Le 28/12/2020 à 04:10, Defang Bo a écrit :
> Similar to commit<0dc294f717d4>("powerpc/mm: bail out early when flushing TLB
> page"),
> there should be a check for 'mm' to prevent Null pointer dereference
> in case of 'mm' argument was legitimately passed.
I don't understand what you are trying to do here.
In commit 0dc294f717d4, a call with a NULL mm is expected. The commit
just reorganises the code in order to bail out earlier and more explicitely.
In your patch, you are adding WARN_ON, which means that being called
with a NULL mm is unexpected. Why do you think it is necessary to do it
? Have you encountered any problem ?
Thanks
Christophe
>
> Signed-off-by: Defang Bo <bodef...@126.com>
> ---
> arch/powerpc/mm/nohash/tlb.c | 11 ++++++++++-
> 1 file changed, 10 insertions(+), 1 deletion(-)
>
> diff --git a/arch/powerpc/mm/nohash/tlb.c b/arch/powerpc/mm/nohash/tlb.c
> index 5872f69..1d89335 100644
> --- a/arch/powerpc/mm/nohash/tlb.c
> +++ b/arch/powerpc/mm/nohash/tlb.c
> @@ -192,6 +192,9 @@ void local_flush_tlb_mm(struct mm_struct *mm)
> {
> unsigned int pid;
>
> + if (WARN_ON(!mm))
> + return;
> +
> preempt_disable();
> pid = mm->context.id;
> if (pid != MMU_NO_CONTEXT)
> @@ -205,8 +208,11 @@ void __local_flush_tlb_page(struct mm_struct *mm,
> unsigned long vmaddr,
> {
> unsigned int pid;
>
> + if (WARN_ON(!mm))
> + return;
> +
> preempt_disable();
> - pid = mm ? mm->context.id : 0;
> + pid = mm->context.id;
> if (pid != MMU_NO_CONTEXT)
> _tlbil_va(vmaddr, pid, tsize, ind);
> preempt_enable();
> @@ -268,6 +274,9 @@ void flush_tlb_mm(struct mm_struct *mm)
> {
> unsigned int pid;
>
> + if (WARN_ON(!mm))
> + return;
> +
> preempt_disable();
> pid = mm->context.id;
> if (unlikely(pid == MMU_NO_CONTEXT))
