| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <alpine.LFD.2.00.1108151612120.19095@dhcp-27-109.brq.redhat.com>
Date: Mon, 15 Aug 2011 16:21:27 +0200 (CEST)
From: Lukas Czerner <lczerner@...hat.com>
To: Andreas Dilger <adilger@...ger.ca>
cc: Lukas Czerner <lczerner@...hat.com>, linux-ext4@...r.kernel.org,
tytso@....edu, Jan Kara <jack@...e.cz>
Subject: Re: [PATCH] ext4: remove deprecated oldalloc
On Thu, 11 Aug 2011, Andreas Dilger wrote:
> On 2011-08-11, at 8:58 AM, Lukas Czerner wrote:
> > On Tue, 7 Jun 2011, Lukas Czerner wrote:
> >> For a long time now orlov is the default block allocator in the ext4. It
> >> performs better than the old one and no one seems to claim otherwise so
> >> we can safely drop it and make oldalloc and orlov mount option
> >> deprecated.
> >>
> >> This is a part of the effort to reduce number of ext4 options hence the
> >> test matrix.
> >>
> >> Signed-off-by: Lukas Czerner <lczerner@...hat.com>
> >
> > ping
>
> I'm OK with removing this, I don't think anyone uses it, and it has almost
> no meaning with flex_bg anyway.
>
> That said, "orlov" is also mostly meaningless with flex_bg as well, since
> there is very little real benefit/affinity from inodes being "close" to
> their data blocks. We gain far more benefit from keeping the inodes
> together than spreading them out and keeping them close to the data blocks.
What about removing it for ext3 as well ? I can prepare a patch.
Also note that there is a bug in the OLDALLOC where if there is
approximately the same number of inodes in all of the allocation groups
it might result in the state where no group has less free inode count
than the average, hence we get ENOSPC even though there is enough space
for the inode to be allocated. It is unlikely, but it is there.
So Ted, could you take the patch ?
Thanks!
-Lukas
>
>
> >> ---
> >> Documentation/filesystems/ext4.txt | 8 --
> >> fs/ext4/ext4.h | 1 -
> >> fs/ext4/ialloc.c | 136 +-----------------------------------
> >> fs/ext4/super.c | 8 +-
> >> 4 files changed, 7 insertions(+), 146 deletions(-)
> >>
> >> diff --git a/Documentation/filesystems/ext4.txt b/Documentation/filesystems/ext4.txt
> >> index 3ae9bc9..ec469fa 100644
> >> --- a/Documentation/filesystems/ext4.txt
> >> +++ b/Documentation/filesystems/ext4.txt
> >> @@ -201,14 +201,6 @@ inode_readahead_blks=n This tuning parameter controls the maximum
> >> table readahead algorithm will pre-read into
> >> the buffer cache. The default value is 32 blocks.
> >>
> >> -orlov (*) This enables the new Orlov block allocator. It is
> >> - enabled by default.
> >> -
> >> -oldalloc This disables the Orlov block allocator and enables
> >> - the old block allocator. Orlov should have better
> >> - performance - we'd like to get some feedback if it's
> >> - the contrary for you.
> >> -
> >> user_xattr Enables Extended User Attributes. Additionally, you
> >> need to have extended attribute support enabled in the
> >> kernel configuration (CONFIG_EXT4_FS_XATTR). See the
> >> diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
> >> index 1921392..7e0b8aa 100644
> >> --- a/fs/ext4/ext4.h
> >> +++ b/fs/ext4/ext4.h
> >> @@ -884,7 +884,6 @@ struct ext4_inode_info {
> >> /*
> >> * Mount flags
> >> */
> >> -#define EXT4_MOUNT_OLDALLOC 0x00002 /* Don't use the new Orlov allocator */
> >> #define EXT4_MOUNT_GRPID 0x00004 /* Create files with directory's group */
> >> #define EXT4_MOUNT_DEBUG 0x00008 /* Some debugging messages */
> >> #define EXT4_MOUNT_ERRORS_CONT 0x00010 /* Continue on errors */
> >> diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
> >> index 21bb2f6..0b5ec23 100644
> >> --- a/fs/ext4/ialloc.c
> >> +++ b/fs/ext4/ialloc.c
> >> @@ -293,118 +293,6 @@ error_return:
> >> ext4_std_error(sb, fatal);
> >> }
> >>
> >> -/*
> >> - * There are two policies for allocating an inode. If the new inode is
> >> - * a directory, then a forward search is made for a block group with both
> >> - * free space and a low directory-to-inode ratio; if that fails, then of
> >> - * the groups with above-average free space, that group with the fewest
> >> - * directories already is chosen.
