From: Andrew Morton <akpm@digeo.com>
To: Daniel Phillips <phillips@arcor.de>
Cc: Linus Torvalds <torvalds@transmeta.com>,
"Martin J. Bligh" <mbligh@aracnet.com>,
Oliver Neukum <oliver@neukum.name>,
Rob Landley <landley@trommello.org>,
linux-kernel@vger.kernel.org
Subject: Re: The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 not 3.0 - (NUMA))
Date: Mon, 07 Oct 2002 15:12:19 -0700 [thread overview]
Message-ID: <3DA206C3.9AD2941A@digeo.com> (raw)
In-Reply-To: E17yfxq-0003vd-00@starship
Daniel Phillips wrote:
>
> On Monday 07 October 2002 23:55, Linus Torvalds wrote:
> > On Mon, 7 Oct 2002, Daniel Phillips wrote:
> > > >
> > > > Sure. The mey is:
> > > ^^^ <---- "bet" ?
> >
> > Yeah. What the heck happened to my fingers?
>
> Apparently, one of them missed the key it was aiming for and the other one
> changed hands.
>
They don't call him Kubys for nothing.
I dug out and dusted off Al's Orlov allocator patch. And found
a comment which rather helps explain how it works.
I performance tested this back in November. See
http://www.uwsg.iu.edu/hypermail/linux/kernel/0111.1/0281.html
Bottom line: it's as good as the use-first-fit-everywhere
approach, and appears to have better long-term antifragmentation
characteristics.
I shall test it.
fs/ext2/ext2.h | 1
fs/ext2/ialloc.c | 164 ++++++++++++++++++++++++++++++++++++++++++++-
fs/ext2/super.c | 8 ++
include/linux/ext2_fs_sb.h | 2
4 files changed, 172 insertions(+), 3 deletions(-)
--- 2.5.41/fs/ext2/ialloc.c~orlov-allocator Mon Oct 7 14:31:50 2002
+++ 2.5.41-akpm/fs/ext2/ialloc.c Mon Oct 7 15:04:09 2002
@@ -18,6 +18,7 @@
#include <linux/sched.h>
#include <linux/backing-dev.h>
#include <linux/buffer_head.h>
+#include <linux/random.h>
/*
* ialloc.c contains the inodes allocation and deallocation routines
@@ -205,6 +206,7 @@ static void ext2_preread_inode(struct in
* For other inodes, search forward from the parent directory\'s block
* group to find a free inode.
*/
+#if 0
static int find_group_dir(struct super_block *sb, int parent_group)
{
@@ -238,9 +240,141 @@ static int find_group_dir(struct super_b
mark_buffer_dirty(best_bh);
return best_group;
}
+#endif
+
+/*
+ * Orlov's allocator for directories.
+ *
+ * We always try to spread first-level directories.
+ *
+ * If there are blockgroups with both free inodes and free blocks counts
+ * not worse than average we return one with smallest directory count.
+ * Otherwise we simply return a random group.
+ *
+ * For the rest rules look so:
+ *
+ * It's OK to put directory into a group unless
+ * it has too many directories already (max_dirs) or
+ * it has too few free inodes left (min_inodes) or
+ * it has too few free blocks left (min_blocks) or
+ * it's already running too large debt (max_debt).
+ * Parent's group is prefered, if it doesn't satisfy these
+ * conditions we search cyclically through the rest. If none
+ * of the groups look good we just look for a group with more
+ * free inodes than average (starting at parent's group).
+ *
+ * Debt is incremented each time we allocate a directory and decremented
+ * when we allocate an inode, within 0--255.
