[Cluster-devel] Re: [PATCH] dlm: fix basts for granted PR waiting CW

2007-07-30 Thread Steven Whitehouse
Hi, Now in the GFS2 -nmw git tree. Thanks, Steve. On Wed, 2007-07-25 at 11:01 -0500, David Teigland wrote: Fix a long standing bug where a blocking callback would be missed when there's a granted lock in PR mode and waiting locks in both PR and CW modes (and the PR lock was added to the

[Cluster-devel] [PATCH] [GFS2] Clean up run_queue()

2007-07-30 Thread Steven Whitehouse
A further clean up to make the logic a bit more obvious and remove a local variable which is no longer required. Signed-off-by: Steven Whitehouse [EMAIL PROTECTED] diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 65d6eba..79d61ff 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -564,24

[Cluster-devel] [PATCH] [GFs2] Remove reclaim logic from rq_promote()

2007-07-30 Thread Steven Whitehouse
There is no need to reclaim glocks in this code path since the glockd does this for us anyway. Also the limit of 5000 is entirely aribtrary and it doesn't make sense to slow down this (fast path) when its not actually responsible for generating reclaimable glocks. Signed-off-by: Steven Whitehouse

[Cluster-devel] [GFS2] Remove HIF_PROMOTE flag

2007-07-31 Thread Steven Whitehouse
Hi, This and the following patch are the next two of my glock clean up patches, Steve. From 175876174e4652899c53c311850070b18a50ce36 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Mon, 30 Jul 2007 19:36:05 +0100 Subject: [PATCH] [GFS2] Remove HIF_PROMOTE flag

[Cluster-devel] [GFS2] Introduce workqueue into glock.c

2007-08-01 Thread Steven Whitehouse
From f2186755db781995d7dbc4b2f1993e503a64bde7 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Wed, 1 Aug 2007 13:42:51 +0100 Subject: [PATCH] [GFS2] Introduce workqueue into glock.c The workqueue is used to run the callbacks from the lock modules so that these no longer

Re: [Cluster-devel] [PATCH] dlm: More othercon fixes

2007-08-02 Thread Steven Whitehouse
Hi, On Thu, 2007-08-02 at 14:58 +0100, Patrick Caulfield wrote: The last patch to clean out 'othercon' structures only fixed half the problem. The attached addresses the other situations too, and fixes bz#238490 Signed-Off-By: Patrick Caulfield [EMAIL PROTECTED] Patrick Now in the

[Cluster-devel] [GFS2] Don't use journal lock type

2007-08-07 Thread Steven Whitehouse
Hi, In GFS2 journals are just normal inodes, so we should use the normal inode lock type rather than the journal lock type, Steve. - diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c index 777ca46..976a395 100644 ---

Re: [Cluster-devel] [GFS2] Don't use journal lock type

2007-08-07 Thread Steven Whitehouse
Hi, On Tue, 2007-08-07 at 09:02 -0500, David Teigland wrote: On Tue, Aug 07, 2007 at 02:52:31PM +0100, Steven Whitehouse wrote: Hi, On Tue, 2007-08-07 at 08:45 -0500, David Teigland wrote: On Tue, Aug 07, 2007 at 10:26:06AM +0100, Steven Whitehouse wrote: error

Re: [Cluster-devel] [PATCH 5 of 5] Bz #248176: GFS2: invalid metadata block - REVISED

2007-08-09 Thread Steven Whitehouse
Hi, This one and the previous patch are now both in the -nmw git tree. Thanks, Steve. On Wed, 2007-08-08 at 17:08 -0500, Bob Peterson wrote: This is for bugzilla bug #248176: GFS2: invalid metadata block Patches 1 thru 3 were accepted upstream, but there were problems with 4 and 5. Those

[Cluster-devel] Re: [PATCH][GFS2] bz250772 Force unstuff of hidden quota inode

2007-08-15 Thread Steven Whitehouse
Hi, On Tue, 2007-08-14 at 15:34 -0500, Abhijith Das wrote: This patch forcibly unstuffs (if stuffed) the hidden quota inode at the first availble opportunity. In any practical scenario the quota inode won't be stuffed, so this is ok to do. Unstuffing the quota inode allows us to ignore the

[Cluster-devel] GFS2/DLM -nmw git tree rebase

2007-08-15 Thread Steven Whitehouse
Hi, Linus has accepted the patches from the -fixes tree, so I've rebased the -nmw git tree to his latest tree, Steve.

