linux-kernel - Re: perf kvm segfaults while reporting events

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

Message-ID: <20120620061423.GB3488@linux.vnet.ibm.com>
Date:	Wed, 20 Jun 2012 11:44:23 +0530
From:	Pradeep Kumar Surisetty <psuriset@...ux.vnet.ibm.com>
To:	David Ahern <dsahern@...il.com>
Cc:	LKML <linux-kernel@...r.kernel.org>, ryanh@...ux.vnet.ibm.com
Subject: Re: perf kvm segfaults while reporting events

* David Ahern <dsahern@...il.com> [2012-06-01 09:40:30]:

> On 6/1/12 7:56 AM, Pradeep Kumar Surisetty wrote:
> >Hello
> >
> >perf kvm generats segfaults while reporting events.
> >
> >host&  guest:  3.4.0-rc4
> >
> >1) Copied kallsyms, modules from guest to host.
> >
> >2) recorded  events on host and guest&  report it using below commands.
> >
> >
> >[root@...t]# ./perf kvm --host --guest --guestkallsyms=/tmp/guest-kallsyms --guestmodules=/tmp/guest-modules record -a -o perf.data
> >^C[ perf record: Woken up 1 times to write data ]
> >[ perf record: Captured and wrote 0.381 MB perf.data (~16659 samples) ]
> >
> >[root@...t]# ./perf kvm --host --guest --guestkallsyms=/tmp/guest-kallsyms --guestmodules=/tmp/guest-modules report -i perf.data --force
> >perf: Segmentation fault
> 
> The short of it is that perf-kvm needs some work.
> 
> More specifically, the guest 'machine' is not created on the report
> path for the above arguments. This will prevent the segfault:
> 
> diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
> index 93d355d..c8e1323 100644
> --- a/tools/perf/util/session.c
> +++ b/tools/perf/util/session.c
> @@ -918,6 +918,8 @@ static int perf_session_deliver_event(struct
> perf_session *session,
>         }
>         return tool->sample(tool, event, sample, evsel, machine);
>     case PERF_RECORD_MMAP:
> +       if (machine == NULL)
> +           return 0;


Ahhh perf kvm works fine if we record guest events for enough time. 
above patch works better, if some one doesnt record for adequate time. 


>         return tool->mmap(tool, event, sample, machine);
>     case PERF_RECORD_COMM:
>         return tool->comm(tool, event, sample, machine);
> 
> 
> The --guestmount option (which has its own problems) combined with
> specific pid is more stable . e.g.,
> 
> mkdir /tmp/guestmount/<vm-pid>
> sshfs root@<vm-ip>:/  /tmp/guestmount/<vm-pid> -o direct_io
> 
> perf kvm --host --guest --guestmount=/tmp/guestmount \
>   record -p <vm-pid>  -o /tmp/perf.data -- sleep 1
> 
> perf kvm --host --guest --guestmount=/tmp/guestmount  report -i
> /tmp/perf.data --force --stdio
> 
> If you use the -a argument on record with multiple VMs running
> without an entry in /tmp/guestmount you'll see:
> 
> Warning:
> 7 unprocessable samples recorded.
> Do you have a KVM guest running and not using 'perf kvm'?
> 
> David
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/