[APS-24624] Missing patch in K5.15 against kernel panic

[APS-24624] Missing patch in K5.15 against kernel panic

[Bilge Aydin]

Dear Linux community,

I am leading a project at Samsung where we utilized your kernel 5.15 in our system. Recently we faced an important kernel panic issue whose fix must be following patch:
https://lore.kernel.org/all/1946ef9f774851732eed78760a78ec40dbc6d178.1667591503.git.robin.murphy@arm.com/

The problem for us is that this patch must be available from kernel 6.1 onwards. 
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/drivers/iommu/iommu.c?h=v6.1.119 
Clear to us is that it is not integrated into K5.15 yet: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/drivers/iommu/iommu.c?h=v5.15.173 

On the other side we can see that this patch was already applied to Android Common Kernel android13-5.15  https://android-review.googlesource.com/q/I51dbcdc5dc536b0e69c6187b2d7ac6a2031a305b. In particular, we can see the same implementation in the recent ACK release https://android.googlesource.com/kernel/common/+/refs/tags/android13-5.15-2024-11_r2/drivers/iommu/iommu.c  given on https://source.android.com/docs/core/architecture/kernel/gki-android13-5_15-release-builds 

Do you have any plan to adjust your implementation in drivers/iommu/iommu.c in K5.15 as present in K6.1? If not, since the issue in our project is very critical for us, may I ask you kindly to trigger the sync of K5.15 and K6.1 soon?

I would appreciate highly your response and any support.


Kind regards / Mit freundlichen Grüßen

-Confidential-
Bilge Aydin
Project Manager
Phone:: +49-89-45578-1055 · Mobile:: +49-160-8855875 · Email: b.aydin@samsung.com 
Samsung Semiconductor Europe GmbH
Einsteinstrasse 174, 81677 München, Deutschland
Jurisdiction and registered Munich / Germany, HRB Nr. 253778
Managing Director: Dermot Ryan

www.samsung.com/semiconductor/solutions/automotive




This communication (including any attachments) contains information which may be confidential and/or privileged. It is for the exclusive use of the intended recipient(s) so, if you have received this communication in error, please do not distribute, copy or use this communication or its contents but notify the sender immediately and then destroy any copies of it. Due to the nature of the Internet, the sender is unable to ensure the integrity of this message. Any views expressed in this communication are those of the individual sender, except where the sender specifically states them to be the views of the company the sender works for.

 

Re: [APS-24624] Missing patch in K5.15 against kernel panic

[Greg KH]

On Tue, Dec 10, 2024 at 02:23:35PM +0100, Bilge Aydin wrote:
> Dear Linux community,
> 
> I am leading a project at Samsung where we utilized your kernel 5.15 in our system. Recently we faced an important kernel panic issue whose fix must be following patch:
> https://lore.kernel.org/all/1946ef9f774851732eed78760a78ec40dbc6d178.1667591503.git.robin.murphy@arm.com/
> 
> The problem for us is that this patch must be available from kernel 6.1 onwards. 
> https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/drivers/iommu/iommu.c?h=v6.1.119 
> Clear to us is that it is not integrated into K5.15 yet: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/drivers/iommu/iommu.c?h=v5.15.173 
> 
> On the other side we can see that this patch was already applied to Android Common Kernel android13-5.15  https://android-review.googlesource.com/q/I51dbcdc5dc536b0e69c6187b2d7ac6a2031a305b. In particular, we can see the same implementation in the recent ACK release https://android.googlesource.com/kernel/common/+/refs/tags/android13-5.15-2024-11_r2/drivers/iommu/iommu.c  given on https://source.android.com/docs/core/architecture/kernel/gki-android13-5_15-release-builds 
> 
> Do you have any plan to adjust your implementation in drivers/iommu/iommu.c in K5.15 as present in K6.1? If not, since the issue in our project is very critical for us, may I ask you kindly to trigger the sync of K5.15 and K6.1 soon?

If you wish for a commit to be backported to the stable kernels, just
ask us, and better yet, provide a working patch to do so.

The reason we did NOT backport this specific commit is that it is marked
as a fix for a commit that shows up in the 6.1 release, so why would it
be applicable to 5.15.y at all?  So why do you feel it is required in
5.15.y?  (Note, we have no idea why it's in an Android kernel tree,
perhaps they have the offending commit in there as well so the backport
is required?  Have you checked?)

If you need or want it applied to older kernels, please send us a
working patch and we will be glad to take it.

thanks,

greg k-h

RE: [APS-24624] Missing patch in K5.15 against kernel panic

[Bilge Aydin]

Hello Greg,

Thank you very much for your prompt reply.

"So why do you feel it is required in 5.15.y?"
--> We have faced the issue in our K5.15 based platform. Our technical team has made an investigation and found your given patch. Afterwards we verified it by running a stress test. The result was positive and we did not face the kernel panic again.

