linux-kernel - Re: [PATCH] selftests: breakpoints: Add "WINE" test for x86

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>] [day] [month] [year] [list]

Message-ID: <875z3fd0g4.fsf@nanos.tec.linutronix.de>
Date:   Fri, 29 Jan 2021 20:09:47 +0100
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     x86@...nel.org, tdevries@...e.de, linux-kernel@...r.kernel.org,
        andrew.cooper3@...rix.com, Frederic Weisbecker <fweisbec@...il.com>
Subject: Re: [PATCH] selftests: breakpoints: Add "WINE" test for x86

On Fri, Jan 29 2021 at 10:08, Peter Zijlstra wrote:
> On Fri, Jan 29, 2021 at 12:28:41AM +0100, Thomas Gleixner wrote:
>
>> Add a test case which covers this scenario. This is modeled after the
>> WINE testcase, but changes the expect in step #2 to:
>> 
>>    - Expect DR6::BS == 1 and DR6::BR0 == 1 and IP == second instruction
>> 
>> to ensure that the GDB expectations are met as well.
>
>> +	/*
>> +	 * Expect: DR6::BS == 1 DR6::BR0 == 1  IP == instr[1]
>> +	 * Wine does not care about BR0 here but GDB does ...
>> +	 */
>> +	wine_test_step(1, addr, 1, 1, buf);
>
>
> So my v2 patch will fail this, while it will pass the actual gdb
> testcase.
>
> The thing it does is process _data_ breakpoints along with TF, but it
> will exclude instruction breakpoints and TF.
>
> Since the above test is using instruction breakpoints, it will report
> 0x4000 and 0x0001 respectively for two consequtive exceptions.

Yes, I'm a moron....