[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <e248c5c8-e55c-46eb-ae8d-f7539cdbeab2@infradead.org>
Date: Sun, 17 Sep 2023 08:06:51 -0700
From: Randy Dunlap <rdunlap@...radead.org>
To: Max Filippov <jcmvbkbc@...il.com>,
LKML <linux-kernel@...r.kernel.org>,
Linux Memory Management List <linux-mm@...ck.org>
Cc: Matthew Wilcox <willy@...radead.org>, Yu Zhao <yuzhao@...gle.com>,
"Liam R. Howlett" <Liam.Howlett@...cle.com>,
Vlastimil Babka <vbabka@...e.cz>,
Davidlohr Bueso <dave@...olabs.net>,
Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: /proc/*/maps broken on nommu
On 9/17/23 04:38, Max Filippov wrote:
> Hello,
>
> As of v6.5 cat /proc/*/maps produces empty output on nommu xtensa systems.
> Bisection points to the commit 0c563f148043 ("proc: remove VMA rbtree
> use from nommu").
> It looks like the issue may be not in the m_start (where find_vma(mm,
> addr) returns
> NULL), but in the nommu version of find_vma itself. I've tried replacing nommu
> find_vma implementation with an implementation of the MMU version and it
> fixes the /proc/*/maps output. Kind of (it works as expected on one of
> my systems,
> but consistently locks up on the other).
>
> Any suggestions of what might be wrong/how to fix it properly?
>
Hi Max,
There is already a proposed patch and some discussion:
https://lore.kernel.org/lkml/20230915160055.971059-2-ben.wolsieffer@hefring.com/
--
~Randy
Powered by blists - more mailing lists