Date: Tue, 15 Nov 2016 16:38:19 -0500
From: Steven Rostedt <rostedt@goodmis.org>
To: Luiz Capitulino <lcapitulino@redhat.com>
Cc: linux-kernel@vger.kernel.org
Subject: Re: [PATCH 3/3] trace-cmd record: add --cpu-list option
Message-ID: <20161115163819.5291b752@gandalf.local.home>
In-Reply-To: <20161028171456.3643d626@redhat.com>
References: <1475858831-32687-1-git-send-email-lcapitulino@redhat.com>
        <1475858831-32687-4-git-send-email-lcapitulino@redhat.com>
        <20161028154922.64e97208@gandalf.local.home>
        <20161028171456.3643d626@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 677
Lines: 25

On Fri, 28 Oct 2016 17:14:56 -0400
Luiz Capitulino <lcapitulino@redhat.com> wrote:


> > 
> > cpu_set(ret_mask, i) sets bits from begin to end in uint64_t cpumask
> > from below.
> > 
> > What happens if this range is greater than 64? We already have boxes
> > that run this with 80 CPUs. Looks to be out of memory range to me.  
> 
> The best solution is probably to detect the number of CPUs at run-time
> and use the CPU_SET() API. The lazy and ugly solution is to just fail.
> 
> Any objections to the CPU_SET() solution?
> 

I thought I replied to this, but I'm guessing the reply is still
pending.

I have no objecting. I'd like to see a patch though.

Thanks,

-- Steve