[PATCH v3 03/24] arm: Remove COMMAND_LINE_SIZE from uapi

[Palmer Dabbelt]

On Thu, 23 Feb 2023 05:09:17 PST (-0800), Arnd Bergmann wrote:
> On Thu, Feb 23, 2023, at 10:54, Alexandre Ghiti wrote:
>> On Wed, Feb 15, 2023 at 2:05 PM Arnd Bergmann <arnd at arndb.de> wrote:
>>>
>>> On Wed, Feb 15, 2023, at 13:59, Russell King (Oracle) wrote:
>>> > On Tue, Feb 14, 2023 at 08:49:04AM +0100, Alexandre Ghiti wrote:
>>> >> From: Palmer Dabbelt <palmer at rivosinc.com>
>>> >>
>>> >> As far as I can tell this is not used by userspace and thus should not
>>> >> be part of the user-visible API.
>>> >>
>>> >> Signed-off-by: Palmer Dabbelt <palmer at rivosinc.com>
>>> >
>>> > Looks good to me. What's the merge plan for this?
>>>
>>> The easiest way is probably if I merge it through the whole
>>> series through the asm-generic tree. The timing is a bit
>>> unfortunate as we're just ahead of the merge window, so unless
>>> we really need this in 6.3, I'd suggest that Alexandre resend
>>> the series to me in two weeks with the Acks added in and I'll
>>> pick it up for 6.4.
>>
>> Sorry for the response delay, I was waiting to see if Palmer would
>> merge my KASAN patchset in 6.3 (which he does): I have to admit that
>> fixing the command line size + the KASAN patchset would allow 6.3 to
>> run on syzkaller, which would be nice.
>>
>> If I don't see this merged in 6.3, I'll send another round as you
>> suggested in 1 week now :)
>
> Hi Alexandre,
>
> I have no plans to still pick up the series for 6.3. The patches
> all look fine to me, but it's clearly too late now. What is the
> actual dependency for KASAN, do you just need a longer command
> line or something else? If it's just the command line size,
> I would suggest that Palmer can still pick up a oneline change
> to increase it and refer to this thread in the changelog as a
> reference for why it is not an actual UAPI break.

Sorry for being slow here, I just queued up the original patch in the 
RISC-V tree and intend on sending it for 6.3 -- the main worry was that 
it's a uABi change and we're confident it's not.  It's late, but I'd 
prefer to have this as it should let us start running syzkaller now and 
that'll probably find more bugs than this is likely to trigger.

https://lore.kernel.org/r/mhng-b5f934ff-a9bb-4c2b-9ba6-3ab68312077a@palmer-ri-x1c9a/