Re: potential malloc problem with recent ATA commits ?

From: Pascal Hofstee <caelian_at_gmail.com>
Date: Wed, 16 Apr 2008 22:26:21 +0200
On Wed, 16 Apr 2008 22:00:53 +0200
Pascal Hofstee <caelian_at_gmail.com> wrote:

> After some suggestions on ##FreeBSD on FreeNode i opted to checkout
> "vmstat -m" statistics and i did notice one single entry that i felt i
> should mention just in case it may be relevant.
> 
> The entry in question is the vmstat's "InUse" statistics regarding the
> memorytype "devbuf". It's value seems to consistently increase itself
> by about 140k over a period of 10 seconds (while the system is doing a
> buildworld) and has at the time of writing reached a value of 75082k
> and rising.
> 
> I do not know If this is alarming or not but i felt i should at least
> mention it.

Damned ... I Hate it when i have to keep replying to my own emails
within 10 minutes after sending them.

It looks like i was able to trace down the source of the steadily
increasing amount of "devbuf" usage. I opted to simply shut down any
port-installed daemons to make sure there was no software running
besides the FreeBSD base-system (and a zsh).

And through that i managed to find out that the process that seems to
be responsible for the steady increase in "devbuf" usage is "hald".

As soon as i stopped hald the amount of "InUse" memory of type "devbuf"
remained stable over more than a minute. As soon as i started hald
again the increase resumed by 140k every 10 seconds again.

I'll try to leave the system running now with hald turned off to see if
the memory problem persists (after a fresh reboot devbuf usage seems
solid at 4254k).

As a precautionary measure i'll probably try to take this hall issue to
the FreeBSD-GNOME team to see what they have to say about this :)

I'll keep the list posted on any further details, but it looks like
this may have been a false alarm.

With kind regards,
  Pascal Hofstee
Received on Wed Apr 16 2008 - 18:26:30 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:29 UTC