Re: + syscallh-add-syscall_set_arguments-fix.patch added to mm-new branch

[Nathan Chancellor <nathan@kernel.org>]

On Thu, Apr 10, 2025 at 09:06:34PM -0700, Andrew Morton wrote:
> On Thu, 10 Apr 2025 19:47:26 -0700 Nathan Chancellor <nathan@kernel.org> wrote:
> 
> > Hi Andrew,
> > 
> > On Thu, Apr 10, 2025 at 06:43:03PM -0700, Andrew Morton wrote:
> > > 
> > > The patch titled
> > >      Subject: riscv: avoid fortify warning in syscall_get_arguments()
> > > has been added to the -mm mm-new branch.  Its filename is
> > >      syscallh-add-syscall_set_arguments-fix.patch
> > ...
> > > ------------------------------------------------------
> > > From: Nathan Chancellor <nathan@kernel.org>
> > > Subject: riscv: avoid fortify warning in syscall_get_arguments()
> > 
> > I do not think I made myself clear enough in my request when reading it
> > back, so I do apologize for that. I need the diff at the end of [1]
> > applied to syscallh-add-syscall_set_arguments.patch, not [2], which is a
> > standalone patch for the existing syscall_get_arguments() that should go
> > via the RISC-V tree because it is an existing issue (or at least be kept
> > in standalone patch form).
> > 
> > [1]: https://lore.kernel.org/20250408213131.GA2872426@ax162/
> 
> This fixed syscall_set_arguments()_ only.
> 
> > [2]: https://lore.kernel.org/20250409-riscv-avoid-fortify-warning-syscall_get_arguments-v1-1-7853436d4755@kernel.org/
> 
> Whereas https://lkml.kernel.org/r/20250409003803.GA2876360@ax162 fixed
> syscall_get_arguments() also.

Yes, I merely meant for that to be a demonstration of what resolves both
fortify warnings.

> I queued both fixes as separate patches.  And I reverted your mail
> client's conversion of tabs to spaces and I added signoffs and all the
> usual whatever.
> 
> Please check mm-new once I get it pushed out.

The fix for syscall_set_arguments() looks properly done in mm-unstable
but I do not understand why the fix for syscall_get_arguments() is
marked as another fix to be squashed into the patch that adds
syscall_set_arguments()? It is an issue that is present in Linus's tree
currently, not one that appears after Dmitry's patch. Just drop the
patch that fixes syscall_get_arguments(), it is my understanding that
the RISC-V folks will handle it (or if you are adamant about carrying
it, keep it as a standalone patch, not one to be squashed). I just need
syscall_set_arguments() fixed by you because you are carrying the patch
that adds it per the plan at
https://lore.kernel.org/20250409155207.GA1506425@ax162/

Cheers,
Nathan