New disk format in -unstable (compression!)

New disk format in -unstable (compression!)

[Chris Mason]

Hello everyone,

I've pushed out the compression code along with a new disk format to the
unstable branches.  A while back I also created a stand alone btrfs repo
that is automatically generated from the unstable git repo (with some
help from David Woodhouse's script).

You can find the standalone repo here:

http://git.kernel.org/?p=linux/kernel/git/mason/btrfs-unstable-standalone.git;a=summary

And the full kernel repo here:

http://git.kernel.org/?p=linux/kernel/git/mason/btrfs-unstable.git;a=summary

I managed to miss the guilt refresh on the compression patch (it was
missing a good description in the commit log) and had to do a quick
rebase of the last few commits on the kernel-unstable repos.  There was
only a 10 minute window where the mistake was in there.

Compression is off by default and enabled by mount -o compress.  Even
when the -o compress mount option is not used, it is possible to read
compressed extents off the disk.

If compression for a given set of pages fails to make them smaller, the
file is flagged to avoid future compression attempts later.

I made some big changes to the writeback paths:

* While finding delalloc extents, the pages are locked before being sent down
to the delalloc handler.  This allows the delalloc handler to do complex things
such as cleaning the pages, marking them writeback and starting IO on their
behalf.

* Inline extents are inserted at delalloc time now.  This allows us to compress
the data before inserting the inline extent, and it allows us to insert
an inline extent that spans multiple pages.

* All of the in-memory extent representations (extent_map.c, ordered-data.c etc)
are changed to record both an in-memory size and an on disk size, as well
as a flag for compression.

>From a disk format point of view, the extent pointers in the file are changed
to record the on disk size of a given extent and some encoding flags.
Space in the disk format is allocated for compression encoding, as well
as encryption and a generic 'other' field.  Neither the encryption or the
'other' field are currently used.

In order to limit the amount of data read for a single random read in the
file, the size of a compressed extent is limited to 128k.  This is a
software only limit, the disk format supports u64 sized compressed extents.

In order to limit the ram consumed while processing extents, the uncompressed
size of a compressed extent is limited to 256k.  This is a software only limit
and will be subject to tuning later.

Checksumming is still done on compressed extents, and it is done on the
uncompressed version of the data.  This way additional encodings can be
layered on without having to figure out which encoding to checksum.

Compression happens at delalloc time, which is basically singled threaded because
it is usually done by a single pdflush thread.  This makes it tricky to
spread the compression load across all the cpus on the box.  We'll have to
look at parallel pdflush walks of dirty inodes at a later time.

Decompression is hooked into readpages and it does spread across CPUs nicely.

-chris

Re: New disk format in -unstable (compression!)

[Paul P Komkoff Jr]

Replying to Chris Mason:
> Hello everyone,

Hi.

> I've pushed out the compression code along with a new disk format to the

First!

This is when restoring a tar file with 2.5M small files.

