Re: [PATCH] fs/resctrl: Convert lock to guard

[Reinette Chatre <reinette.chatre@intel.com>]

On 9/3/25 7:37 AM, Dave Martin wrote:
> Hi,
> 
> You seem already to have been told that patches of this sort are
> unlikely to be helpful.  Greg summed it up pretty well, e.g. [1], [2].
> 
> 
> With regard to this specific patch:
> 
> On Wed, Sep 03, 2025 at 03:20:15PM +0300, Erick Karanja wrote:
>> Convert manual lock/unlock calls to guard and tidy up the code.
> 
> The very first sentence under "Describe your changes"
> in submitting-patches.rst is: "Describe your problem."
> 
> So, what problem is being addressed by this patch?
> 
>> Generated-by: Coccinelle SmPL
>> Signed-off-by: Erick Karanja <karanja99erick@gmail.com>
> 
> Coccinelle is a powerful tool, but ultimately somebody has to take
> responsibility for the correctness of the output.  How have you
> verified that the result is correct?
> 
> Also, this is a lot of diffstat for what is really just a change of
> coding style that fixes nothing (or at least, you don't claim that it
> fixes anything).
> 
> While all contributions are welcome, they do need to deliver a net
> benefit.

Well said. Thanks Dave.

> 
> I can't speak for the maintainers, but given the pain that they would
> likely have fixing up all the merge conflicts that this patch would
> cause, I think that the benefit would need to be substantial.

There just happened to be a similar change suggested for SGX that provides
insight into x86 maintainers' view that I fully agree with:
https://lore.kernel.org/lkml/d9a76e90-7723-49ee-b3ec-85c7533d8023@intel.com/

> 
> 
> If you found actual bugs as part of this process, then that would
> definitely be worth looking at (?)

ack.

Reinette