[Cluster-devel] [GFS2] Add a missing gfs2_trans_add_bh()

2007-08-16 Thread Steven Whitehouse
From 523fd9e576bf1aa1007948742d86e2080cc7811c Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Thu, 16 Aug 2007 17:08:20 +0100 Subject: [PATCH] [GFS2] Add a missing gfs2_trans_add_bh() This was missing from the dir_split_leaf() function although in most cases its

Re: [Cluster-devel] Re: [gfs2][RFC] readdir caused ls process into D (uninterruptible) state, under testing with Samba 3.0.25

2007-08-17 Thread Steven Whitehouse
Hi, On Fri, 2007-08-17 at 15:43 +0800, rae l wrote: [some comments trimmed for brevity] then I start a simple ls command on the gfs2 mouting point: $ ls /mnt/gfs2 the ls process is also changed to D state, I think it's problems about readdir implementation in gfs2, and I want to

Re: [Cluster-devel] [PATCH] [GFS2] bz253289 Add NULL entry to token table

2007-08-18 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Fri, 2007-08-17 at 20:22 -0500, Benjamin Marzinski wrote: match_token() was returning garbage data instead of a fail value. This data happened to match a valid option id for an option that required an argument (in this case, lockproto=%s) For

Re: [Cluster-devel] [GFS2 PATCH] Reduce truncate IO traffic

2007-08-20 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Mon, 2007-08-20 at 09:29 -0400, Wendy Cheng wrote: Current GFS2 setattr call unconditionally invokes do_shrink even the requested size and actual file size are equal. This has generated large amount of extra IOs found during NFS benchmark

[Cluster-devel] [GFS2] Use write_one_page() for writeback of in-place data

2007-08-20 Thread Steven Whitehouse
From fe2c07e31e0ed90422fab685eef97997d5befcac Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Mon, 20 Aug 2007 15:48:00 +0100 Subject: [PATCH] [GFS2] Use write_one_page() for writeback of in-place data When writing back in-place data, we need to hold the page lock when

[Cluster-devel] Re: [PATCH][GFS2] - fix for bz253238 + Wendy's dump lockname in hex

2007-08-22 Thread Steven Whitehouse
-by: Steven Whitehouse [EMAIL PROTECTED] Signed-off-by: S. Wendy Cheng [EMAIL PROTECTED] Signed-off-by: Abhijith Das [EMAIL PROTECTED] This is now in the -nmw git tree. Thanks, Steve.

Re: [Cluster-devel] [PATCH] GFS2: Patch to protect sd_log_num_jdata

2007-08-22 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Wed, 2007-08-22 at 11:15 -0500, Bob Peterson wrote: Hi, This is a patch to GFS2 to protect sd_log_num_jdata with the gfs2_log_lock. Without this patch, there is a timing window where you can get hit the following assert from function

[Cluster-devel] Re: [GFS2] Inode meta data corruption

2007-08-22 Thread Steven Whitehouse
Hi, On Wed, 2007-08-22 at 13:11 -0400, Wendy Cheng wrote: Patch to fix the create-truncate meta data corruption problem reported via bugzilla 253590 ... Wendy Now in the -nmw git tree. Thanks, Steve.

[Cluster-devel] [GFS2] Fix for lock ordering in unlink

2007-08-23 Thread Steven Whitehouse
Hi, The ordering of locks in unlink was wrong as it was using the _nq_m varient of the glock functions which only for locking multiple locks of the same type. This patch corrects the locking to parent - child - rgrp as required by gfs2's lock ordering rules. Signed-off-by: Steven Whitehouse

Re: [Cluster-devel] [PATCH] [GFS2] bz253768 (bz248480 try 2) delay glock demote for a minimum hold time

2007-08-24 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Thu, 2007-08-23 at 13:19 -0500, Benjamin Marzinski wrote: When a lot of IO, with some distributed mmap IO, is run on a GFS2 filesystem in a cluster, it will deadlock. The reason is that do_no_page() will repeatedly call

Re: [Cluster-devel] [GFS2] fix inode meta data corruption

2007-08-24 Thread Steven Whitehouse
Hi, On Fri, 2007-08-24 at 09:15 -0400, Wendy Cheng wrote: Previous patch needs to be revised: https://www.redhat.com/archives/cluster-devel/2007-August/msg00252.html This is the new one. -- Wendy Now in the -nmw git tree. Thanks, Steve.

[Cluster-devel] [GFS2] Patch to correct lock ordering of unlink

2007-08-26 Thread Steven Whitehouse
Hi, The following corrects the unlink lock ordering to parent - child - rgrp in common with the rest of GFS2, Steve. diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index 5b8b994..2cbe5a3 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -278,17 +278,25 @@ static int

[Cluster-devel] [GFS2] Introduce gfs2_remove_from_ail

2007-08-27 Thread Steven Whitehouse
From 1e5a49fb61b78f5df2065f48a33b04102a8a083f Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Mon, 27 Aug 2007 09:45:26 +0100 Subject: [PATCH] [GFS2] Introduce gfs2_remove_from_ail This collects together the operations required to remove a gfs2_bufdata from the ail lists

[Cluster-devel] [GFS2] Move pin/unpin into lops.c, clean up locking

2007-08-27 Thread Steven Whitehouse
From 4eb6b7e0299a35ecb1ed7e8c186d6d24e9a3ac4e Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Mon, 27 Aug 2007 13:54:05 +0100 Subject: [PATCH] [GFS2] Move pin/unpin into lops.c, clean up locking gfs2_pin and gfs2_unpin are only used in lops.c, despite being defined

[Cluster-devel] [GFS2] Remove ail2 list from the ai

2007-08-27 Thread Steven Whitehouse
From 2a666f519dd12e8b3a82d1e16cad3114cfdd917d Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Mon, 27 Aug 2007 16:42:29 +0100 Subject: [PATCH] [GFS2] Remove ail2 list from the ai The ail2 list wasn't actually used for anything other than gathering buffers at the end