+ */
+
+#define INODE_COST 64
+#define BLOCK_COST 256
+
+static int find_group_orlov(struct super_block *sb, struct inode *parent)
+{
+ int parent_group = EXT2_I(parent)->i_block_group;
+ struct ext2_sb_info *sbi = EXT2_SB(sb);
+ struct ext2_super_block *es = sbi->s_es;
+ int ngroups = sbi->s_groups_count;
+ int inodes_per_group = EXT2_INODES_PER_GROUP(sb);
+ int avefreei = le32_to_cpu(es->s_free_inodes_count) / ngroups;
+ int avefreeb = le32_to_cpu(es->s_free_blocks_count) / ngroups;
+ int blocks_per_dir;
+ int ndirs = sbi->s_dir_count;
+ int max_debt, max_dirs, min_blocks, min_inodes;
+ int group = -1, i;
+ struct ext2_group_desc *desc;
+ struct buffer_head *bh;
+
+ if (parent == sb->s_root->d_inode) {
+ struct ext2_group_desc *best_desc = NULL;
+ struct buffer_head *best_bh = NULL;
+ int best_ndir = inodes_per_group;
+ int best_group = -1;
+
+ get_random_bytes(&group, sizeof(group));
+ parent_group = (unsigned)group % ngroups;
+ for (i = 0; i < ngroups; i++) {
+ group = (parent_group + i) % ngroups;
+ desc = ext2_get_group_desc (sb, group, &bh);
+ if (!desc || !desc->bg_free_inodes_count)
+ continue;
+ if (le16_to_cpu(desc->bg_used_dirs_count) >= best_ndir)
+ continue;
+ if (le16_to_cpu(desc->bg_free_inodes_count) < avefreei)
+ continue;
+ if (le16_to_cpu(desc->bg_free_blocks_count) < avefreeb)
+ continue;
+ best_group = group;
+ best_ndir = le16_to_cpu(desc->bg_used_dirs_count);
+ best_desc = desc;
+ best_bh = bh;
+ }
+ if (best_group >= 0) {
+ desc = best_desc;
+ bh = best_bh;
+ group = best_group;
+ goto found;
+ }
+ goto fallback;
+ }
+
+ blocks_per_dir = (le32_to_cpu(es->s_blocks_count) -
+ le32_to_cpu(es->s_free_blocks_count)) / ndirs;
+
+ max_dirs = ndirs / ngroups + inodes_per_group / 16;
+ min_inodes = avefreei - inodes_per_group / 4;
+ min_blocks = avefreeb - EXT2_BLOCKS_PER_GROUP(sb) / 4;
+
+ max_debt = EXT2_BLOCKS_PER_GROUP(sb) / max(blocks_per_dir, BLOCK_COST);
+ if (max_debt * INODE_COST > inodes_per_group)
+ max_debt = inodes_per_group / INODE_COST;
+ if (max_debt > 255)
+ max_debt = 255;
+ if (max_debt == 0)
+ max_debt = 1;
+
+ for (i = 0; i < ngroups; i++) {
+ group = (parent_group + i) % ngroups;
+ desc = ext2_get_group_desc (sb, group, &bh);
+ if (!desc || !desc->bg_free_inodes_count)
+ continue;
+ if (sbi->debts[group] >= max_debt)
+ continue;
+ if (le16_to_cpu(desc->bg_used_dirs_count) >= max_dirs)
+ continue;
+ if (le16_to_cpu(desc->bg_free_inodes_count) < min_inodes)
+ continue;
+ if (le16_to_cpu(desc->bg_free_blocks_count) < min_blocks)
+ continue;
+ goto found;
+ }
+
+fallback:
+ for (i = 0; i < ngroups; i++) {
+ group = (parent_group + i) % ngroups;
+ desc = ext2_get_group_desc (sb, group, &bh);
+ if (!desc || !desc->bg_free_inodes_count)
+ continue;
+ if (le16_to_cpu(desc->bg_free_inodes_count) >= avefreei)
+ goto found;
+ }
+
+ return -1;
+
+found:
+ desc->bg_free_inodes_count =
+ cpu_to_le16(le16_to_cpu(desc->bg_free_inodes_count) - 1);
+ desc->bg_used_dirs_count =
+ cpu_to_le16(le16_to_cpu(desc->bg_used_dirs_count) + 1);
+ sbi->s_dir_count++;
+ mark_buffer_dirty(bh);
+ return group;
+}
-static int find_group_other(struct super_block *sb, int parent_group)
+static int find_group_other(struct super_block *sb, struct inode *parent)
{
+ int parent_group = EXT2_I(parent)->i_block_group;