btrfs: use compression
BUG: unable to handle kernel paging request at ffffffff812a0d4d
IP: [<ffffffffa04fed42>] btrfs_submit_compressed_write+0x120/0x262 [btrfs]
PGD 203067 PUD 207063 PMD 21c11f161 PTE 4a0161
Oops: 0003 [1] SMP 
CPU 1 
Modules linked in: btrfs deflate zlib_deflate crc32c libcrc32c fuse ipv6 cpufreq_ondemand acpi_cpufreq freq_table dm_multipath raid1 snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq ppdev snd_seq_device snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore floppy snd_page_alloc pcspkr i2c_i801 radeon drm parport_pc parport i2c_algo_bit iTCO_wdt i3000_edac iTCO_vendor_support edac_core i2c_core shpchp e1000e raid456 async_xor async_memcpy async_tx xor raid10 [last unloaded: microcode]
Pid: 247, comm: pdflush Not tainted 2.6.27.4-61.fc10.x86_64 #1
RIP: 0010:[<ffffffffa04fed42>]  [<ffffffffa04fed42>] btrfs_submit_compressed_write+0x120/0x262 [btrfs]
RSP: 0018:ffff88021cdc17a0  EFLAGS: 00010206
RAX: ffff88018f438638 RBX: ffff88021c498300 RCX: 0000000000001000
RDX: 0000000000000000 RSI: ffff88021c498300 RDI: ffff88021c0a0f30
RBP: ffff88021cdc1800 R08: 0000000000000000 R09: 0000000000000400
R10: ffff88021c498300 R11: 0000080000000001 R12: ffff8801eed37120
R13: ffffffff812a0d35 R14: 0000000041d71000 R15: ffff88018f438528
FS:  0000000000000000(0000) GS:ffff88021fc04980(0000) knlGS:0000000000000000
CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
CR2: ffffffff812a0d4d CR3: 00000002190ea000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process pdflush (pid: 247, threadinfo ffff88021cdc0000, task ffff88021f171740)
Stack:  000000000001c000 0000000041d61000 ffff880218156c00 000000000000c000
 ffff88021ec27000 ffff88018f4383b0 ffff880016089c80 000000000001c000
 0000000000000000 0000000000040000 0000000000040000 0000000000040000
Call Trace:
 [<ffffffffa04e0136>] cow_file_range+0x6e0/0x7c7 [btrfs]
 [<ffffffffa04e0542>] run_delalloc_range+0x325/0x33b [btrfs]
 [<ffffffffa04f1d1d>] ? find_lock_delalloc_range+0xfc/0x151 [btrfs]
 [<ffffffffa04f266b>] __extent_writepage+0x14e/0x648 [btrfs]
 [<ffffffff8116ffc1>] ? __lookup_tag+0xa9/0x110
 [<ffffffff8103baac>] ? try_to_wake_up+0x26f/0x281
 [<ffffffff8109d344>] ? __dec_zone_page_state+0x29/0x2b
 [<ffffffffa04efe73>] extent_write_cache_pages+0x1dd/0x341 [btrfs]
 [<ffffffffa04f251d>] ? __extent_writepage+0x0/0x648 [btrfs]
 [<ffffffffa04f0006>] extent_writepages+0x2f/0x51 [btrfs]
 [<ffffffffa04de375>] ? btrfs_get_extent+0x0/0x733 [btrfs]
 [<ffffffffa04de24e>] btrfs_writepages+0x23/0x25 [btrfs]
 [<ffffffff81097bad>] do_writepages+0x28/0x38
 [<ffffffff810df344>] __writeback_single_inode+0x185/0x2f9
 [<ffffffff81010a07>] ? restore_args+0x0/0x30
 [<ffffffff810df89d>] generic_sync_sb_inodes+0x229/0x309
 [<ffffffff810dfc06>] writeback_inodes+0xa4/0xfd
 [<ffffffff810981ce>] background_writeout+0x92/0xcb
 [<ffffffff81098756>] pdflush+0x171/0x234
 [<ffffffff8109813c>] ? background_writeout+0x0/0xcb
 [<ffffffff810985e5>] ? pdflush+0x0/0x234
 [<ffffffff810985e5>] ? pdflush+0x0/0x234
 [<ffffffff8105684d>] kthread+0x49/0x76
 [<ffffffff810116e9>] child_rip+0xa/0x11
 [<ffffffff81010a07>] ? restore_args+0x0/0x30
 [<ffffffff81056804>] ? kthread+0x0/0x76
 [<ffffffff810116df>] ? child_rip+0x0/0x11