Re: [Cluster-devel] [GFS2] Remove ail2 list from the ai

2007-08-27 Thread Steven Whitehouse
Hi, On Mon, 2007-08-27 at 11:30 -0500, David Teigland wrote: On Mon, Aug 27, 2007 at 05:00:06PM +0100, Steven Whitehouse wrote: From 2a666f519dd12e8b3a82d1e16cad3114cfdd917d Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Mon, 27 Aug 2007 16:42:29 +0100 Subject

[Cluster-devel] [GFS2] Use slab operations for all gfs2_bufdata allocations

2007-09-02 Thread Steven Whitehouse
From 380abca2c5489948bb81247df57a97cb5660f198 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Sun, 2 Sep 2007 16:47:38 +0100 Subject: [PATCH] [GFS2] Use slab operations for all gfs2_bufdata allocations The old revoke structure was allocated using kalloc/kfree

Re: [Cluster-devel] [PATCH GFS2] Move inode deletion out of blocking_cb

2007-09-14 Thread Steven Whitehouse
Hi, On Thu, 2007-09-13 at 17:52 -0400, Wendy Cheng wrote: Move inode deletion code out of blocking_cb handle_callback route to avoid racy conditions that end up blocking lock_dlm1 thread. Fix bugzilla 286821. -- Wendy Now in the -nmw git tree. Thanks, Steve.

[Cluster-devel] Re: [PATCH][GFS2] - bz 272021: GFS2 - flocks from same process trip kernel BUG at fs/gfs2/glock.c:1118!

2007-09-14 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Thu, 2007-09-13 at 23:35 -0500, Abhijith Das wrote: This patch adds a new flag to the gfs2_holder structure GL_FLOCK. It is set on holders of glocks representing flocks. This flag is checked in add_to_queue() and a process is permitted to

Re: [Cluster-devel] [PATCH] [DLM] bz#250464: Make dlm_sendd cond_resched more

2007-09-14 Thread Steven Whitehouse
Hi, On Fri, 2007-09-14 at 08:49 +0100, Patrick Caulfield wrote: Under high recovery loads dlm_sendd can monopolise the CPU and cause soft lockups. This one extra and one moved cond_resched() make it yield a little more during such times keeping work moving. Signed-Off-By: Patrick

Re: [Cluster-devel] [PATCH] [GFS2] bz 276631 : GFS2: chmod hung - TRY 2

2007-09-14 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Thu, 2007-09-13 at 23:04 -0500, Bob Peterson wrote: Josef's right--my bad. Here is the corrected patch for 276631. The problem boiled down to a race between the gdlm_init_threads() function initializing thread1 and its setting of blist = 1.

[Cluster-devel] [GFS2] Don't try to remove buffers that don't exist

2007-09-20 Thread Steven Whitehouse
From da5dd3052aded73757af74005b1aa8bfcd00bf3e Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Thu, 20 Sep 2007 15:26:33 +0100 Subject: [PATCH] [GFS2] Don't try to remove buffers that don't exist Signed-off-by: Steven Whitehouse [EMAIL PROTECTED] diff --git a/fs/gfs2

[Cluster-devel] GFS2 git tree update

2007-09-25 Thread Steven Whitehouse
Hi, Linus has released 2.6.23-rc8, so I've updated the -nmw git tree to that level, Steve.

[Cluster-devel] [GFS2] Get superblock a different way

2007-09-26 Thread Steven Whitehouse
From 24647918647d19bb02dfa3e7d92382921e4cbcd1 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Wed, 26 Sep 2007 09:39:31 +0100 Subject: [PATCH] [GFS2] Get superblock a different way The mapping may be NULL by the time the I/O has completed, so we now get the superblock

[Cluster-devel] Re: [PATCH] dlm: block dlm_recv in recovery transition

2007-10-02 Thread Steven Whitehouse
Hi, Sorry for the delay. Now in the -nmw git tree, Steve. On Thu, 2007-09-27 at 15:53 -0500, David Teigland wrote: Introduce a per-lockspace rwsem that's held in read mode by dlm_recv threads while working in the dlm. This allows dlm_recv activity to be suspended when the lockspace

Re: [Cluster-devel] [PATCH] dlm: don't overwrite castparam if it's NULL

2007-10-02 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Mon, 2007-10-01 at 15:28 +0100, Patrick Caulfield wrote: If the castaddr passed to the userland API is NULL then don't overwrite the existing castparam. This allows a different thread to cancel a lock request and the CANCEL AST gets delivered

[Cluster-devel] GFS2 git tree rebase

2007-10-10 Thread Steven Whitehouse
Hi, Linus has released 2.6.23, so I've rebased against that. I'm not now, expecting to put any more patches into the -nmw tree until after the merge window, so that any patches that I've been sent recently and which haven't made it into the current tree will go into the next -nmw tree unless they