int ngroups = EXT2_SB(sb)->s_groups_count;
struct ext2_group_desc *desc;
struct buffer_head *bh;
@@ -312,9 +446,9 @@ struct inode * ext2_new_inode(struct ino
es = EXT2_SB(sb)->s_es;
repeat:
if (S_ISDIR(mode))
- group = find_group_dir(sb, EXT2_I(dir)->i_block_group);
+ group = find_group_orlov(sb, dir);
else
- group = find_group_other(sb, EXT2_I(dir)->i_block_group);
+ group = find_group_other(sb, dir);
err = -ENOSPC;
if (group == -1)
@@ -349,6 +483,15 @@ repeat:
es->s_free_inodes_count =
cpu_to_le32(le32_to_cpu(es->s_free_inodes_count) - 1);
+
+ if (S_ISDIR(mode)) {
+ if (EXT2_SB(sb)->debts[i] < 255)
+ EXT2_SB(sb)->debts[i]++;
+ } else {
+ if (EXT2_SB(sb)->debts[i])
+ EXT2_SB(sb)->debts[i]--;
+ }
+
mark_buffer_dirty(EXT2_SB(sb)->s_sbh);
sb->s_dirt = 1;
inode->i_uid = current->fsuid;
@@ -478,6 +621,21 @@ unsigned long ext2_count_free_inodes (st
#endif
}
+/* Called at mount-time, super-block is locked */
+unsigned long ext2_count_dirs (struct super_block * sb)
+{
+ unsigned long count = 0;
+ int i;
+
+ for (i = 0; i < EXT2_SB(sb)->s_groups_count; i++) {
+ struct ext2_group_desc *gdp = ext2_get_group_desc (sb, i, NULL);
+ if (!gdp)
+ continue;
+ count += le16_to_cpu(gdp->bg_used_dirs_count);
+ }
+ return count;
+}
+
#ifdef CONFIG_EXT2_CHECK
/* Called at mount-time, super-block is locked */
void ext2_check_inodes_bitmap (struct super_block * sb)
--- 2.5.41/fs/ext2/super.c~orlov-allocator Mon Oct 7 14:31:58 2002
+++ 2.5.41-akpm/fs/ext2/super.c Mon Oct 7 14:52:38 2002
@@ -665,6 +665,12 @@ static int ext2_fill_super(struct super_
printk ("EXT2-fs: not enough memory\n");
goto failed_mount;
}
+ sbi->debts = kmalloc(sbi->s_groups_count, GFP_KERNEL);
+ if (!sbi->debts) {
+ printk ("EXT2-fs: not enough memory\n");
+ goto failed_mount_group_desc;
+ }
+ memset(sbi->debts, 0, sbi->s_groups_count);
for (i = 0; i < db_count; i++) {
sbi->s_group_desc[i] = sb_bread(sb, logic_sb_block + i + 1);
if (!sbi->s_group_desc[i]) {
@@ -681,6 +687,7 @@ static int ext2_fill_super(struct super_
goto failed_mount2;
}
sbi->s_gdb_count = db_count;
+ sbi->s_dir_count = ext2_count_dirs(sb);
get_random_bytes(&sbi->s_next_generation, sizeof(u32));
/*
* set up enough so that it can read an inode
@@ -706,6 +713,7 @@ static int ext2_fill_super(struct super_
failed_mount2:
for (i = 0; i < db_count; i++)
brelse(sbi->s_group_desc[i]);
+failed_mount_group_desc:
kfree(sbi->s_group_desc);
failed_mount:
brelse(bh);
--- 2.5.41/include/linux/ext2_fs_sb.h~orlov-allocator Mon Oct 7 14:32:07 2002
+++ 2.5.41-akpm/include/linux/ext2_fs_sb.h Mon Oct 7 14:38:23 2002
@@ -43,6 +43,8 @@ struct ext2_sb_info {
int s_inode_size;
int s_first_ino;
u32 s_next_generation;
+ unsigned long s_dir_count;
+ u8 *debts;
};
#endif /* _LINUX_EXT2_FS_SB */
--- 2.5.41/fs/ext2/ext2.h~orlov-allocator Mon Oct 7 14:37:36 2002
+++ 2.5.41-akpm/fs/ext2/ext2.h Mon Oct 7 14:37:51 2002
@@ -45,6 +45,7 @@ extern int ext2_new_block (struct inode
extern void ext2_free_blocks (struct inode *, unsigned long,
unsigned long);
extern unsigned long ext2_count_free_blocks (struct super_block *);
+extern unsigned long ext2_count_dirs (struct super_block *);
extern void ext2_check_blocks_bitmap (struct super_block *);
extern struct ext2_group_desc * ext2_get_group_desc(struct super_block * sb,
unsigned int block_group,
.