Code: ea 4f a0 f0 41 ff 04 24 48 8b 55 a0 4c 89 75 d0 4c 8b 75 a8 48 89 55 b8 e9 e9 00 00 00 48 8b 45 d0 4c 8b 28 49 8b 87 08 01 00 00 <49> 89 45 18 83 7b 30 00 74 1f 48 8b 55 c8 45 31 c0 31 f6 48 89 
RIP  [<ffffffffa04fed42>] btrfs_submit_compressed_write+0x120/0x262 [btrfs]
 RSP <ffff88021cdc17a0>
CR2: ffffffff812a0d4d
---[ end trace 1844b0f2613c00dd ]---
general protection fault: 0000 [2] SMP 
CPU 1 
Modules linked in: btrfs deflate zlib_deflate crc32c libcrc32c fuse ipv6 cpufreq_ondemand acpi_cpufreq freq_table dm_multipath raid1 snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq ppdev snd_seq_device snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore floppy snd_page_alloc pcspkr i2c_i801 radeon drm parport_pc parport i2c_algo_bit iTCO_wdt i3000_edac iTCO_vendor_support edac_core i2c_core shpchp e1000e raid456 async_xor async_memcpy async_tx xor raid10 [last unloaded: microcode]
Pid: 6319, comm: tar Tainted: G      D   2.6.27.4-61.fc10.x86_64 #1
RIP: 0010:[<ffffffff810bc782>]  [<ffffffff810bc782>] kmem_cache_alloc+0x56/0xc6
RSP: 0018:ffff88021c409688  EFLAGS: 00010082
RAX: 0000000000000000 RBX: c5ffff8801eed37f RCX: 0000000000001000
RDX: ffff8800280531b0 RSI: 0000000000000050 RDI: 0000000000000060
RBP: ffff88021c4096b8 R08: 0000000000000000 R09: 0000000000000400
R10: ffff880192627800 R11: 0000000000010000 R12: 0000000000000296
R13: ffffffff816475f0 R14: ffffffffa04d8552 R15: 0000000000000050
FS:  00007f97c2c99780(0000) GS:ffff88021fc04980(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: ffffffff812a0d4d CR3: 0000000219121000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process tar (pid: 6319, threadinfo ffff88021c408000, task ffff8801f15dc5c0)
Stack:  0000006042001000 ffff880192627800 ffff88021a832000 0000000000000000
 0000000042002000 ffff880218408178 ffff88021c4096e8 ffffffffa04d8552
 0000000042001000 ffff880192627800 ffff8801eed37120 ffffe2000596e1d8
Call Trace:
 [<ffffffffa04d8552>] btrfs_bio_wq_end_io+0x27/0x6d [btrfs]
 [<ffffffffa04fee41>] btrfs_submit_compressed_write+0x21f/0x262 [btrfs]
 [<ffffffffa04e0136>] cow_file_range+0x6e0/0x7c7 [btrfs]
 [<ffffffffa04e0542>] run_delalloc_range+0x325/0x33b [btrfs]
 [<ffffffffa04f1d1d>] ? find_lock_delalloc_range+0xfc/0x151 [btrfs]
 [<ffffffffa04f266b>] __extent_writepage+0x14e/0x648 [btrfs]
 [<ffffffff8116ffc1>] ? __lookup_tag+0xa9/0x110
 [<ffffffff8109d344>] ? __dec_zone_page_state+0x29/0x2b
 [<ffffffffa04efe73>] extent_write_cache_pages+0x1dd/0x341 [btrfs]
 [<ffffffffa04f251d>] ? __extent_writepage+0x0/0x648 [btrfs]
 [<ffffffffa04f0006>] extent_writepages+0x2f/0x51 [btrfs]
 [<ffffffffa04de375>] ? btrfs_get_extent+0x0/0x733 [btrfs]
 [<ffffffffa04de24e>] btrfs_writepages+0x23/0x25 [btrfs]
 [<ffffffff81097bad>] do_writepages+0x28/0x38
 [<ffffffff810df344>] __writeback_single_inode+0x185/0x2f9
 [<ffffffff8116f6b8>] ? prop_fraction_single+0x3c/0x5e
 [<ffffffff810df89d>] generic_sync_sb_inodes+0x229/0x309
 [<ffffffff810dfc06>] writeback_inodes+0xa4/0xfd
 [<ffffffff810983f5>] balance_dirty_pages_ratelimited_nr+0x15a/0x285
 [<ffffffffa04e4ff5>] btrfs_file_write+0x471/0x64c [btrfs]
 [<ffffffff810c267e>] vfs_write+0xab/0x105
 [<ffffffff810c279c>] sys_write+0x47/0x6f
 [<ffffffff8101024a>] system_call_fastpath+0x16/0x1b


