Re: NFS client/buffer cache deadlock

2005-04-23 Thread Garrett Wollman
< said: > Can you find any evidence that it's acceptable to interleave multiple > writers that are doing O_APPEND? At best, to do what you're asking, > they could be kept from being interleaved from the context of one > specific NFS client host... As far as POSIX goes, the standard says that ap

Re: NFS client/buffer cache deadlock

2005-04-22 Thread Brian Fundakowski Feldman
On Fri, Apr 22, 2005 at 11:28:15AM -0400, Garrett Wollman wrote: > < PROTECTED]> said: > > > > Can you find any evidence that it's acceptable to interleave multiple > > writers that are doing O_APPEND? At best, to do what you're asking, > > they could be kept from being interleaved from the cont

Re: NFS client/buffer cache deadlock

2005-04-22 Thread Brian Fundakowski Feldman
On Wed, Apr 20, 2005 at 01:29:10PM -0400, Garrett Wollman wrote: > < PROTECTED]> said: > > > I think the first is more useful behavior than the last. Supporting it > > should be exactly the same as supporting what happens if the actual > > filesystem fills up. In this case, the filesystem is bei

Re: NFS client/buffer cache deadlock

2005-04-22 Thread Garrett Wollman
< POSIX == SUSv3 these days. Not quite. POSIX and SUSv3 use the same specification, but don't require the same things. (Specifically, SUSv3 requires the XSI option to be implemented.) -GAWollman ___ freebsd-hackers@freebsd.org mailing list http://lis

Re: NFS client/buffer cache deadlock

2005-04-21 Thread Garrett Wollman
< said: > I think the first is more useful behavior than the last. Supporting it > should be exactly the same as supporting what happens if the actual > filesystem fills up. In this case, the filesystem is being requested to > write more "than there is room for." Returning a short write for ope

Re: NFS client/buffer cache deadlock

2005-04-21 Thread Garrett Wollman
< said: > Btw.: I'm not sure write(),writev() and pwrite() are allowed to do short > writes on regular files... ? I believe it is the intent of the Standard to prohibit this (a paragraph in the rationale says that short writes can only happen if O_NONBLOCK is set, but this is clearly wrong becaus

Re: NFS client/buffer cache deadlock

2005-04-21 Thread Dag-Erling Smørgrav
Marc Olzheim <[EMAIL PROTECTED]> writes: > Ah, I was reading the SUSv2 page: > > http://www.opengroup.org/onlinepubs/009695399/functions/write.html > > instead of the POSIX version. POSIX == SUSv3 these days. DES -- Dag-Erling Smørgrav - [EMAIL PROTECTED]

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Marc Olzheim
On Wed, Apr 20, 2005 at 01:28:39PM -0400, Brian Fundakowski Feldman wrote: > > It is ok to return partial success if the first chunk of a large write > > succeeded and a later chunk failed persistently, but not if it cannot be > > performed as a single NFS transaction. > > What is your rationale f

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Brian Fundakowski Feldman
On Wed, Apr 20, 2005 at 01:29:10PM -0400, Garrett Wollman wrote: > < PROTECTED]> said: > > > I think the first is more useful behavior than the last. Supporting it > > should be exactly the same as supporting what happens if the actual > > filesystem fills up. In this case, the filesystem is bei

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Brian Fundakowski Feldman
On Wed, Apr 20, 2005 at 07:12:20PM +0200, Jilles Tjoelker wrote: > On Wed, Apr 20, 2005 at 11:52:33AM -0400, Brian Fundakowski Feldman wrote: > > On Wed, Apr 20, 2005 at 05:35:28PM +0200, Marc Olzheim wrote: > > > On Wed, Apr 20, 2005 at 11:20:38AM -0400, Brian Fundakowski Feldman wrote: > > > > >

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Jilles Tjoelker
On Wed, Apr 20, 2005 at 11:52:33AM -0400, Brian Fundakowski Feldman wrote: > On Wed, Apr 20, 2005 at 05:35:28PM +0200, Marc Olzheim wrote: > > On Wed, Apr 20, 2005 at 11:20:38AM -0400, Brian Fundakowski Feldman wrote: > > > > Btw.: I'm not sure write(),writev() and pwrite() are allowed to do short

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Brian Fundakowski Feldman
On Wed, Apr 20, 2005 at 05:35:28PM +0200, Marc Olzheim wrote: > On Wed, Apr 20, 2005 at 11:20:38AM -0400, Brian Fundakowski Feldman wrote: > > Reads should be totally unaffected... > > The server was misbehaving. Fixed. :-) > > > > Btw.: I'm not sure write(),writev() and pwrite() are allowed to d

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Marc Olzheim
On Wed, Apr 20, 2005 at 11:20:38AM -0400, Brian Fundakowski Feldman wrote: > Reads should be totally unaffected... The server was misbehaving. Fixed. :-) > > Btw.: I'm not sure write(),writev() and pwrite() are allowed to do short > > writes on regular files... ? > > Our manpage is incorrect; PO

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Brian Fundakowski Feldman
On Wed, Apr 20, 2005 at 04:38:42PM +0200, Marc Olzheim wrote: > On Wed, Apr 20, 2005 at 10:24:48AM -0400, Brian Fundakowski Feldman wrote: > > > It does and it seems to work. The NFS performance drops considerably > > > though, from 8/9 MByte/s to 3/4 on sequential reads for instance. > > > > > >

