NetBSD-Bugs archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: kern/57690: XEN pvh guest GENERIC: xennet0: no rx clusters (possible mbuf/pool leak)



Yes after some running time on a freshly startet VM I do

see freed MBUFs and pages.

Memory resource pool statistics
Name Size Requests Fail Releases Pgreq Pgrel Npage Hiwat Minpg Maxpg Idle xnfrx 4096 222214 17 222040 18628 18418 210 873 0 inf 36

There are also fails (17) here.

Looking at vmstat -m it could very well be ZFS is currently eating up almost all

kernel memory and I am running plain NetBSD10 without my local ZFS fixes.

Frank


On 11/10/23 15:00, Manuel Bouyer wrote:
The following reply was made to PR kern/57690; it has been noted by GNATS.

From: Manuel Bouyer <bouyer%antioche.eu.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: kern-bug-people%netbsd.org@localhost, gnats-admin%netbsd.org@localhost, netbsd-bugs%netbsd.org@localhost,
         kardel%netbsd.org@localhost
Subject: Re: kern/57690: XEN pvh guest GENERIC: xennet0: no rx clusters
  (possible mbuf/pool leak)
Date: Fri, 10 Nov 2023 14:58:22 +0100

  On Fri, Nov 10, 2023 at 01:50:01PM +0000, Frank Kardel wrote:
  > The following reply was made to PR kern/57690; it has been noted by GNATS.
  >
  > From: Frank Kardel <kardel%netbsd.org@localhost>
  > To: gnats-bugs%netbsd.org@localhost
  > Cc:
  > Subject: Re: kern/57690: XEN pvh guest GENERIC: xennet0: no rx clusters
  >  (possible mbuf/pool leak)
  > Date: Fri, 10 Nov 2023 14:48:33 +0100
  >
  >  Well, after running a bit with a freshly started kernel I do see
  >  releases after some more traffic, so that part does
  >  seem to work.
  >
  >  Memory resource pool statistics
  >  Name        Size Requests Fail Releases Pgreq Pgrel Npage Hiwat Minpg
  >  Maxpg Idle
  >  xnfrx       4096   203576    6   203234  9728  9358   370   873 0   inf   28
  >
  >
  >  Why there are allocation failures and reduced performance issues still
  >  needs to be examined.
maybe it's a general RAM shortage ? The allocations happens from interrupt
  context, so if there is no free page it can fail.
Both MGETHDR and pool_cache_get_paddr() from if_xennetrxbuf_cache fails, so
  it seems just that there is no free page at some point.
I do see there diagnostif printfs from time to time but the system
  recovers.
--
  Manuel Bouyer <bouyer%antioche.eu.org@localhost>
       NetBSD: 26 ans d'experience feront toujours la difference
  --



Home | Main Index | Thread Index | Old Index