mbox series

[0/5] mm: Stop alaising VM_FAULT_HINDEX_MASK in arch code

Message ID 20221203030356.3917-1-palmer@rivosinc.com (mailing list archive)
Headers show
Series mm: Stop alaising VM_FAULT_HINDEX_MASK in arch code | expand

Message

Palmer Dabbelt Dec. 3, 2022, 3:03 a.m. UTC 
When reviewing
<https://lore.kernel.org/r/20221029084715.3669204-1-tongtiangen@huawei.com>
I noticed that the arch-specific VM_FAULT flags used by arm and s390
alias with VM_FAULT_HINDEX_MASK.  I'm not sure if it's possible to
manifest this as a bug, but it certainly seems fragile.

The RISC-V patch is on top of the linked patch above, which isn't in any
proper tree yet.  Everything else should apply to 6.1-rc1, but I'm in no
particular rush to get that cleanup in so rush on my end for 6.2.  That
said I figured it would be easier to send along this now, just in case
someone who knows the MM code better thinks it can manifest as a bug.

Comments

Palmer Dabbelt Feb. 3, 2023, 3:31 a.m. UTC | #1 
On Fri, 02 Dec 2022 19:03:51 PST (-0800), Palmer Dabbelt wrote:
> When reviewing
> <https://lore.kernel.org/r/20221029084715.3669204-1-tongtiangen@huawei.com>
> I noticed that the arch-specific VM_FAULT flags used by arm and s390
> alias with VM_FAULT_HINDEX_MASK.  I'm not sure if it's possible to
> manifest this as a bug, but it certainly seems fragile.
>
> The RISC-V patch is on top of the linked patch above, which isn't in any
> proper tree yet.  Everything else should apply to 6.1-rc1, but I'm in no
> particular rush to get that cleanup in so rush on my end for 6.2.  That
> said I figured it would be easier to send along this now, just in case
> someone who knows the MM code better thinks it can manifest as a bug.

Just pinging this one, I'm not sure if it was just bad timing on the 
last one.  That RISC-V patch is still sitting around outside my tree.