Re: [PATCH] fs/proc/task_mmu: fix missed VMAs when reading /proc/PID/maps and friends

[Jose A. Perez de Azpillaga]

On 10/25/26 11:43 AM, David Hildenbrand (Arm) wrote:
[snip]

> Fixes: 5631da56c9a8 ("fs/proc/task_mmu: read proc/pid/maps under per-vma lock")
> Fixes: ea3085dd7a4e ("fs/proc/task_mmu: read proc/pid/{smaps|numa_maps} under per-vma lock")
> Fixes: 0e985559d72a ("proc/task_mmu: read proc/pid/smaps_rollup under per-vma lock")

PROCMAP_QUERY does the same fresh lookup, shouldn't d9d1c2d81797
("fs/proc/task_mmu: execute PROCMAP_QUERY ioctl under per-vma locks") be
here as well? (commit tracked down by an LLM)

[snip]

> There is the issue with seqcount overflows that Lorenzo is tackling. It
> probably interacts here as well? Likely, but that's an independent issue
> to resolve.

the saved seq is reset on every m_start() and query, so a wrap would
need 2^31 write locks inside one read() or ioctl, I don't think it
matters in practice :)

I tested it on x86_64 in qemu with KASAN, lockdep and DEBUG_VM on top of
mm-unstable, the maps holes I hit there are gone. Also, the code LGTM:

Reviewed-by: Jose A. Perez de Azpillaga <azpijr@xxxxxxxxx>
Tested-by: Jose A. Perez de Azpillaga <azpijr@xxxxxxxxx>

-- 
cheers,
jose a. p-a