[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20201027091504.712183781@infradead.org>
Date: Tue, 27 Oct 2020 10:15:04 +0100
From: Peter Zijlstra <peterz@...radead.org>
To: tglx@...utronix.de, luto@...nel.org, me@...ehuey.com
Cc: x86@...nel.org, linux-kernel@...r.kernel.org,
torvalds@...ux-foundation.org, rocallahan@...il.com,
alexandre.chartre@...cle.com, paulmck@...nel.org,
frederic@...nel.org, pbonzini@...hat.com,
sean.j.christopherson@...el.com, mhiramat@...nel.org,
pmladek@...e.com, joel@...lfernandes.org, rostedt@...dmis.org,
boris.ostrovsky@...cle.com, jgross@...e.com, brgerst@...il.com,
jpoimboe@...hat.com, daniel.thompson@...aro.org,
julliard@...ehq.org, pgofman@...eweavers.com, peterz@...radead.org
Subject: [PATCH 0/3] x86/debug: Fixes
Hi,
Triggered by the x86/entry rework, the resulting #DB cleanup (obviously :/)
broke something. Kyle reported that ptrace_get_debugreg(6) no longer contained
DR_STEP after PTRACE_SINGLESTEP which broke RR.
While looking at this, I realized that a kernel #DB should not consume a
userspace BTF, and equally a kernel #DB should not clobber the (userspace)
ptrace DR6 state. Both these have been busted since forever afaict.
I've added a few Wine folks to Cc, with the hope that they can test Wine
on 5.10-rc and make sure it all still works as expected. There have been
significant changes. Although hopefully it all works again after these
patches.
Powered by blists - more mailing lists