From: Mingming Cao Subject: Re: [RFC PATCH -v2] ext4: Retry block reservation Date: Mon, 25 Aug 2008 14:06:33 -0700 Message-ID: <1219698393.6394.11.camel@mingming-laptop> References: <1219663233-21849-1-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <1219663233-21849-2-git-send-email-aneesh.kumar@linux.vnet.ibm.com> <1219663233-21849-3-git-send-email-aneesh.kumar@linux.vnet.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: tytso@mit.edu, sandeen@redhat.com, linux-ext4@vger.kernel.org To: "Aneesh Kumar K.V" Return-path: Received: from e2.ny.us.ibm.com ([32.97.182.142]:48682 "EHLO e2.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753538AbYHYVGi (ORCPT ); Mon, 25 Aug 2008 17:06:38 -0400 Received: from d01relay02.pok.ibm.com (d01relay02.pok.ibm.com [9.56.227.234]) by e2.ny.us.ibm.com (8.13.8/8.13.8) with ESMTP id m7PL6YE0002222 for ; Mon, 25 Aug 2008 17:06:34 -0400 Received: from d01av02.pok.ibm.com (d01av02.pok.ibm.com [9.56.224.216]) by d01relay02.pok.ibm.com (8.13.8/8.13.8/NCO v9.0) with ESMTP id m7PL6YQ2211842 for ; Mon, 25 Aug 2008 17:06:34 -0400 Received: from d01av02.pok.ibm.com (loopback [127.0.0.1]) by d01av02.pok.ibm.com (8.12.11.20060308/8.13.3) with ESMTP id m7PL6X7E022347 for ; Mon, 25 Aug 2008 17:06:34 -0400 In-Reply-To: <1219663233-21849-3-git-send-email-aneesh.kumar@linux.vnet.ibm.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: =E5=9C=A8 2008-08-25=E4=B8=80=E7=9A=84 16:50 +0530=EF=BC=8CAneesh Kumar= K.V=E5=86=99=E9=81=93=EF=BC=9A > During block reservation if we don't have enough > blocks left, retry block reservation with smaller > block count. This make sure we try fallocate > and DIO with smaller request size and don't fail early. > The delayed allocation reservation cannot try with smaller > block count. So retry block reservation to handle temporary > disk full conditions. Also print free blocks details if we > fail block allocation during writepages. >=20 > Signed-off-by: Aneesh Kumar K.V > --- > fs/ext4/balloc.c | 8 +++++++- > fs/ext4/inode.c | 33 ++++++++++++++++++++++++++++++--- > fs/ext4/mballoc.c | 7 ++++++- > 3 files changed, 43 insertions(+), 5 deletions(-) >=20 > diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c > index 4a53541..b7d1347 100644 > --- a/fs/ext4/balloc.c > +++ b/fs/ext4/balloc.c > @@ -1760,10 +1760,16 @@ ext4_fsblk_t ext4_old_new_blocks(handle_t *ha= ndle, struct inode *inode, > /* > * With delalloc we already reserved the blocks > */ > - if (ext4_claim_free_blocks(sbi, *count)) { > + while (*count && ext4_claim_free_blocks(sbi, *count)) { > + /* let others to free the space */ > + yield(); > + *count =3D *count >> 1; > + } > + if (!*count) { Could ext4_claim_free_blocks() returns the total number of blocks shortage, so that we don't need to "guess" how much to attempt to allocate? > *errp =3D -ENOSPC; > return 0; /*return with ENOSPC error */ > } > + num =3D *count; > } > /* > * Check quota for allocation of this block. > diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c > index d965a05..72a4a71 100644 > --- a/fs/ext4/inode.c > +++ b/fs/ext4/inode.c > @@ -1521,6 +1521,7 @@ static int ext4_journalled_write_end(struct fil= e *file, >=20 > static int ext4_da_reserve_space(struct inode *inode, int nrblocks) > { > + int retries =3D 0; > struct ext4_sb_info *sbi =3D EXT4_SB(inode->i_sb); > unsigned long md_needed, mdblocks, total =3D 0; >=20 > @@ -1529,6 +1530,7 @@ static int ext4_da_reserve_space(struct inode *= inode, int nrblocks) > * in order to allocate nrblocks > * worse case is one extent per block > */ > +repeat: > spin_lock(&EXT4_I(inode)->i_block_reservation_lock); > total =3D EXT4_I(inode)->i_reserved_data_blocks + nrblocks; > mdblocks =3D ext4_calc_metadata_amount(inode, total); > @@ -1539,6 +1541,10 @@ static int ext4_da_reserve_space(struct inode = *inode, int nrblocks) >=20 > if (ext4_claim_free_blocks(sbi, total)) { > spin_unlock(&EXT4_I(inode)->i_block_reservation_lock); > + if (ext4_should_retry_alloc(inode->i_sb, &retries)) { > + yield(); > + goto repeat; > + } > return -ENOSPC; > } > EXT4_I(inode)->i_reserved_data_blocks +=3D nrblocks; > @@ -1813,6 +1819,26 @@ static void ext4_da_block_invalidatepages(stru= ct mpage_da_data *mpd, > return; > } >=20 > +static void ext4_print_free_blocks(struct super_block *sb) > +{ > +#ifdef EXT4FS_DEBUG > + ext4_group_t i; > + struct ext4_group_desc *gdp; > + ext4_group_t ngroups =3D EXT4_SB(sb)->s_groups_count; > + printk(KERN_DEBUG "Free blocks details\n"); > + for (i =3D 0; i < ngroups; i++) { > + gdp =3D ext4_get_group_desc(sb, i, NULL); > + if (!gdp) > + continue; > + printk(KERN_DEBUG "Free blocks in group %lu is %d\n", > + i, le16_to_cpu(gdp->bg_free_blocks_count)); > + } > +#endif > + printk(KERN_CRIT "Total free blocks count %lld\n", > + ext4_count_free_blocks(sb)); > + return; > +} > + Ext4_count_free_blocks() has the extra debugging code too. I think you could calling ext4_count_free_blocks() directly and don't need the new function ext4_print_free_blocks(). > /* > * mpage_da_map_blocks - go through given space > * > @@ -1825,20 +1851,18 @@ static void ext4_da_block_invalidatepages(str= uct mpage_da_data *mpd, > static int mpage_da_map_blocks(struct mpage_da_data *mpd) > { > int err =3D 0; > + struct buffer_head new; > struct buffer_head *lbh =3D &mpd->lbh; > sector_t next =3D lbh->b_blocknr; > - struct buffer_head new; >=20 > /* > * We consider only non-mapped and non-allocated blocks > */ > if (buffer_mapped(lbh) && !buffer_delay(lbh)) > return 0; > - > new.b_state =3D lbh->b_state; > new.b_blocknr =3D 0; > new.b_size =3D lbh->b_size; > - > /* > * If we didn't accumulate anything > * to write simply return > @@ -1871,6 +1895,9 @@ static int mpage_da_map_blocks(struct mpage_da= _data *mpd) > lbh->b_size >> mpd->inode->i_blkbits, err); > printk(KERN_EMERG "This should not happen.!! " > "Data will be lost\n"); could we print out the error code with the warning message here? > + if (err =3D=3D -ENOSPC) { > + ext4_print_free_blocks(mpd->inode->i_sb); > + } > /* invlaidate all the pages */ > ext4_da_block_invalidatepages(mpd, next, > lbh->b_size >> mpd->inode->i_blkbits); > diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c > index 4404b46..419009f 100644 > --- a/fs/ext4/mballoc.c > +++ b/fs/ext4/mballoc.c > @@ -4397,7 +4397,12 @@ ext4_fsblk_t ext4_mb_new_blocks(handle_t *hand= le, > /* > * With delalloc we already reserved the blocks > */ > - if (ext4_claim_free_blocks(sbi, ar->len)) { > + while (ar->len && ext4_claim_free_blocks(sbi, ar->len)) { > + /* let others to free the space */ > + yield(); > + ar->len =3D ar->len >> 1; > + } > + if (!ar->len) { > *errp =3D -ENOSPC; > return 0; > } -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html