Code: 00 00 00 e8 e1 27 fd ff 65 8b 04 25 24 00 00 00 48 98 49 8b 94 c5 f0 10 00 00 8b 7a 18 89 7d d4 48 8b 1a 48 85 db 74 0c 8b 42 14 <48> 8b 04 c3 48 89 02 eb 17 49 89 d0 4c 89 f1 83 ca ff 44 89 fe 
RIP  [<ffffffff810bc782>] kmem_cache_alloc+0x56/0xc6
 RSP <ffff88021c409688>
---[ end trace 1844b0f2613c00dd ]---
general protection fault: 0000 [3] SMP 
CPU 1 
Modules linked in: btrfs deflate zlib_deflate crc32c libcrc32c fuse ipv6 cpufreq_ondemand acpi_cpufreq freq_table dm_multipath raid1 snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq ppdev snd_seq_device snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore floppy snd_page_alloc pcspkr i2c_i801 radeon drm parport_pc parport i2c_algo_bit iTCO_wdt i3000_edac iTCO_vendor_support edac_core i2c_core shpchp e1000e raid456 async_xor async_memcpy async_tx xor raid10 [last unloaded: microcode]
Pid: 6264, comm: btrfs-transacti Tainted: G      D   2.6.27.4-61.fc10.x86_64 #1
RIP: 0010:[<ffffffff810bc782>]  [<ffffffff810bc782>] kmem_cache_alloc+0x56/0xc6
RSP: 0018:ffff8802191a9b80  EFLAGS: 00010082
RAX: 0000000000000000 RBX: c5ffff8801eed37f RCX: ffff88021995c880
RDX: ffff8800280531b0 RSI: 0000000000000050 RDI: 0000000000000060
RBP: ffff8802191a9bb0 R08: 0000000000000000 R09: 0000000000000000
R10: ffff88006c5b8178 R11: ffff8801f92969b0 R12: 0000000000000282
R13: ffffffff816475f0 R14: ffffffffa04d9fe2 R15: 0000000000000050
FS:  0000000000000000(0000) GS:ffff88021fc04980(0000) knlGS:0000000000000000
CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
CR2: 00007f1d11438000 CR3: 000000021c810000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process btrfs-transacti (pid: 6264, threadinfo ffff8802191a8000, task ffff8802191c5d00)
Stack:  00000060191a9cd0 ffff8802191a9cd0 ffff88021a832000 ffff88006c5b8178
 0000000000000001 ffff88021995c880 ffff8802191a9c40 ffffffffa04d9fe2
 0000000000000000 0000000000000000 0000000000000004 0000000407640520
Call Trace:
 [<ffffffffa04d9fe2>] btrfs_wq_submit_bio+0x4e/0x242 [btrfs]
 [<ffffffffa04dae66>] btree_submit_bio_hook+0x44/0x46 [btrfs]
 [<ffffffffa04dad6c>] ? __btree_submit_bio_hook+0x0/0xb6 [btrfs]
 [<ffffffffa04eebf0>] submit_one_bio+0x61/0x8b [btrfs]
 [<ffffffffa04f2c11>] extent_write_full_page+0xac/0xbc [btrfs]
 [<ffffffffa04dae68>] ? btree_get_extent+0x0/0x1e0 [btrfs]
 [<ffffffffa04d841b>] btree_writepage+0x52/0x57 [btrfs]
 [<ffffffff810977ec>] write_one_page+0x88/0xd7
 [<ffffffffa04dbd09>] btrfs_write_and_wait_marked_extents+0xc2/0x1c7 [btrfs]
 [<ffffffffa04dbe4b>] btrfs_write_and_wait_transaction+0x3d/0x3f [btrfs]
 [<ffffffffa04dcbaa>] btrfs_commit_transaction+0x50a/0x6a0 [btrfs]
 [<ffffffff81056ba1>] ? autoremove_wake_function+0x0/0x38
 [<ffffffffa04d872d>] transaction_kthread+0x195/0x233 [btrfs]
 [<ffffffffa04d8598>] ? transaction_kthread+0x0/0x233 [btrfs]
 [<ffffffff8105684d>] kthread+0x49/0x76
 [<ffffffff810116e9>] child_rip+0xa/0x11
 [<ffffffff81010a07>] ? restore_args+0x0/0x30
 [<ffffffff81056804>] ? kthread+0x0/0x76
 [<ffffffff810116df>] ? child_rip+0x0/0x11


Code: 00 00 00 e8 e1 27 fd ff 65 8b 04 25 24 00 00 00 48 98 49 8b 94 c5 f0 10 00 00 8b 7a 18 89 7d d4 48 8b 1a 48 85 db 74 0c 8b 42 14 <48> 8b 04 c3 48 89 02 eb 17 49 89 d0 4c 89 f1 83 ca ff 44 89 fe 
RIP  [<ffffffff810bc782>] kmem_cache_alloc+0x56/0xc6
 RSP <ffff8802191a9b80>
---[ end trace 1844b0f2613c00dd ]---

-- 
Paul P 'Stingray' Komkoff Jr // http://stingr.net/key <- my pgp key
 This message represents the official view of the voices in my head

Re: New disk format in -unstable (compression!)

[Chris Mason]

On Thu, 2008-10-30 at 16:01 +0300, Paul P Komkoff Jr wrote:
> Replying to Chris Mason:
> > Hello everyone,
> 
> Hi.
> 
> > I've pushed out the compression code along with a new disk format to the
> 
> First!
> 
> This is when restoring a tar file with 2.5M small files.
> 

Thanks, any chance I can get my hands on this magic tar file?  If not
how big were the files?

-chris