> >> - *
> >> - * For other inodes, search forward from the parent directory\'s block
> >> - * group to find a free inode.
> >> - */
> >> -static int find_group_dir(struct super_block *sb, struct inode *parent,
> >> - ext4_group_t *best_group)
> >> -{
> >> - ext4_group_t ngroups = ext4_get_groups_count(sb);
> >> - unsigned int freei, avefreei;
> >> - struct ext4_group_desc *desc, *best_desc = NULL;
> >> - ext4_group_t group;
> >> - int ret = -1;
> >> -
> >> - freei = percpu_counter_read_positive(&EXT4_SB(sb)->s_freeinodes_counter);
> >> - avefreei = freei / ngroups;
> >> -
> >> - for (group = 0; group < ngroups; group++) {
> >> - desc = ext4_get_group_desc(sb, group, NULL);
> >> - if (!desc || !ext4_free_inodes_count(sb, desc))
> >> - continue;
> >> - if (ext4_free_inodes_count(sb, desc) < avefreei)
> >> - continue;
> >> - if (!best_desc ||
> >> - (ext4_free_blks_count(sb, desc) >
> >> - ext4_free_blks_count(sb, best_desc))) {
> >> - *best_group = group;
> >> - best_desc = desc;
> >> - ret = 0;
> >> - }
> >> - }
> >> - return ret;
> >> -}
> >> -
> >> -#define free_block_ratio 10
> >> -
> >> -static int find_group_flex(struct super_block *sb, struct inode *parent,
> >> - ext4_group_t *best_group)
> >> -{
> >> - struct ext4_sb_info *sbi = EXT4_SB(sb);
> >> - struct ext4_group_desc *desc;
> >> - struct flex_groups *flex_group = sbi->s_flex_groups;
> >> - ext4_group_t parent_group = EXT4_I(parent)->i_block_group;
> >> - ext4_group_t parent_fbg_group = ext4_flex_group(sbi, parent_group);
> >> - ext4_group_t ngroups = ext4_get_groups_count(sb);
> >> - int flex_size = ext4_flex_bg_size(sbi);
> >> - ext4_group_t best_flex = parent_fbg_group;
> >> - int blocks_per_flex = sbi->s_blocks_per_group * flex_size;
> >> - int flexbg_free_blocks;
> >> - int flex_freeb_ratio;
> >> - ext4_group_t n_fbg_groups;
> >> - ext4_group_t i;
> >> -
> >> - n_fbg_groups = (ngroups + flex_size - 1) >>
> >> - sbi->s_log_groups_per_flex;
> >> -
> >> -find_close_to_parent:
> >> - flexbg_free_blocks = atomic_read(&flex_group[best_flex].free_blocks);
> >> - flex_freeb_ratio = flexbg_free_blocks * 100 / blocks_per_flex;
> >> - if (atomic_read(&flex_group[best_flex].free_inodes) &&
> >> - flex_freeb_ratio > free_block_ratio)
> >> - goto found_flexbg;
> >> -
> >> - if (best_flex && best_flex == parent_fbg_group) {
> >> - best_flex--;
> >> - goto find_close_to_parent;
> >> - }
> >> -
> >> - for (i = 0; i < n_fbg_groups; i++) {
> >> - if (i == parent_fbg_group || i == parent_fbg_group - 1)
> >> - continue;
> >> -
> >> - flexbg_free_blocks = atomic_read(&flex_group[i].free_blocks);
> >> - flex_freeb_ratio = flexbg_free_blocks * 100 / blocks_per_flex;
> >> -
> >> - if (flex_freeb_ratio > free_block_ratio &&
> >> - (atomic_read(&flex_group[i].free_inodes))) {
> >> - best_flex = i;
> >> - goto found_flexbg;
> >> - }
> >> -
> >> - if ((atomic_read(&flex_group[best_flex].free_inodes) == 0) ||
> >> - ((atomic_read(&flex_group[i].free_blocks) >
> >> - atomic_read(&flex_group[best_flex].free_blocks)) &&
> >> - atomic_read(&flex_group[i].free_inodes)))
> >> - best_flex = i;