[Cluster-devel] [GFS2/DLM] Pull request

2007-10-12 Thread Steven Whitehouse
[DLM] don't overwrite castparam if it's NULL Steve French (1): [GFS2] GFS2 not checking pointer on create when running under nfsd Steven Whitehouse (16): [GFS2] Fix two races relating to glock callbacks [GFS2] Fix an oops in glock dumping [GFS2] Reduce number

[Cluster-devel] [GFS2] Clean up internal read function

2007-10-16 Thread Steven Whitehouse
d8923f7c2e0d08cce31996f639284b5083338cf4 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Mon, 15 Oct 2007 14:42:35 +0100 Subject: [PATCH] [GFS2] Clean up internal read function As requested by Christoph, this patch cleans up GFS2's internal read function so

[Cluster-devel] [GFS2] Use -page_mkwrite() for mmap()

2007-10-16 Thread Steven Whitehouse
From 548d3cbf986d8092f8e13ff9960eb3ec1e46d79f Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Mon, 15 Oct 2007 15:40:33 +0100 Subject: [PATCH] [GFS2] Use -page_mkwrite() for mmap() This cleans up the mmap() code path for GFS2 by implementing the page_mkwrite function

[Cluster-devel] [GFS2] Remove useless i_cache from inodes

2007-10-16 Thread Steven Whitehouse
From aa974896eb5c1a70d4be6df1e3f9f5e12b8887f9 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Mon, 15 Oct 2007 16:29:05 +0100 Subject: [PATCH] [GFS2] Remove useless i_cache from inodes The i_cache was designed to keep references to the indirect blocks used during block

[Cluster-devel] [GFS2] Remove unused field in struct gfs2_inode

2007-10-16 Thread Steven Whitehouse
From a0cd4c9eb46a79829a60a37743fc0089882d3af6 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Tue, 16 Oct 2007 11:47:04 +0100 Subject: [PATCH] [GFS2] Remove unused field in struct gfs2_inode Removes a field that is not used. Signed-off-by: Steven Whitehouse [EMAIL

Re: [Cluster-devel] [GFS2] Clean up internal read function

2007-10-16 Thread Steven Whitehouse
. On Tue, 2007-10-16 at 12:50 +0100, Steven Whitehouse wrote: Hi, Since it looks like the new AOPs patches which might clash with this are going to be merged tonight (Andrew Morton has sent them to Linus) once I see them in Linus tree, then I'll merge up add the patches which

Re: [Cluster-devel] [GFS2] Clean up internal read function

2007-10-16 Thread Steven Whitehouse
Hi, On Tue, 2007-10-16 at 09:16 -0500, Kevin Anderson wrote: On Tue, 2007-10-16 at 16:11 +0100, Steven Whitehouse wrote: Hi, On Tue, 2007-10-16 at 08:54 -0500, Kevin Anderson wrote: Steve, What is the impact on performance due to this change? Seems to be messing with where

[Cluster-devel] [GFS2] Introduce gfs2_set_aops()

2007-10-17 Thread Steven Whitehouse
From 8b777b51c2aebedc469fe7c6b0f41b50746c8281 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Wed, 17 Oct 2007 08:47:38 +0100 Subject: [PATCH] [GFS2] Introduce gfs2_set_aops() Just like ext3 we now have three sets of address space operations to cover the cases

[Cluster-devel] [GFS2] Add gfs2_is_writeback()

2007-10-17 Thread Steven Whitehouse
5f5c60307feda8319fa3285c000dc07c24e84007 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Wed, 17 Oct 2007 08:35:19 +0100 Subject: [PATCH] [GFS2] Add gfs2_is_writeback() This adds a function gfs2_is_writeback() along the lines of the existing gfs2_is_jdata() in order to clean up the code and make

[Cluster-devel] [GFS2] Split gfs2_writepage into three cases

2007-10-17 Thread Steven Whitehouse
From da6ae988ab5a8f1a111b1dc857c705c2e143c309 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Fri, 28 Sep 2007 13:49:05 +0100 Subject: [PATCH] [GFS2] Split gfs2_writepage into three cases This patch splits gfs2_writepage into separate functions for each of the three cases

[Cluster-devel] GFS2/DLM git tree

2007-10-18 Thread Steven Whitehouse
Hi, Since a number of patches got pushed into Linus' tree last night which affect GFS2 and/or DLM in one way or another, I've rebased again today. I don't know of any more patches which are likely to hit the tree shortly and which affect GFS2/DLM, so I hope the next rebase will be at -rc1 but I

[Cluster-devel] Re: [PATCH] [DLM] Bind connections from known local address when using TCP

2007-10-29 Thread Steven Whitehouse
Hi, On Thu, 2007-10-25 at 18:49 -0400, Lon Hohberger wrote: A common problem occurs when multiple IP addresses within the same subnet are assigned to the same NIC. If we make a connection attempt to another address on the same subnet as one of those addresses, the connection attempt will not

[Cluster-devel] [GFS2] Remove reclaim limit

2007-10-31 Thread Steven Whitehouse
This call to reclaim glocks is not needed, and in particular we don't want it in the fast path for locking glocks. The limit was entirely arbitrary anyway and we can't expect users to adjust things like this, the remaining code will do the right thing on its own. Signed-off-by: Steven Whitehouse

