From 8d280b98cfe3c0b69c37d355218975c1c0279bb0 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Tue, 14 Mar 2006 13:13:09 +1100 Subject: [XFS] On machines with more than 8 cpus, when running parallel I/O threads, the incore superblock lock becomes the limiting factor for buffered write throughput. Make the contended fields in the incore superblock use per-cpu counters so that there is no global lock to limit scalability. SGI-PV: 946630 SGI-Modid: xfs-linux-melb:xfs-kern:25106a Signed-off-by: David Chinner Signed-off-by: Nathan Scott --- fs/xfs/linux-2.6/xfs_linux.h | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'fs/xfs/linux-2.6/xfs_linux.h') diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h index 67389b74552..377a9f54a04 100644 --- a/fs/xfs/linux-2.6/xfs_linux.h +++ b/fs/xfs/linux-2.6/xfs_linux.h @@ -100,6 +100,11 @@ */ #undef HAVE_REFCACHE /* reference cache not needed for NFS in 2.6 */ #define HAVE_SENDFILE /* sendfile(2) exists in 2.6, but not in 2.4 */ +#if CONFIG_SMP +#define HAVE_PERCPU_SB /* per cpu superblock counters are a 2.6 feature */ +#else +#undef HAVE_PERCPU_SB /* per cpu superblock counters are a 2.6 feature */ +#endif /* * State flag for unwritten extent buffers. -- cgit From d2c32edf64a7e9bc8dfb5cb3a8f7bf7be94c93ae Mon Sep 17 00:00:00 2001 From: Nathan Scott Date: Tue, 14 Mar 2006 13:20:13 +1100 Subject: [XFS] When compiling with gcc 4.0 and CONFIG_SMP unset, there are many warnings along the lines: xfs_linux.h:103:5: warning: "CONFIG_SMP" is not defined. SGI-PV: 946630 SGI-Modid: xfs-linux-melb:xfs-kern:25171a Signed-off-by: Adrian Bunk Signed-off-by: Andrew Morton Signed-off-by: Nathan Scott --- fs/xfs/linux-2.6/xfs_linux.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/xfs/linux-2.6/xfs_linux.h') diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h index 377a9f54a04..e2be64982bc 100644 --- a/fs/xfs/linux-2.6/xfs_linux.h +++ b/fs/xfs/linux-2.6/xfs_linux.h @@ -100,7 +100,7 @@ */ #undef HAVE_REFCACHE /* reference cache not needed for NFS in 2.6 */ #define HAVE_SENDFILE /* sendfile(2) exists in 2.6, but not in 2.4 */ -#if CONFIG_SMP +#ifdef CONFIG_SMP #define HAVE_PERCPU_SB /* per cpu superblock counters are a 2.6 feature */ #else #undef HAVE_PERCPU_SB /* per cpu superblock counters are a 2.6 feature */ -- cgit From e8234a6871aa0de1ed0aeeecb5230ecf3ab414e2 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Tue, 14 Mar 2006 13:23:52 +1100 Subject: [XFS] Add support for hotplug CPUs to the per-CPU superblock counters by registering a notifier callback that listens to CPU up/down events to modify the counters appropriately. SGI-PV: 949726 SGI-Modid: xfs-linux-melb:xfs-kern:25214a Signed-off-by: David Chinner Signed-off-by: Nathan Scott --- fs/xfs/linux-2.6/xfs_linux.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs/xfs/linux-2.6/xfs_linux.h') diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h index e2be64982bc..9fdc14cffb7 100644 --- a/fs/xfs/linux-2.6/xfs_linux.h +++ b/fs/xfs/linux-2.6/xfs_linux.h @@ -73,6 +73,8 @@ #include #include #include +#include +#include #include #include -- cgit From 01e1b69cfcdcfdd5b405165eaba29428f8b18a7c Mon Sep 17 00:00:00 2001 From: David Chinner Date: Tue, 14 Mar 2006 13:29:16 +1100 Subject: [XFS] using a spinlock per cpu for superblock counter exclusion results in a preēmpt counter overflow at 256p and above. Change the exclusion mechanism to use atomic bit operations and busy wait loops to emulate the spin lock exclusion mechanism but without the preempt count issues. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit SGI-PV: 950027 SGI-Modid: xfs-linux-melb:xfs-kern:25338a Signed-off-by: David Chinner Signed-off-by: Nathan Scott --- fs/xfs/linux-2.6/xfs_linux.h | 1 + 1 file changed, 1 insertion(+) (limited to 'fs/xfs/linux-2.6/xfs_linux.h') diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h index 9fdc14cffb7..bd88ccb0cad 100644 --- a/fs/xfs/linux-2.6/xfs_linux.h +++ b/fs/xfs/linux-2.6/xfs_linux.h @@ -75,6 +75,7 @@ #include #include #include +#include #include #include -- cgit From ec86dc02fdc062d0d298814b1e78b482ab38caf7 Mon Sep 17 00:00:00 2001 From: Nathan Scott Date: Fri, 17 Mar 2006 17:25:36 +1100 Subject: [XFS] Complete transition away from linvfs naming convention, finally. SGI-PV: 947038 SGI-Modid: xfs-linux-melb:xfs-kern:25474a Signed-off-by: Nathan Scott --- fs/xfs/linux-2.6/xfs_linux.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/xfs/linux-2.6/xfs_linux.h') diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h index bd88ccb0cad..1fe09f2d651 100644 --- a/fs/xfs/linux-2.6/xfs_linux.h +++ b/fs/xfs/linux-2.6/xfs_linux.h @@ -234,7 +234,7 @@ BUFFER_FNS(PrivateStart, unwritten); #define xfs_sort(a,n,s,fn) sort(a,n,s,fn,NULL) #define xfs_stack_trace() dump_stack() #define xfs_itruncate_data(ip, off) \ - (-vmtruncate(LINVFS_GET_IP(XFS_ITOV(ip)), (off))) + (-vmtruncate(vn_to_inode(XFS_ITOV(ip)), (off))) #define xfs_statvfs_fsid(statp, mp) \ ({ u64 id = huge_encode_dev((mp)->m_ddev_targp->bt_dev); \ __kernel_fsid_t *fsid = &(statp)->f_fsid; \ -- cgit