next prev parent reply other threads:[~2002-10-07 22:07 UTC|newest]
Thread overview: 206+ messages / expand[flat|nested] mbox.gz Atom feed top
2002-09-24 1:54 [PATCH-RFC] 4 of 4 - New problem logging macros, SCSI RAID device driver Larry Kessler
2002-09-24 2:22 ` Jeff Garzik
2002-09-26 15:52 ` Alan Cox
2002-09-26 22:55 ` [PATCH-RFC] 4 of 4 - New problem logging macros, SCSI RAIDdevice driver Larry Kessler
2002-09-26 22:58 ` Jeff Garzik
2002-09-26 23:07 ` Linus Torvalds
2002-09-27 2:27 ` Jeff Garzik
2002-09-27 4:45 ` Linus Torvalds
2002-09-28 7:46 ` Ingo Molnar
2002-09-28 9:16 ` jw schultz
2002-09-30 14:05 ` Denis Vlasenko
2002-09-30 10:22 ` Tomas Szepe
2002-09-30 11:10 ` jw schultz
2002-09-30 11:17 ` Adrian Bunk
2002-09-30 19:48 ` Rik van Riel
2002-09-30 20:30 ` Christoph Hellwig
2002-09-28 15:40 ` Kernel version [Was: Re: [PATCH-RFC] 4 of 4 - New problem logging macros, SCSI RAIDdevice driver] Horst von Brand
2002-09-29 1:31 ` v2.6 vs v3.0 Linus Torvalds
2002-09-29 6:14 ` james
2002-09-29 6:55 ` Andre Hedrick
2002-09-29 12:59 ` Gerhard Mack
2002-09-29 13:46 ` Dr. David Alan Gilbert
2002-09-29 14:06 ` Wakko Warner
2002-09-29 15:42 ` Jens Axboe
2002-09-29 16:21 ` Alan Cox
2002-09-29 16:17 ` Jens Axboe
2002-09-30 0:39 ` Jeff Chua
2002-09-29 16:22 ` Dave Jones
2002-09-29 16:26 ` Jens Axboe
2002-09-29 21:46 ` Matthias Andree
2002-09-30 7:05 ` Michael Clark
2002-09-30 7:22 ` Andrew Morton
2002-09-30 13:08 ` Kevin Corry
2002-09-30 13:05 ` Kevin Corry
2002-09-30 13:49 ` Michael Clark
2002-09-30 14:26 ` Kevin Corry
2002-09-30 13:59 ` Michael Clark
2002-09-30 15:50 ` Kevin Corry
2002-09-29 17:06 ` Jochen Friedrich
2002-09-29 15:18 ` Trever L. Adams
2002-09-29 15:45 ` Jens Axboe
2002-09-29 15:59 ` Trever L. Adams
2002-09-29 16:06 ` Jens Axboe
2002-09-29 16:13 ` Trever L. Adams
2002-09-30 6:54 ` Kai Henningsen
2002-09-30 18:40 ` Bill Davidsen
2002-10-01 12:38 ` Matthias Andree
2002-10-04 19:58 ` Bill Davidsen
2002-09-29 17:42 ` Linus Torvalds
2002-09-29 17:54 ` Rik van Riel
2002-09-29 18:24 ` Alan Cox
2002-09-30 7:56 ` Jens Axboe
2002-09-30 9:53 ` Andre Hedrick
2002-09-30 11:54 ` Jens Axboe
2002-09-30 12:58 ` Alan Cox
2002-09-30 13:05 ` Jens Axboe
2002-10-01 2:17 ` Andre Hedrick
2002-09-30 16:39 ` jbradford
2002-09-30 16:47 ` Pau Aliagas
2002-09-29 7:16 ` jbradford
2002-09-29 8:08 ` Jeff Garzik
2002-09-29 8:17 ` David S. Miller
2002-09-29 9:12 ` Jens Axboe
2002-09-29 11:19 ` Murray J. Root
2002-09-29 15:50 ` Jens Axboe
2002-09-30 7:01 ` Kai Henningsen
2002-09-29 16:04 ` Zwane Mwaikambo
2002-09-29 14:56 ` Alan Cox
2002-09-29 15:38 ` Jens Axboe
2002-09-29 16:30 ` Dave Jones
2002-09-29 16:42 ` Bjoern A. Zeeb
2002-09-29 21:16 ` Russell King
2002-09-29 21:32 ` Alan Cox
2002-09-29 21:49 ` steve
2002-09-29 21:52 ` Matthias Andree
2002-09-30 7:31 ` Tomas Szepe
