

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/stable-kernel-rules.rst?id=v6.4#n64

Re: When it rains it pours

[Limonciello, Mario]

On 6/30/2023 14:39, Bjorn Helgaas wrote:
> On Fri, Jun 30, 2023 at 02:30:56PM -0500, Limonciello, Mario wrote:
>> Hi Bjorn,
>>
>> For the _REG change that went into Linus' tree I was recently made aware of
>> another system that it helps.
>>
>> This system was appearing to hang during bootup which evaluating the USB4
>> _OSC.
>> This hang happened on both the 6.1 LTS kernel and 6.4 final kernel.
>>
>> In looking at the BIOS debug log shared by the reporter I noticed that
>> the kernel isn't hung it's just that the BIOS was waiting to be given the
>> ability to access the config space.
>>
>> Backporting just that _REG patch onto 6.1 LTS kernel fixes the issue.
>>
>> I'm encouraging the BIOS team to try to come up with a cleaner failure path
>> for the lack of _REG being called.  However there is always the possibility
>> they can't or choose not to and people try to boot older kernels and fail.
>>
>> Given how severe this boot issue is compared to the original suspend issue
>> that prompted the patch I wanted to gauge how you feel about the risk of
>> taking this change back to stable.
> 
> I think we can do that.  But the patch is already in the pull request
> for v6.5, so we'll have to wait until Linus pulls it and then ask the
> stable folks to pick it up.  I don't think it should be a big deal; we
> just need a mainline SHA1 for it.
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/stable-kernel-rules.rst?id=v6.4#n64

OK thanks, I saw your PR was sent out but I didn't realize it wasn't 
picked yet.

I'll keep an eye out for when the SHA1 is in Linus' tree and I'll 
request it for stable when it is.