linux-kernel - Re: [BUG?] perf_event_open fails on synthetic event created through dynamic

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

Message-ID: <20250512102222.2361a57a@gandalf.local.home>
Date: Mon, 12 May 2025 10:22:22 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: Tomas Glozar <tglozar@...hat.com>
Cc: Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...hat.com>,
 Arnaldo Carvalho de Melo <acme@...nel.org>, Namhyung Kim
 <namhyung@...nel.org>, Masami Hiramatsu <mhiramat@...nel.org>, Mathieu
 Desnoyers <mathieu.desnoyers@...icios.com>, Linux Trace Kernel
 <linux-trace-kernel@...r.kernel.org>, linux-perf-users@...r.kernel.org,
 LKML <linux-kernel@...r.kernel.org>
Subject: Re: [BUG?] perf_event_open fails on synthetic event created through
 dynamic_events

On Mon, 12 May 2025 16:08:42 +0200
Tomas Glozar <tglozar@...hat.com> wrote:

> It seems that it is trying to access a struct tracepoint object whose
> func field is NULL. I have not tried attaching to dynamic events
> through perf before, so I don't know whether this is just
> unimplemented and poorly handled or a regression. For what it's worth,
> it also happens on CentOS Stream 9 kernel (5.14.0-578.el9).

I think it is unimplemented and poorly handled ;-)

If this is something you want, we can add the feature for it, but for this
release and stable kernels, as it's not a regression, the fix is to have it
not produce a warning and give something like "-EINVAL" back to the perf
system call.

-- Steve