2002-09-30 15:33 ` Jan Harkes
2002-09-30 18:13 ` Jeff Willis
2002-09-29 17:48 ` Linus Torvalds
2002-09-29 18:13 ` Jaroslav Kysela
2002-09-30 19:32 ` Bill Davidsen
2002-10-01 6:26 ` Jens Axboe
2002-10-01 7:54 ` Mikael Pettersson
2002-10-01 8:27 ` Jens Axboe
2002-10-01 8:44 ` jbradford
2002-10-01 11:31 ` Alan Cox
2002-10-01 11:25 ` Jens Axboe
2002-09-29 15:34 ` Andi Kleen
2002-09-29 17:26 ` Jochen Friedrich
2002-09-29 17:35 ` Jeff Garzik
2002-09-30 0:00 ` Andi Kleen
2002-10-01 19:28 ` IPv6 stability (success story ;) Petr Baudis
2002-09-29 9:15 ` v2.6 vs v3.0 Jens Axboe
2002-09-29 19:53 ` james
2002-09-29 15:26 ` Matthias Andree
2002-09-29 16:24 ` Alan Cox
2002-09-29 22:00 ` Matthias Andree
2002-09-30 19:02 ` Bill Davidsen
2002-09-30 18:37 ` Bill Davidsen
2002-10-03 15:51 ` [OT] 2.6 not 3.0 - (WAS Re: [PATCH-RFC] 4 of 4 - New problem logging macros, SCSI RAIDdevice) jbradford
2002-10-03 15:57 ` Linus Torvalds
2002-10-03 16:16 ` [OT] 2.6 not 3.0 - (WAS Re: [PATCH-RFC] 4 of 4 - New problem jbradford
2002-10-03 22:30 ` Greg KH
2002-10-04 6:33 ` jbradford
2002-10-04 6:37 ` Greg KH
2002-10-04 7:17 ` jbradford
2002-10-04 7:30 ` Greg KH
2002-10-03 16:37 ` [OT] 2.6 not 3.0 - (WAS Re: [PATCH-RFC] 4 of 4 - New problem logging macros, SCSI RAIDdevice) Alan Cox
2002-10-03 16:56 ` Linus Torvalds
2002-10-03 17:40 ` Alan Cox
2002-10-03 19:55 ` jlnance
2002-10-03 16:51 ` Dave Jones
2002-10-03 17:04 ` Alan Cox
2002-10-03 20:43 ` Andrew Morton
2002-10-03 22:05 ` Dave Jones
2002-10-04 3:46 ` Andreas Boman
2002-10-04 7:44 ` jbradford
2002-10-03 19:51 ` Rik van Riel
2002-10-04 22:26 ` [OT] 2.6 not 3.0 - (NUMA) Martin J. Bligh
2002-10-04 23:13 ` Linus Torvalds
2002-10-05 0:21 ` Martin J. Bligh
2002-10-05 0:36 ` Linus Torvalds
2002-10-05 1:25 ` Michael Hohnbaum
2002-10-05 20:30 ` The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 not 3.0 - (NUMA)) Rob Landley
2002-10-06 2:15 ` Andrew Morton
2002-10-06 9:42 ` Russell King
2002-10-06 17:06 ` Alan Cox
2002-10-06 13:44 ` Oliver Neukum
2002-10-06 15:19 ` Martin J. Bligh
2002-10-06 15:14 ` Oliver Neukum
2002-10-07 8:08 ` Helge Hafting
2002-10-07 9:18 ` Oliver Neukum
2002-10-07 14:11 ` Jan Hudec
2002-10-07 15:01 ` Jesse Pollard
2002-10-07 15:34 ` Jan Hudec
2002-10-08 3:12 ` [OT] " Scott Mcdermott
2002-10-10 23:49 ` Mike Fedyk
2002-10-07 15:15 ` Martin J. Bligh
2002-10-08 13:49 ` Helge Hafting
2002-10-07 17:43 ` Daniel Phillips
2002-10-07 18:31 ` Andrew Morton
2002-10-07 18:51 ` Linus Torvalds
2002-10-07 20:14 ` Alan Cox
2002-10-07 20:31 ` The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 not3.0 " Andrew Morton
2002-10-07 20:46 ` Linus Torvalds
2002-10-07 20:44 ` The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 not 3.0 " Linus Torvalds
