Re: [PATCH 6.6 000/166] 6.6.84-rc1 review

[Jens Axboe <axboe@kernel.dk>]

On 3/19/25 5:51 PM, Greg Kroah-Hartman wrote:
> On Wed, Mar 19, 2025 at 10:37:20AM -0600, Jens Axboe wrote:
>> On 3/19/25 10:33 AM, Naresh Kamboju wrote:
>>> On Wed, 19 Mar 2025 at 20:09, Greg Kroah-Hartman
>>> <gregkh@linuxfoundation.org> wrote:
>>>>
>>>> This is the start of the stable review cycle for the 6.6.84 release.
>>>> There are 166 patches in this series, all will be posted as a response
>>>> to this one.  If anyone has any issues with these being applied, please
>>>> let me know.
>>>>
>>>> Responses should be made by Fri, 21 Mar 2025 14:29:55 +0000.
>>>> Anything received after that time might be too late.
>>>>
>>>> The whole patch series can be found in one patch at:
>>>>         https://www.kernel.org/pub/linux/kernel/v6.x/stable-review/patch-6.6.84-rc1.gz
>>>> or in the git tree and branch at:
>>>>         git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-6.6.y
>>>> and the diffstat can be found below.
>>>>
>>>> thanks,
>>>>
>>>> greg k-h
>>>
>>> Regressions on mips the rt305x_defconfig builds failed with gcc-12
>>> the stable-rc v6.6.83-167-gd16a828e7b09
>>>
>>> First seen on the v6.6.83-167-gd16a828e7b09
>>>  Good: v6.6.83
>>>  Bad: v6.6.83-167-gd16a828e7b09
>>>
>>> * mips, build
>>>   - gcc-12-rt305x_defconfig
>>>
>>> Regression Analysis:
>>>  - New regression? Yes
>>>  - Reproducibility? Yes
>>>
>>> Build regression: mips implicit declaration of function 'vunmap'
>>> Reported-by: Linux Kernel Functional Testing <lkft@linaro.org>
>>
>> Ah that's my fault, forgot to include the backport of:
>>
>> commit 62346c6cb28b043f2a6e95337d9081ec0b37b5f5
>> Author: Jens Axboe <axboe@kernel.dk>
>> Date:   Sat Mar 16 07:21:43 2024 -0600
>>
>>     mm: add nommu variant of vm_insert_pages()
>>
>> for 6.1-stable and 6.6-stable. Greg, can you just cherry pick that one?
>> It'll pick cleanly into both, should go before the io_uring mmap series
>> obviously.
>>
>> Sorry about that! I did have it in my local trees, but for some reason
>> forgot to include it in the sent in series.
> 
> Wait, this is already in the 6.6.y and 6.1.y queues, so this can't be
> the fix.  Was there a fixup for that commit somewhere else that I'm
> missing?

Huh indeed, guess I didn't mess up in the first place. What is going on
here indeed... Is that mips config NOMMU yet doesn't link in mm/nommu.o?

Checking, and no, it definitely has MMU=y in the config. Guess I
should've read the initial report more closely, it's simply missing the
vunmap definition. Adding linux/vmalloc.h to io_uring/io_uring.c should
fix it.

How do we want to deal with this?

diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index 058c89f72e8c..efa7849b82c1 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -70,6 +70,7 @@
 #include <linux/io_uring.h>
 #include <linux/audit.h>
 #include <linux/security.h>
+#include <linux/vmalloc.h>
 #include <asm/shmparam.h>
 
 #define CREATE_TRACE_POINTS

-- 
Jens Axboe