pcp
[Top] [All Lists]

Re: [pcp] pmatop crashes when it encounters pid >= 1000000 (#58)

To: performancecopilot/pcp <pcp@xxxxxxxxxxxxxxxxxx>
Subject: Re: [pcp] pmatop crashes when it encounters pid >= 1000000 (#58)
From: Lukas Berk <notifications@xxxxxxxxxx>
Date: Wed, 16 Dec 2015 08:17:48 -0800
Authentication-results: i-otc1esa.syd1.mail-filtering.com.au; spf=Pass smtp.mailfrom=noreply@xxxxxxxxxx; spf=None smtp.helo=postmaster@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Delivered-to: pcp@xxxxxxxxxxx
Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=github.com; s=pf2014; t=1450282668; bh=jlYwHA29Hjq0FxNXFCFeagNTXX2vM0WA4ufpfkfVauw=; h=From:Reply-To:To:In-Reply-To:References:Subject:List-ID: List-Archive:List-Post:List-Unsubscribe:From; b=ez4iAh3A98qQ2j8FVlxY4JTLaZNP+WZRiHnKVYjGXuqhyp/j7WertvT7klhVhJUwI 1y4tWkobxjnrZEUL1YShiaJBzyRdM35V/VRMqvgwzj6e4xGBGLnPwM6NpJnpY/7CPB Se1vwpNMd/wS0NKHwSgh2XkQbLJ0f9uVnh1QpA0o=
In-reply-to: <performancecopilot/pcp/issues/58@xxxxxxxxxx>
Ironport-phdr: 9a23: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
List-archive: https://github.com/performancecopilot/pcp
List-id: performancecopilot/pcp <pcp.performancecopilot.github.com>
List-post: <mailto:reply+00bd08b605470b81b0695fb2e735bd99a6f84e148a0693a492cf00000001128950ac92a169ce074d3659@reply.github.com>
List-unsubscribe: <mailto:unsub+00bd08b605470b81b0695fb2e735bd99a6f84e148a0693a492cf00000001128950ac92a169ce074d3659@reply.github.com>, <https://github.com/notifications/unsubscribe/AL0Ithzu5cLeBLPZfRGirT_nC7cGIqJpks5pQYYsgaJpZM4G2fcF>
References: <performancecopilot/pcp/issues/58@xxxxxxxxxx>
Reply-to: performancecopilot/pcp <reply+00bd08b605470b81b0695fb2e735bd99a6f84e148a0693a492cf00000001128950ac92a169ce074d3659@xxxxxxxxxxxxxxxx>

Hi,

Thanks for reporting this issue! I'm working to reproduce this now, however, I'm weary that the large PID might be a red herring here. From the output, it appears that the 'name' var is empty, which in the context of the update_task function:

strsep(&name, " "); /* remove process identifier prefix */
strncpy(task->gen.cmdline, name, CMDLEN);

Means, the return of strsep would be NULL, and name would be empty (running through this with gdb and

set name = ""

lead to a similar SEGV. Regardless of whether or not the large PID is causing this issue (which I'm still looking into), it appears we need to be more careful with our assumptions of the name being passed to update_task.


Reply to this email directly or view it on GitHub.

<Prev in Thread] Current Thread [Next in Thread>