Since we are not OS experts, we fully respect Linux community's decisions. That's why we do not see ourselves in that high position to give any guidance and share any patch with you. During our investigation, we only recognized that the given patch was already integrated into the Android common kernel. Outgoing from that, I just want to share the indication with you.

Would it be possible for you to start an internal communication within the Linux community for the synchronization of the K5.15 and K6.1 mainlines explicitly for iommu driver?


Best regards
Bilge




-----Original Message-----
From: Greg KH <gregkh@linuxfoundation.org> 
Sent: Tuesday, December 10, 2024 2:33 PM
To: Bilge Aydin <b.aydin@samsung.com>
Cc: stable@vger.kernel.org
Subject: Re: [APS-24624] Missing patch in K5.15 against kernel panic

On Tue, Dec 10, 2024 at 02:23:35PM +0100, Bilge Aydin wrote:
> Dear Linux community,
> 
> I am leading a project at Samsung where we utilized your kernel 5.15 in our system. Recently we faced an important kernel panic issue whose fix must be following patch:
> https://lore.kernel.org/all/1946ef9f774851732eed78760a78ec40dbc6d178.1
> 667591503.git.robin.murphy@arm.com/
> 
> The problem for us is that this patch must be available from kernel 6.1 onwards. 
> https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/
> drivers/iommu/iommu.c?h=v6.1.119 Clear to us is that it is not 
> integrated into K5.15 yet: 
> https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/
> drivers/iommu/iommu.c?h=v5.15.173
> 
> On the other side we can see that this patch was already applied to 
> Android Common Kernel android13-5.15  
> https://android-review.googlesource.com/q/I51dbcdc5dc536b0e69c6187b2d7
> ac6a2031a305b. In particular, we can see the same implementation in 
> the recent ACK release 
> https://android.googlesource.com/kernel/common/+/refs/tags/android13-5
> .15-2024-11_r2/drivers/iommu/iommu.c  given on 
> https://source.android.com/docs/core/architecture/kernel/gki-android13
> -5_15-release-builds
> 
> Do you have any plan to adjust your implementation in drivers/iommu/iommu.c in K5.15 as present in K6.1? If not, since the issue in our project is very critical for us, may I ask you kindly to trigger the sync of K5.15 and K6.1 soon?

If you wish for a commit to be backported to the stable kernels, just ask us, and better yet, provide a working patch to do so.

The reason we did NOT backport this specific commit is that it is marked as a fix for a commit that shows up in the 6.1 release, so why would it be applicable to 5.15.y at all?  So why do you feel it is required in 5.15.y?  (Note, we have no idea why it's in an Android kernel tree, perhaps they have the offending commit in there as well so the backport is required?  Have you checked?)

If you need or want it applied to older kernels, please send us a working patch and we will be glad to take it.

thanks,

greg k-h

Re: [APS-24624] Missing patch in K5.15 against kernel panic

['Greg KH']

On Tue, Dec 10, 2024 at 02:48:52PM +0100, Bilge Aydin wrote:
> Hello Greg,
> 
> Thank you very much for your prompt reply.
> 
> "So why do you feel it is required in 5.15.y?"
> --> We have faced the issue in our K5.15 based platform. Our technical team has made an investigation and found your given patch. Afterwards we verified it by running a stress test. The result was positive and we did not face the kernel panic again.

That's great, then send us a backported patch (and the fix for that, you
did catch that one, right?)

> Since we are not OS experts, we fully respect Linux community's decisions. That's why we do not see ourselves in that high position to give any guidance and share any patch with you. During our investigation, we only recognized that the given patch was already integrated into the Android common kernel. Outgoing from that, I just want to share the indication with you.
> 
> Would it be possible for you to start an internal communication within the Linux community for the synchronization of the K5.15 and K6.1 mainlines explicitly for iommu driver?

That is not how the stable kernel branches work at all, sorry.  Please
read:
    https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
for more information.

Also, if you need the iommu code in 6.1, then just use the 6.1.y kernel
please.  There's nothing keeping you from doing that, right?  That way
you get more support for more hardware and more fixes that didn't happen
to get backported to the really old 5.15.y release.  Also, you will have
to move off of 5.15.y in a few years anyway, so might as well do it now
when you have a very good reason to do so.

good luck!

greg k-h

Re: [APS-24624] Missing patch in K5.15 against kernel panic

[Greg KH]

On Tue, Dec 10, 2024 at 02:23:35PM +0100, Bilge Aydin wrote:
> This communication (including any attachments) contains information which may be confidential and/or privileged. It is for the exclusive use of the intended recipient(s) so, if you have received this communication in error, please do not distribute, copy or use this communication or its contents but notify the sender immediately and then destroy any copies of it. Due to the nature of the Internet, the sender is unable to ensure the integrity of this message. Any views expressed in this communication are those of the individual sender, except where the sender specifically states them to be the views of the company the sender works for.

Note, this footer is not compatible with kernel development at all, and
normally causes emails to be automatically deleted.  For some reason it
snuck through my filters, but for many others, it might not have.
Please remove it on your responses.

thanks,

greg k-h