> >> - }
> >> -
> >> - if (!atomic_read(&flex_group[best_flex].free_inodes) ||
> >> - !atomic_read(&flex_group[best_flex].free_blocks))
> >> - return -1;
> >> -
> >> -found_flexbg:
> >> - for (i = best_flex * flex_size; i < ngroups &&
> >> - i < (best_flex + 1) * flex_size; i++) {
> >> - desc = ext4_get_group_desc(sb, i, NULL);
> >> - if (ext4_free_inodes_count(sb, desc)) {
> >> - *best_group = i;
> >> - goto out;
> >> - }
> >> - }
> >> -
> >> - return -1;
> >> -out:
> >> - return 0;
> >> -}
> >> -
> >> struct orlov_stats {
> >> __u32 free_inodes;
> >> __u32 free_blocks;
> >> @@ -817,7 +705,6 @@ struct inode *ext4_new_inode(handle_t *handle, struct inode *dir, int mode,
> >> struct inode *ret;
> >> ext4_group_t i;
> >> int free = 0;
> >> - static int once = 1;
> >> ext4_group_t flex_group;
> >>
> >> /* Cannot create files in a deleted directory */
> >> @@ -843,26 +730,9 @@ struct inode *ext4_new_inode(handle_t *handle, struct inode *dir, int mode,
> >> goto got_group;
> >> }
> >>
> >> - if (sbi->s_log_groups_per_flex && test_opt(sb, OLDALLOC)) {
> >> - ret2 = find_group_flex(sb, dir, &group);
> >> - if (ret2 == -1) {
> >> - ret2 = find_group_other(sb, dir, &group, mode);
> >> - if (ret2 == 0 && once) {
> >> - once = 0;
> >> - printk(KERN_NOTICE "ext4: find_group_flex "
> >> - "failed, fallback succeeded dir %lu\n",
> >> - dir->i_ino);
> >> - }
> >> - }
> >> - goto got_group;
> >> - }
> >> -
> >> - if (S_ISDIR(mode)) {
> >> - if (test_opt(sb, OLDALLOC))
> >> - ret2 = find_group_dir(sb, dir, &group);
> >> - else
> >> - ret2 = find_group_orlov(sb, dir, &group, mode, qstr);
> >> - } else
> >> + if (S_ISDIR(mode))
> >> + ret2 = find_group_orlov(sb, dir, &group, mode, qstr);
> >> + else
> >> ret2 = find_group_other(sb, dir, &group, mode);
> >>
> >> got_group:
> >> diff --git a/fs/ext4/super.c b/fs/ext4/super.c
> >> index cc5c157..e1f8f73 100644
> >> --- a/fs/ext4/super.c
> >> +++ b/fs/ext4/super.c
> >> @@ -1031,8 +1031,6 @@ static int ext4_show_options(struct seq_file *seq, struct vfsmount *vfs)
> >> seq_puts(seq, ",nouid32");
> >> if (test_opt(sb, DEBUG) && !(def_mount_opts & EXT4_DEFM_DEBUG))
> >> seq_puts(seq, ",debug");
> >> - if (test_opt(sb, OLDALLOC))
> >> - seq_puts(seq, ",oldalloc");
> >> #ifdef CONFIG_EXT4_FS_XATTR
> >> if (test_opt(sb, XATTR_USER))
> >> seq_puts(seq, ",user_xattr");
> >> @@ -1541,10 +1539,12 @@ static int parse_options(char *options, struct super_block *sb,
> >> set_opt(sb, DEBUG);
> >> break;
> >> case Opt_oldalloc:
> >> - set_opt(sb, OLDALLOC);
> >> + ext4_msg(sb, KERN_WARNING,
> >> + "Ignoring deprecated oldalloc option");
> >> break;
> >> case Opt_orlov:
> >> - clear_opt(sb, OLDALLOC);
> >> + ext4_msg(sb, KERN_WARNING,
> >> + "Ignoring deprecated orlov option");
> >> break;
> >> #ifdef CONFIG_EXT4_FS_XATTR
> >> case Opt_user_xattr:
> >>
> >
> > --
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> > the body of a message to majordomo@...r.kernel.org
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
>
>
> Cheers, Andreas
>
>
>
>
>
>
--
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists