Date: Thu, 10 Jan 2008 07:33:03 +0200 From: Kostik Belousov <kostikbel@gmail.com> To: "J.R. Oldroyd" <fbsd@opal.com> Cc: Kris Kennaway <kris@freebsd.org>, freebsd-stable@freebsd.org Subject: Re: 7.0BETA4 desktop system also periodically freezes Message-ID: <20080110053303.GC57756@deviant.kiev.zoral.com.ua> In-Reply-To: <20080109235819.7707133d@linwhf.opal.com> References: <4783D41B.3000204@FreeBSD.org> <4783D748.1050401@raad.tartu.ee> <4783D824.1050502@FreeBSD.org> <4783DB72.6030605@raad.tartu.ee> <4783DCAA.1080108@FreeBSD.org> <4783EA7E.1050008@raad.tartu.ee> <4783ED15.1060606@FreeBSD.org> <20080109141859.74d08743@linwhf.opal.com> <478522B5.5030101@FreeBSD.org> <20080109235819.7707133d@linwhf.opal.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--U06nEXQpEf10f9fB
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
On Wed, Jan 09, 2008 at 11:58:19PM -0500, J.R. Oldroyd wrote:
> On Wed, 09 Jan 2008 20:38:29 +0100, Kris Kennaway <kris@freebsd.org> wrot=
e:
> >
> >=20
> > OK, same requests as to the others then.
> >=20
>=20
> I presume you mean hwpmc...
>=20
> In setting that up, I may have stumbled upon a possible cause.
> I ran pmcstat for a while and ended up with a very large output
> file. Having not noticed any freezes during that time, I decided
> to start over with the output to a different filesystem with more
> free space. When I removed that first output file, the resulting
> disk i/o for the block deallocation caused a very similar freeze
> for a few seconds. And when it came back, the load average graph
> peaked up, just as before.
>=20
> This appears to be repeatable:
>=20
> 1. create a very large file, say 1Gb or so
> 2. remove it
> 3. observe freeze during file dealloc i/o activity
>=20
> This on an ata drive, ufs filesystem with softupdates.
>=20
> Folk complaining of jerky mouse syndrome (e.g., during compilations)
> may be seeing the same - the compiler creates and removes lots of
> files.
>=20
> I'm not sure how this would explain those longer freezes (30s or
> several minutes) though.
For the SU-imposed jerkiness, try the patch below. Progress on developing
the change stalled somehow, and I would like to process with it.
diff --git a/sys/kern/vfs_mount.c b/sys/kern/vfs_mount.c
index 248ce22..005f915 100644
--- a/sys/kern/vfs_mount.c
+++ b/sys/kern/vfs_mount.c
@@ -2001,6 +2001,12 @@ __mnt_vnode_next(struct vnode **mvp, struct mount *m=
p)
mtx_assert(MNT_MTX(mp), MA_OWNED);
=20
KASSERT((*mvp)->v_mount =3D=3D mp, ("marker vnode mount list mismatch"));
+ if ((*mvp)->v_yield++ =3D=3D 500) {
+ MNT_IUNLOCK(mp);
+ (*mvp)->v_yield =3D 0;
+ uio_yield();
+ MNT_ILOCK(mp);
+ }
vp =3D TAILQ_NEXT(*mvp, v_nmntvnodes);
while (vp !=3D NULL && vp->v_type =3D=3D VMARKER)
vp =3D TAILQ_NEXT(vp, v_nmntvnodes);
diff --git a/sys/sys/vnode.h b/sys/sys/vnode.h
index 0525a43..1ac289a 100644
--- a/sys/sys/vnode.h
+++ b/sys/sys/vnode.h
@@ -131,6 +131,7 @@ struct vnode {
struct socket *vu_socket; /* v unix domain net (VSOCK) */
struct cdev *vu_cdev; /* v device (VCHR, VBLK) */
struct fifoinfo *vu_fifoinfo; /* v fifo (VFIFO) */
+ int vu_yield; /* yield count (VMARKER) */
} v_un;
=20
/*
@@ -185,6 +186,7 @@ struct vnode {
#define v_socket v_un.vu_socket
#define v_rdev v_un.vu_cdev
#define v_fifoinfo v_un.vu_fifoinfo
+#define v_yield v_un.vu_yield
=20
/* XXX: These are temporary to avoid a source sweep at this time */
#define v_object v_bufobj.bo_object
diff --git a/sys/ufs/ffs/ffs_softdep.c b/sys/ufs/ffs/ffs_softdep.c
index a221b66..02cf45b 100644
--- a/sys/ufs/ffs/ffs_softdep.c
+++ b/sys/ufs/ffs/ffs_softdep.c
@@ -864,6 +864,7 @@ softdep_process_worklist(mp, full)
*/
if (loopcount++ % 128 =3D=3D 0) {
FREE_LOCK(&lk);
+ uio_yield();
bwillwrite();
ACQUIRE_LOCK(&lk);
}
--U06nEXQpEf10f9fB
Content-Type: application/pgp-signature
Content-Disposition: inline
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (FreeBSD)
iD8DBQFHha4OC3+MBN1Mb4gRAuJqAJ9ddrTwSREYVsRzoDZOYqWT+YCubgCgnCc7
ma8oolIVP5ccUHdi6nZsQd4=
=j8hC
-----END PGP SIGNATURE-----
--U06nEXQpEf10f9fB--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20080110053303.GC57756>