[Cluster-devel] [GFS2] Removed unused variables

2007-10-31 Thread Steven Whitehouse
Remove unused variables. Signed-off-by: Steven Whitehouse [EMAIL PROTECTED] diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 82dfe9b..f7a50fe 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -412,9 +412,6 @@ struct gfs2_args { struct gfs2_tune { spinlock_t gt_spin

[Cluster-devel] Re: [PATCH] [DLM] Bind connections from known local address when using TCP

2007-10-31 Thread Steven Whitehouse
Hi, On Tue, 2007-10-30 at 14:02 +, Patrick Caulfield wrote: Steven Whitehouse wrote: Hi, On Thu, 2007-10-25 at 18:49 -0400, Lon Hohberger wrote: A common problem occurs when multiple IP addresses within the same subnet are assigned to the same NIC. If we make a connection attempt

Re: [Cluster-devel] [GFS2] Git tree update

2007-10-31 Thread Steven Whitehouse
Hi, On Mon, 2007-10-29 at 11:25 +0100, Fabio Massimo Di Nitto wrote: Steven Whitehouse wrote: Hi, Since Linus released -rc1 when I was on holiday last week, I've just updated the git tree. There have been quite a lot of changes in the core kernel which affect GFS2 but which were

[Cluster-devel] [GFS2] Add sync_page to metadata address space operations

2007-11-01 Thread Steven Whitehouse
From ac8280e7b89b4f84019db9da34e6401024a820cf Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Thu, 1 Nov 2007 09:34:14 + Subject: [PATCH] [GFS2] Add sync_page to metadata address space operations This set of address space operations was missing a sync_page operation

[Cluster-devel] [GFS2] Remove flags no longer required

2007-11-02 Thread Steven Whitehouse
From abadedd83f6f70d5acb32d0fcfad564557d50797 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Fri, 2 Nov 2007 09:14:31 + Subject: [PATCH] [GFS2] Remove flags no longer required The HIF_MUTEX and HIF_PROMOTE flags were set on the glock holders depending upon which

[Cluster-devel] Re: [PATCH][GFS2] Given device ID rather than s_id in id sysfs file

2007-11-02 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve On Fri, 2007-11-02 at 09:37 -0500, Bob Peterson wrote: Hi, This patch changes the /sys/fs/gfs2/s_id/id file to give the device id major:minor rather than the s_id. That enables gfs2_tool to match devices properly (by id, not name) when locating

[Cluster-devel] [GFS2] Reorder writeback for glock sync

2007-11-02 Thread Steven Whitehouse
From 601765436dd953ef0749bf90d1f467aa35f9dd77 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Fri, 2 Nov 2007 08:39:34 + Subject: [PATCH] [GFS2] Reorder writeback for glock sync Previously we were doing (write data, wait for data, write metadata, wait for metadata

Re: [Cluster-devel] Re: [2.6 patch] fs/dlm/: proper prototypes

2007-11-07 Thread Steven Whitehouse
Hi, I've added this to the GFS2 -nmw git tree and since Linus has released -rc2, I've rebased it to that at the same time. Thanks, Steve. On Mon, 2007-11-05 at 08:42 -0600, David Teigland wrote: On Sat, Nov 03, 2007 at 01:04:30AM +0100, Adrian Bunk wrote: This patch adds a proper prototype

[Cluster-devel] [GFS2] Various patches

2007-11-13 Thread Steven Whitehouse
Hi, Following this message will be a variety of patches which constitute my current patch queue. They are aimed at solving a number of issues which have been reported recently relating to performance and memory usage. A number of the patches are standalone, if you ignore the last patch in the

[Cluster-devel] [GFS2] Use atomic_t for journal free blocks counter

2007-11-13 Thread Steven Whitehouse
From 4b9510f05a676623069d22f49cd9da64ea6f708b Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Thu, 8 Nov 2007 14:55:03 + Subject: [PATCH] [GFS2] Use atomic_t for journal free blocks counter This patch changes the counter which keeps track of the free blocks

[Cluster-devel] [GFS2] Move gfs2_logd into log.c

2007-11-13 Thread Steven Whitehouse
From fbf7c723b65e7eccf462b80bb378bbe5ea8037b0 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Fri, 9 Nov 2007 10:01:41 + Subject: [PATCH] [GFS2] Move gfs2_logd into log.c This means that we can mark gfs2_ail1_empty static and prepares the way for further changes

[Cluster-devel] [GFS2] Don't periodically update the jindex

2007-11-13 Thread Steven Whitehouse
From b5dc3ccdb88e58656aadac428452d2c3050802bc Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Fri, 9 Nov 2007 10:07:21 + Subject: [PATCH] [GFS2] Don't periodically update the jindex We only care about the content of the jindex in two cases, one is when we mount the fs

[Cluster-devel] [GFS2] Update journal wake up/flushing

2007-11-13 Thread Steven Whitehouse
This patch has several important features: o Log flushing removed from gfs2_log_reserve so that this just waits in the case that we try to start a transaction when there are not enough free blocks in the journal. We use an exclusive wait queue to avoid the thundering herd and to ensure

