Have been doing additional testing of these - and also verified that David's most recent patch, now in my for-next branch ("cifs: Fix memory leak in direct I/O") fixes the remaining problem (the issue with xfstest 208 that Murphy pointed out). Will send the three small cifs fixes from David later this week, along with some unrelated fixes from Paulo and Shyam.
-----Original Message-----
From: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Sent: Friday, February 24, 2023 1:06 PM
To: David Howells <dhowells@xxxxxxxxxx>
Cc: Steven French <Steven.French@xxxxxxxxxxxxx>; Vishal Moola <vishal.moola@xxxxxxxxx>; Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>; Jan Kara <jack@xxxxxxx>; pc <pc@xxxxxx>; Matthew Wilcox <willy@xxxxxxxxxxxxx>; Huang Ying <ying.huang@xxxxxxxxx>; Baolin Wang <baolin.wang@xxxxxxxxxxxxxxxxx>; Xin Hao <xhao@xxxxxxxxxxxxxxxxx>; linux-mm@xxxxxxxxx; mm-commits@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
Subject: [EXTERNAL] Re: [RFC][PATCH] cifs: Fix cifs_writepages_region()
[You don't often get email from torvalds@xxxxxxxxxxxxxxxxxxxx. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
On Fri, Feb 24, 2023 at 10:58 AM Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> I'll apply that minimal fix for now - I think the right thing to do is
> your bigger patch, but that needs more thinking (or at least splitting
> up).
Minimal fix applied - that way I can drop this for now, and we can discuss the whole "maybe we can just use write_cache_pages()" instead.
Because that _would_ be lovely, even if it's possible that the generic helper might need some extra love to work better for cifs/afs.
Linus
