Re: [PATCH v6 0/2] uio_hv_generic: Fix ring buffer sysfs creation path

[Naman Jain <namjain@linux.microsoft.com>]

On 5/2/2025 11:43 AM, Greg Kroah-Hartman wrote:
> On Fri, May 02, 2025 at 11:31:03AM +0530, Naman Jain wrote:
>>
>>
>> On 5/1/2025 9:35 PM, Greg Kroah-Hartman wrote:
>>> On Mon, Apr 28, 2025 at 02:37:22PM +0530, Naman Jain wrote:
>>>>
>>>>
>>>> On 4/25/2025 7:30 PM, Greg Kroah-Hartman wrote:
>>>>> On Thu, Apr 24, 2025 at 11:05:22AM +0530, Naman Jain wrote:
>>>>>> Hi,
>>>>>> This patch series aims to address the sysfs creation issue for the ring
>>>>>> buffer by reorganizing the code. Additionally, it updates the ring sysfs
>>>>>> size to accurately reflect the actual ring buffer size, rather than a
>>>>>> fixed static value.
>>>>>>
>>>>>> PFB change logs:
>>>>>>
>>>>>> Changes since v5:
>>>>>> https://lore.kernel.org/all/20250415164452.170239-1-namjain@linux.microsoft.com/
>>>>>> * Added Reviewed-By tags from Dexuan. Also, addressed minor comments in
>>>>>>      commit msg of both patches.
>>>>>> * Missed to remove check for "primary_channel->device_obj->channels_kset" in
>>>>>>      hv_create_ring_sysfs in earlier patch, as suggested by Michael. Did it
>>>>>>      now.
>>>>>> * Changed type for declaring bin_attrs due to changes introduced by
>>>>>>      commit 9bec944506fa ("sysfs: constify attribute_group::bin_attrs") which
>>>>>>      merged recently. Did not use bin_attrs_new since another change is in
>>>>>>      the queue to change usage of bin_attrs_new to bin_attrs
>>>>>>      (sysfs: finalize the constification of 'struct bin_attribute').
>>>>>
>>>>> Please fix up to apply cleanly without build warnings:
>>>>>
>>>>> drivers/hv/vmbus_drv.c:1893:15: error: initializing 'struct bin_attribute **' with an expression of type 'const struct bin_attribute *const[2]' discards qualifiers in nested pointer types [-Werror,-Wincompatible-pointer-types-discards-qualifiers]
>>>>>     1893 |         .bin_attrs = vmbus_chan_bin_attrs,
>>>>>          |                      ^~~~~~~~~~~~~~~~~~~~
>>>>> 1 error generated.
>>>>
>>>> Hi Greg,
>>>> I tried reproducing this error but could not see it. Should I rebase the
>>>> change to some other tree or use some specific config option, gcc version,
>>>> compilation flag etc.?
>>>>
>>>> I tried the following:
>>>> * Rebased to latest linux-next tip with below base commit:
>>>> 393d0c54cae31317deaa9043320c5fd9454deabc
>>>> * Regular compilation with gcc: make -j8
>>>> * extra flags:
>>>>     make -j8  EXTRA_CFLAGS="-Wall -O2"
>>>>     make -j8 EXTRA_CFLAGS="-Wincompatible-pointer-types-discards-qualifiers
>>>> -Werror"
>>>> * Tried gcc 11.4, 13.3
>>>> * Tried clang/LLVM with version 18.1.3 : make LLVM=1
>>>
>>> I tried this against my char-misc-linus branch (which is pretty much
>>> just 6.15.0-rc4 plus some iio patches), and it fails with that error
>>> above.
>>>
>>>> BTW I had to edit the type for bin_attrs as this change got merged recently:
>>>> 9bec944506fa ("sysfs: constify attribute_group::bin_attrs")
>>>>
>>>> diff --git a/include/linux/sysfs.h b/include/linux/sysfs.h
>>>> index 576b8b3c60af..f418aae4f113 100644
>>>> --- a/include/linux/sysfs.h
>>>> +++ b/include/linux/sysfs.h
>>>> @@ -107,7 +107,7 @@ struct attribute_group {
>>>>                                               int);
>>>>           struct attribute        **attrs;
>>>>           union {
>>>> -               struct bin_attribute            **bin_attrs;
>>>> +               const struct bin_attribute      *const *bin_attrs;
>>>>                   const struct bin_attribute      *const *bin_attrs_new;
>>>>           };
>>>>    };
>>>
>>> That commit is not in my char-misc branches, that's coming from
>>> somewhere else.
>>>
>>> thanks,
>>>
>>> greg k-h
>>
>> Hi Greg,
>>
>> I can send a patch based on char-misc/6.15.0-rc4 which does not have this
>> patch, but I am worried that it will cause compilation issues when your
>> branch is merged with linux-next since this change is already there in
>> linux-next. Do you want me to proceed with sending a patch on 6.15.0-rc4?
> 
> Yes, because you want this fix in 6.15-final, right?
> 
>> Here are more details of that patch:
>>
>> """
>> sysfs: constify attribute_group::bin_attrs
>> All users of this field have been migrated to bin_attrs_new.
>> It can now be constified.
>>
>> Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
>> Link: https://lore.kernel.org/r/20250313-sysfs-const-bin_attr-final-v2-2-96284e1e88ce@weissschuh.net
>> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>>
>> """
> 
> I know that patch, I will deal with that in linux-next when needed, you
> shouldn't be worrying about it.  I'm more concerned as to why your patch
> was not being tested against Linus's tree if you expected it to be in
> the latest release and backported everywhere as it you asked it to be.
> 
> thanks,
> 
> greg k-h

Sure, thanks. I'll rebase it to 6.15.0-rc4 and send the next version.

Regards,
Naman