[Cluster-devel] [GFS2] Logd patches

2007-11-15 Thread Steven Whitehouse
in the initial version of this which I posted, so this is now the remaining proposed patch for merging. Steve. From a3e7bb2908abb176b30c4ce06658f2d01f9a5c47 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Thu, 15 Nov 2007 10:04:42 + Subject: [PATCH] [GFS2] Revise gfs2_logd

[Cluster-devel] Re: [PATCH] gfs2: tidy up error message

2007-11-15 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Thu, 2007-11-15 at 09:01 -0600, David Teigland wrote: Print error with log_error() to be consistent with others. Signed-off-by: David Teigland [EMAIL PROTECTED] --- fs/gfs2/locking/dlm/mount.c |3 +-- 1 files changed, 1 insertions(+),

[Cluster-devel] Re: [GFS2] Fix runtime issue with UP kernels

2007-11-16 Thread Steven Whitehouse
Hi, On Fri, 2007-11-16 at 09:57 +0100, Fabio Massimo Di Nitto wrote: Hi Steven, the patch fixes the issues we discussed on the mailing list before and IRC. The issue is indeed UP vs SMP and it is totally random. spin_is_locked() is a bad assertion because there is no correct answer on

[Cluster-devel] Re: [PATCH] gfs2: remove unnecessary permission checks

2007-11-27 Thread Steven Whitehouse
Hi, Now applied to the -nmw git tree. Thanks, Steve. On Tue, 2007-11-27 at 06:16 +0100, Fabio Massimo Di Nitto wrote: Hi Steven, Steven Whitehouse wrote: Hi, Now in the -nmw git tree. Thanks, Steve. On Wed, 2007-11-21 at 11:54 -0600, Ryan O'Hara wrote: this patch

[Cluster-devel] Re: [PATCH][GFS2] GFS2 - patch to check for recursive lock requests in gfs2_rename code path

2007-11-30 Thread Steven Whitehouse
Hi, On Thu, 2007-11-29 at 14:13 -0600, Abhijith Das wrote: A certain scenario in the rename code path triggers a kernel BUG() because it accidentally does recursive locking The first lock is requested to unlink an already existing inode (replacing a file) and the second lock is requested when

[Cluster-devel] Re: [PATCH] gfs2: use pid for plock owner for nfs clients

2007-12-07 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Thu, 2007-12-06 at 09:35 -0600, David Teigland wrote: The fl_owner is that of lockd when posix locks arrive from nfs clients, so it can't be used to distinguish between lock holders. Use fl_pid as owner instead; it's the pid of the process on

Re: [Cluster-devel] [PATCH] dlm_controld: add multi-home

2007-12-10 Thread Steven Whitehouse
Hi, On Mon, 2007-12-10 at 15:22 +, Patrick Caulfield wrote: This patch adds multi-home capability to dlm_controld. If a node has more than one address then the DLM will be told about all of them, rather than just the first as at present. It is assumed that sctp will be used as a

Re: [Cluster-devel] [GFS2][Patch] Remove function gfs2_get_block

2007-12-11 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Also I rebased the tree to 2.6.24-rc5 since its a while since I last did that, Steve. On Mon, 2007-12-10 at 14:13 -0600, Bob Peterson wrote: Hi, This patch is just a cleanup. Function gfs2_get_block() just calls function gfs2_block_map reversing the last two

[Cluster-devel] [GFS2] Fix typo in log.c

2007-12-11 Thread Steven Whitehouse
From 51aa9f8ebf99017cdcf6fbaa32d8501d3ddf4e46 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Tue, 11 Dec 2007 14:40:23 + Subject: [PATCH] [GFS2] Fix typo in log.c An inequality was the wrong way around causing gfs2_logd to wake up too often. This fixes it. Signed

Re: [Cluster-devel] [GFS2][Patch 0/10] GFS2 performance tweaks

2007-12-12 Thread Steven Whitehouse
Hi, On Tue, 2007-12-11 at 18:43 -0600, Bob Peterson wrote: Hi GFS2 Folks, The following is a set of ten patches designed to give gfs2 somewhat better performance than before. Some are trivial and may or may not have a real impact on performance. I'll let Steve Whitehouse decide which

Re: [Cluster-devel] [GFS2] [Patch 3/10] Run through full bitmaps quicker in gfs2_bitfit

2007-12-12 Thread Steven Whitehouse
Hi, fs/gfs2/rgrp.c: In function ‘gfs2_bitfit’: fs/gfs2/rgrp.c:139: warning: assignment from incompatible pointer type Please can you send me a fix for the above. Thanks, Steve. On Tue, 2007-12-11 at 19:00 -0600, Bob Peterson wrote: Hi, I eliminated the passing of an unused parameter into

Re: [Cluster-devel] [GFS2] [Patch 4/10] Get rid of sd_statfs_mutex

2007-12-12 Thread Steven Whitehouse
Hi, On Wed, 2007-12-12 at 09:31 -0600, Bob Peterson wrote: On Wed, 2007-12-12 at 09:18 +, Steven Whitehouse wrote: You can't do gfs2_trans_add_bh under a spinlock, but there is no reason why you can't just reverse the order of these two statements to fix it, Steve. Hi Steve

