linux-kernel - [BUGFIX][PATCH] flush icache before set

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20070807172458.59931ef2.kamezawa.hiroyu@jp.fujitsu.com>
Date:	Tue, 7 Aug 2007 17:24:58 +0900
From:	KAMEZAWA Hiroyuki <kamezawa.hiroyu@...fujitsu.com>
To:	linux-kernel@...r.kernel.org
Cc:	linux-ia64@...r.kernel.org,
	"tony.luck@...el.com" <tony.luck@...el.com>,
	"Zoltan.Menyhart@...l.net" <Zoltan.Menyhart@...l.net>,
	"kamezawa.hiroyu@...fujitsu.com" <kamezawa.hiroyu@...fujitsu.com>
Subject: [BUGFIX][PATCH] flush icache before set_pte() in ia64 take7, [0/2]

This patch set is for fixing SIGILL trouble in NFS by adding
sync_icache_dcache() before set_pte(). (see patches for detail)
Tested with my NUMA box and works well.

Chanves V6 -> V7
- dropped add-on patches for cpu family handling.

For this post, I wrote an easy test for this SIGILL/NFS problem.
Please verify.

This is my result.
==
Distro: RHEL5.
CPU: Montecito 1.4GHz, 4 sockets, 8core.

NFS + Montecito + 2.6.23-rc2
23 SIGILLS in 1000 trial

ext3 + Montecito + 2.6.23-rc2
0 erros in 1000 trial

NFS + Itanium2(model2 revision2) + 2.6.23-rc2
0 errors in 1000 trial

NFS + Montecito + 2.6.23-rc2 + patch.
0 errors in 1000 trial
==

Thanks,
-Kame

Download attachment "nfstest.tgz" of type "application/x-gtar" (3980 bytes)