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Marc Olzheim
On Wed, Apr 20, 2005 at 10:24:48AM -0400, Brian Fundakowski Feldman wrote: > > It does and it seems to work. The NFS performance drops considerably > > though, from 8/9 MByte/s to 3/4 on sequential reads for instance. > > > > kern/79208 is fixed by this indeed, in that I get short writes (in case

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Brian Fundakowski Feldman
On Wed, Apr 20, 2005 at 04:04:09PM +0200, Marc Olzheim wrote: > On Tue, Apr 19, 2005 at 04:47:23PM -0400, Brian Fundakowski Feldman wrote: > > This compiles. > > It does and it seems to work. The NFS performance drops considerably > though, from 8/9 MByte/s to 3/4 on sequential reads for instance

Re: NFS client/buffer cache deadlock

2005-04-20 Thread Marc Olzheim
On Tue, Apr 19, 2005 at 04:47:23PM -0400, Brian Fundakowski Feldman wrote: > This compiles. It does and it seems to work. The NFS performance drops considerably though, from 8/9 MByte/s to 3/4 on sequential reads for instance. kern/79208 is fixed by this indeed, in that I get short writes (in ca

Re: NFS client/buffer cache deadlock

2005-04-19 Thread Brian Fundakowski Feldman
On Tue, Apr 19, 2005 at 12:16:16PM -0400, Brian Fundakowski Feldman wrote: > On Tue, Apr 19, 2005 at 06:09:00PM +0200, Marc Olzheim wrote: > > On Tue, Apr 19, 2005 at 06:02:58PM +0200, Marc Olzheim wrote: > > > On Tue, Apr 19, 2005 at 11:18:00AM -0400, Brian Fundakowski Feldman wrote: > > > > Does

Re: NFS client/buffer cache deadlock

2005-04-19 Thread Brian Fundakowski Feldman
On Tue, Apr 19, 2005 at 06:09:00PM +0200, Marc Olzheim wrote: > On Tue, Apr 19, 2005 at 06:02:58PM +0200, Marc Olzheim wrote: > > On Tue, Apr 19, 2005 at 11:18:00AM -0400, Brian Fundakowski Feldman wrote: > > > Does this work for you? > > > > ... > > > > cc -c -O -pipe -Wall -Wredundant-decls -W

Re: NFS client/buffer cache deadlock

2005-04-19 Thread Marc Olzheim
On Tue, Apr 19, 2005 at 06:02:58PM +0200, Marc Olzheim wrote: > On Tue, Apr 19, 2005 at 11:18:00AM -0400, Brian Fundakowski Feldman wrote: > > Does this work for you? > > ... > > cc -c -O -pipe -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes > -Wmissing-prototypes -Wpointer-arith

Re: NFS client/buffer cache deadlock

2005-04-19 Thread Marc Olzheim
On Tue, Apr 19, 2005 at 11:18:00AM -0400, Brian Fundakowski Feldman wrote: > Does this work for you? ... cc -c -O -pipe -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual -fformat-extensions -std=c99 -g -nostdinc -I- -I. -I

Re: NFS client/buffer cache deadlock

2005-04-19 Thread Brian Fundakowski Feldman
On Tue, Apr 19, 2005 at 03:32:27PM +0200, Marc Olzheim wrote: > On Mon, Apr 18, 2005 at 10:33:21PM +0200, Marc Olzheim wrote: > > On Mon, Apr 18, 2005 at 04:22:13PM -0400, Brian Fundakowski Feldman wrote: > > > > > http://green.homeunix.org/~green/nfs_client.deadlock.patch > > > > Hmm, could you ch

Re: NFS client/buffer cache deadlock

2005-04-19 Thread Marc Olzheim
On Mon, Apr 18, 2005 at 10:33:21PM +0200, Marc Olzheim wrote: > On Mon, Apr 18, 2005 at 04:22:13PM -0400, Brian Fundakowski Feldman wrote: > > > > http://green.homeunix.org/~green/nfs_client.deadlock.patch > > > Hmm, could you change it into a diff -u ? > > > > I replaced the patch with one with -

Re: NFS client/buffer cache deadlock

2005-04-18 Thread Marc Olzheim
On Fri, Apr 15, 2005 at 11:07:08AM -0400, Brian Fundakowski Feldman wrote: > > Is this supposed to fix kern/79208 ? > > Yes, it does; would you like to try a more recent version of the patch? > It's actually against -STABLE, but it needs to be tested in -CURRENT if > it's going ot try to make it i

Re: NFS client/buffer cache deadlock

2005-04-15 Thread Brian Fundakowski Feldman
On Fri, Apr 15, 2005 at 03:21:08PM +0200, Marc Olzheim wrote: > On Fri, Apr 15, 2005 at 01:08:21AM -0400, Brian Fundakowski Feldman wrote: > > I'll spare a lengthy write-up because I think the patch documents it well > > enough. It certainly appears to fix things here when doing very large > > blo

Re: NFS client/buffer cache deadlock

2005-04-15 Thread Marc Olzheim
On Fri, Apr 15, 2005 at 01:08:21AM -0400, Brian Fundakowski Feldman wrote: > I'll spare a lengthy write-up because I think the patch documents it well > enough. It certainly appears to fix things here when doing very large > block-sized writes, but it also reduces the throughput with those block >

NFS client/buffer cache deadlock

2005-04-14 Thread Brian Fundakowski Feldman
I'll spare a lengthy write-up because I think the patch documents it well enough. It certainly appears to fix things here when doing very large block-sized writes, but it also reduces the throughput with those block sizes. (I don't think there should be any difference when using reasonable block