[Cluster-devel] GFS2] Mark inode dirty data by association

2007-12-12 Thread Steven Whitehouse
From 873ffcd14f43c51b9030245e3d257679c323e164 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Tue, 11 Dec 2007 09:20:53 + Subject: [PATCH] [GFS2] Mark inode dirty data by association This adds dirty inode data to the b_assoc_buffers list so that the VFS's own metadata

[Cluster-devel] [GFS2] Add bh_to_bufdata() function

2007-12-12 Thread Steven Whitehouse
From efc837e7d13dd81ea7f9b75919c00e0f98b6ddba Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Fri, 30 Nov 2007 14:14:51 + Subject: [PATCH] [GFS2] Add bh_to_bufdata() function This adds an inline function through which all accesses of bh-b_private go for journaled data

[Cluster-devel] Re: Why the gfs2 performance regressed?

2008-01-02 Thread Steven Whitehouse
Hi, On Wed, 2008-01-02 at 15:53 +0800, Cheng Renquan wrote: hello, Steven: I have tested gfs2 in the out-of-box RHEL51 and in the latest gfs2-nmw git repository, finding that the latest gfs2 performance regressed. the testing environment is RHEL5.1, default kernel + gfs2 + samba, I use

[Cluster-devel] GFS2 git update/rebase

2008-01-02 Thread Steven Whitehouse
Hi, I've just updated the GFS2 -nmw git tree. I've rolled out the five patches which were aimed at reducing the memory consumption of GFS2 since they are causing too large a performance hit at the moment. They should reappear once we can solve that problem, but since its likely that a merge

[Cluster-devel] Re: Why the gfs2 performance regressed?

2008-01-02 Thread Steven Whitehouse
Hi, On Thu, 2008-01-03 at 00:44 +0800, Cheng Renquan wrote: On Jan 2, 2008 6:19 PM, Steven Whitehouse [EMAIL PROTECTED] wrote: Are you running single node? if so then use lock_nolock rather than lock_dlm as it will be much faster for fcntl locks. Even if you intend to run as a cluster

[Cluster-devel] [GFS2] Remove unused variable

2008-01-03 Thread Steven Whitehouse
From c6c1c7c8f68d98958a34a03b011e62f6f327656d Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Wed, 2 Jan 2008 10:16:56 + Subject: [PATCH] [GFS2] Remove unused variable The go_drop_th function is never called or referenced. Signed-off-by: Steven Whitehouse [EMAIL

[Cluster-devel] [GFS2] Allow page migration for writeback and ordered pages

2008-01-03 Thread Steven Whitehouse
From c2d6be02fdb6a968e3d5eef1a3c13dbf731fd0e9 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Thu, 3 Jan 2008 11:31:38 + Subject: [PATCH] [GFS2] Allow page migration for writeback and ordered pages To improve performance on NUMA, we use the VM's standard page

Re: [Cluster-devel] [GFS2 patch] Initialize extent_list earlier

2008-01-03 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Thu, 2008-01-03 at 09:24 -0600, Bob Peterson wrote: Hi, Here is a patch for the latest upstream GFS2 code: The journal extent map needs to be initialized sooner than it currently is. Otherwise failed mount attempts (e.g. not enough

[Cluster-devel] [GFS2] Proposed patch to reduce inode size

2008-01-07 Thread Steven Whitehouse
the immediate problem of using too much low-memory on x86 and doesn't add too much overhead. Signed-off-by: Steven Whitehouse [EMAIL PROTECTED] diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 2241141..73dfad7 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -683,7 +683,7 @@ static int do_strip

[Cluster-devel] Re: Why the gfs2 performance regressed?

2008-01-07 Thread Steven Whitehouse
Hi, On Mon, 2008-01-07 at 17:48 +0800, rae l wrote: On Jan 3, 2008 5:36 PM, Denis Cheng [EMAIL PROTECTED] wrote: After some bisecting work, now 2.6.22.15 is verified to be able to support samba high performance, while 2.6.23 cannot, git bisecting work not finished, now v2.6.22 good and

[Cluster-devel] [GFS2] Fix problems relating to execution of files on GFS2

2008-01-07 Thread Steven Whitehouse
, but only on the first lookup, subsequent requests worked correctly. The second was a problem relating to incorrect updating of file sizes which was introduced with the write_begin/end code for GFS2 a little while back. Signed-off-by: Steven Whitehouse [EMAIL PROTECTED] diff --git a/fs/gfs2/ops_address.c

Re: [Cluster-devel] [GFS2] Fix problems relating to execution of files on GFS2

2008-01-08 Thread Steven Whitehouse
Hi, On Mon, 2008-01-07 at 13:11 -0500, Wendy Cheng wrote: Steven Whitehouse wrote: --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -113,8 +113,18 @@ static struct dentry *gfs2_lookup(struct inode *dir, struct dentry *dentry, if (inode IS_ERR(inode)) return

