[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <8f9e2dd8-f83f-0685-5939-33e77d97a6c6@amazon.com>
Date: Fri, 28 Jun 2019 10:27:30 +0200
From: <samcacc@...zon.com>
To: Paolo Bonzini <pbonzini@...hat.com>,
Alexander Graf <graf@...zon.com>,
Sam Caccavale <samcacc@...zon.de>
CC: <samcaccavale@...il.com>, <nmanthey@...zon.de>,
<wipawel@...zon.de>, <dwmw@...zon.co.uk>, <mpohlack@...zon.de>,
<karahmed@...zon.de>, <andrew.cooper3@...rix.com>,
<JBeulich@...e.com>, <rkrcmar@...hat.com>, <tglx@...utronix.de>,
<mingo@...hat.com>, <bp@...en8.de>, <hpa@...or.com>,
<paullangton4@...il.com>, <x86@...nel.org>, <kvm@...r.kernel.org>,
<linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v3 4/5] Added build and install scripts
On 6/28/19 10:17 AM, Paolo Bonzini wrote:
> On 28/06/19 09:59, samcacc@...zon.com wrote:
>>> Surely if it's important to generate core dumps, it's not only important
>>> during installation, no?
>> Yep... missed this. I'll move it to run.sh right before alf-many is
>> invoked. It would be nice to not have to sudo but it seems the only
>> alternative is an envvar AFL_I_DONT_CARE_ABOUT_MISSING_CRASHES which
>> just ignores AFL's warning if your system isn't going to produce core
>> dumps (which will cause AFL to miss some crashes, as the name suggests).
>
> Can you do this only if /proc/sys/kernel/core_pattern starts with a pipe
> sign?
>
I think I'll just remove the `echo > ...core_pattern` step from both the
build.sh and run.sh scripts and instead document it as a setup step in
the README. This also sidesteps the sudo use.
- Sam
> Thanks,
>
> Paolo
>
Powered by blists - more mailing lists