[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <a970a44b-504a-a26c-af8e-c11ee8c8c2fd@physik.fu-berlin.de>
Date: Wed, 12 Aug 2020 08:28:19 +0200
From: John Paul Adrian Glaubitz <glaubitz@...sik.fu-berlin.de>
To: "Edgecombe, Rick P" <rick.p.edgecombe@...el.com>
Cc: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-mm@...ck.org" <linux-mm@...ck.org>,
"linux-ia64@...r.kernel.org" <linux-ia64@...r.kernel.org>,
"Luck, Tony" <tony.luck@...el.com>,
"frank.scheiner@....de" <frank.scheiner@....de>,
"jrtc27@...c27.com" <jrtc27@...c27.com>,
"slyfox@...too.org" <slyfox@...too.org>,
"matorola@...il.com" <matorola@...il.com>
Subject: Re: "mm/vmalloc: Add flag for freeing of special permsissions" broke
udev on ia64
Hi Rick!
On 8/12/20 6:50 AM, Edgecombe, Rick P wrote:
> On non-x86, this patchset shouldn't have changed much, just always
> flush the TLB and purge the vmalloc area in cases where previously this
> would only happen intermittently. Despite this, it also caused a
> problem on sparc. In the end the sparc issue turned out to be an issue
> with sparc's kernel range TLB flush implementation that was just now
> being hit earlier during boot instead of after the system had been up
> for some time. So I'm at first wondering if there might be something
> like that uncovered here.
Yeah, I am also suspecting a bug in the ia64-specific mm code which got
just unconvered by your changes.
> Unfortunately, it's unlikely that I can get access to ia64 hardware in
> the near term to debug. Would you be able to test some debug patches?
I can both test patches for you and arrange access to one of the Itanium
servers I manage if you want to test patches yourself. Both works fine.
> Also, can you post your config?
See attached.
Adrian
--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer - glaubitz@...ian.org
`. `' Freie Universitaet Berlin - glaubitz@...sik.fu-berlin.de
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913
View attachment "config-rx2600-5.1" of type "text/plain" (89985 bytes)
Powered by blists - more mailing lists