[Cluster-devel] [GFS2] Fix assert in log code

2008-01-10 Thread Steven Whitehouse
From 3a46e3bef51f7a1feffafcac855c1b252fcc5877 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse [EMAIL PROTECTED] Date: Thu, 10 Jan 2008 14:49:43 + Subject: [PATCH] [GFS2] Fix assert in log code Although the values were all being calculated correctly, there was a race in the assert due

Re: [Cluster-devel] [GFS2] [Patch] gfs2_alloc_required performance

2008-01-11 Thread Steven Whitehouse
Hi, Now in the -nmw git tree, Steve. On Fri, 2008-01-11 at 13:44 -0600, Bob Peterson wrote: Hi, This is a small I/O performance enhancement to gfs2. (Actually, it is a rework of an earlier version I got wrong). The idea here is to check if the write extends past the last block in

Re: [Cluster-devel] [GFS2] [Patch] Remove unneeded i_spin

2008-01-11 Thread Steven Whitehouse
Hi, Now in the -nmw git tree. Thanks, Steve. On Fri, 2008-01-11 at 13:31 -0600, Bob Peterson wrote: Hi, This patch removes a vestigial variable i_spin from the gfs2_inode structure. This not only saves us memory (30 of these in memory for the oom test) it also saves us time because

Re: [Cluster-devel] [PATCH] dlm: fix cross-platform error values

2008-01-15 Thread Steven Whitehouse
Hi, On Tue, 2008-01-15 at 07:15 +0100, Fabio M. Di Nitto wrote: Hi guys, On Mon, 14 Jan 2008, Patrick Caulfeld wrote: Some errno values differ across platforms. So if we return things like -EINPROGRESS from one node it can get misinterpreted or rejected on another one. This patch

[Cluster-devel] Re: [PATCH] DLM: Fix static buffer alignment

2008-01-15 Thread Steven Whitehouse
Hi, There is something not quite right about this patch... it doesn't seem to apply and I suspect whitespace, but I can't see what the problem is at the moment, Steve. On Tue, 2008-01-15 at 05:51 +0100, Fabio M. Di Nitto wrote: Hi Steven, gcc does not guarantee that a static buffer is 64bit

Re: [Cluster-devel] [GFS2][Patch] Fix typo

2008-01-16 Thread Steven Whitehouse
Hi, This is now in the -nmw git tree. Thanks, Steve. On Wed, 2008-01-16 at 08:45 -0600, Bob Peterson wrote: Hi, This patch fixes a minor typo. Surprisingly, it still compiled. Regards, Bob Peterson Signed-off-by: Bob Peterson [EMAIL PROTECTED] -- fs/gfs2/meta_io.c |2 +-

[Cluster-devel] [GFS2] Fix page_mkwrite truncation race path

2008-01-17 Thread Steven Whitehouse
. If it was due to a real truncation (as opposed to an invalidate because we let a glock go) then the -fault path will pick that up when it gets called again. Signed-off-by: Steven Whitehouse [EMAIL PROTECTED] diff --git a/fs/gfs2/ops_file.c b/fs/gfs2/ops_file.c index d7f4726..f4842f2 100644 --- a/fs

[Cluster-devel] [GFS2] Speed up gfs2_write_alloc_required, deprecate gfs2_extent_map

2008-01-18 Thread Steven Whitehouse
as all the callers have been converted. Signed-off-by: Steven Whitehouse [EMAIL PROTECTED] diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index e4effc4..67d896d 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -548,6 +548,9 @@ out_fail: return error; } +/* + * Deprecated: do not use

Re: [Cluster-devel] [GFS2][PATCH] - Allow journal recovery on read-only mount

2008-01-19 Thread Steven Whitehouse
Hi, On Fri, 2008-01-18 at 14:35 -0600, Abhijith Das wrote: I'm NACKing this on Dave's suggestion that we restrict this behavior to standalone gfs2 filesystems (lock_nolock). I'm working on a patch for that and will post it soon. --Abhi Why? I don't see the need for such a restriction. The

Re: [Cluster-devel] [PATCH][GFS2] Lockup on error

2008-01-21 Thread Steven Whitehouse
Hi, On Sat, 2008-01-19 at 21:50 -0600, Bob Peterson wrote: Hi, I spotted this bug while I was digging around. Looks like it could cause a lockup in some rare error condition. Regards, Bob Peterson -- Signed-off-by: Bob Peterson [EMAIL PROTECTED] -- fs/gfs2/inode.c |2 +- 1

[Cluster-devel] GFS2 git tree

2008-01-21 Thread Steven Whitehouse
Hi, I've just rebased the tree due to the impending merge window. Also I've removed all of the DLM patches since there will shortly be a separate DLM tree. I'm going to hold off accepting any larger patches now until after the merge window, Steve.

[Cluster-devel] [GFS2] Streamline indirect pointer tree height calculation

2008-01-28 Thread Steven Whitehouse
(10). Signed-off-by: Steven Whitehouse [EMAIL PROTECTED] diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 67d896d..2a90084 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -166,7 +166,7 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) di-di_blocks

  1   2   3   4   5   6   7   8   9   10   >