2002-10-07 21:16 ` The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 not3.0 " Andrew Morton
2002-10-07 23:47 ` jw schultz
2002-10-11 0:02 ` Mike Fedyk
2002-10-07 18:58 ` The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 not 3.0 " Chris Friesen
2002-10-07 19:21 ` Daniel Phillips
2002-10-07 19:35 ` Linus Torvalds
2002-10-08 0:39 ` Theodore Ts'o
2002-10-08 2:59 ` Andrew Morton
2002-10-08 16:15 ` Theodore Ts'o
2002-10-08 19:39 ` Andrew Morton
2002-10-08 17:06 ` Rob Landley
2002-10-07 19:36 ` Andrew Morton
2002-10-08 2:36 ` Simon Kirby
2002-10-08 2:47 ` Daniel Phillips
2002-10-08 2:50 ` Andrew Morton
2002-10-08 2:54 ` Simon Kirby
2002-10-08 3:00 ` Andrew Morton
2002-10-08 16:17 ` Theodore Ts'o
2002-10-08 12:49 ` jlnance
2002-10-08 17:09 ` Andrew Morton
2002-10-10 20:53 ` Thomas Zimmerman
2002-10-08 13:54 ` Helge Hafting
2002-10-08 15:31 ` Andreas Dilger
2002-10-07 19:05 ` Daniel Phillips
2002-10-07 19:24 ` Linus Torvalds
2002-10-07 20:02 ` Daniel Phillips
2002-10-07 20:14 ` Andrew Morton
2002-10-07 20:22 ` Daniel Phillips
2002-10-07 20:28 ` Linus Torvalds
2002-10-07 21:16 ` Daniel Phillips
2002-10-07 21:55 ` Linus Torvalds
2002-10-07 22:02 ` Daniel Phillips
2002-10-07 22:12 ` Andrew Morton [this message]
2002-10-08 8:49 ` Padraig Brady
2002-10-07 22:14 ` Charles Cazabon
2002-10-30 18:26 ` Lee Leahu
2002-10-06 6:33 ` Martin J. Bligh
2002-10-07 5:28 ` John Alvord
2002-10-07 8:39 ` The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 n Giuliano Pochini
2002-10-07 13:56 ` The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 not 3.0 - (NUMA)) Jesse Pollard
2002-10-07 14:03 ` Rob Landley
2002-10-08 22:14 ` Jesse Pollard
2002-10-08 19:11 ` Rob Landley
2002-10-09 8:17 ` Alexander Kellett
2002-10-07 18:22 ` Daniel Phillips
2002-10-08 8:19 ` Jan Hudec
2002-10-11 23:53 ` Hans Reiser
2002-10-11 20:26 ` Rob Landley
2002-10-12 4:14 ` Nick LeRoy
2002-10-13 17:27 ` Rob Landley
2002-10-12 10:03 ` Hans Reiser
2002-10-13 17:32 ` Rob Landley
2002-10-13 23:51 ` Hans Reiser
2002-10-14 16:33 ` Rob Landley
2002-10-14 7:10 ` Nikita Danilov
2002-10-21 15:36 ` [OT] Please don't call it 3.0!! (was Re: The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 not 3.0 - (NUMA))) Calin A. Culianu
2002-10-21 16:20 ` Wakko Warner
2002-10-12 11:42 ` The reason to call it 3.0 is the desktop (was Re: [OT] 2.6 not 3.0 - (NUMA)) Matthias Andree
2002-10-12 14:56 ` Hugh Dickins
2002-09-27 11:32 ` [PATCH-RFC] 4 of 4 - New problem logging macros, SCSI RAIDdevice driver Alan Cox
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3DA206C3.9AD2941A@digeo.com \
--to=akpm@digeo.com \
--cc=landley@trommello.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mbligh@aracnet.com \
--cc=oliver@neukum.name \
--cc=phillips@arcor.de \
--cc=torvalds@transmeta.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).