On 18.04.2018 18:10, Brendan Hide wrote:
> Hi, all
>
> I'm looking for some advice re compression with NVME. Compression helps
> performance with a minor CPU hit - but is it still worth it with the far
> higher throughputs offered by newer PCI and NVME-type SSDs?
>
> I've ordered a PCIe-to-M.2
On Wed, Apr 18, 2018 at 02:59:25PM +0800, Anand Jain wrote:
> Adds comments about BTRFS_FS_EXCL_OP to existing comments
> about the device locks.
>
> Signed-off-by: Anand Jain
Reviewed-by: David Sterba
--
To unsubscribe from this list: send the line
Hi, all
I'm looking for some advice re compression with NVME. Compression helps
performance with a minor CPU hit - but is it still worth it with the far
higher throughputs offered by newer PCI and NVME-type SSDs?
I've ordered a PCIe-to-M.2 adapter along with a 1TB 960 Evo drive for my
home
On Tue, Apr 17, 2018 at 12:18:45PM +0200, René Rebe wrote:
> On 16 Apr 2018, at 21:15, David Sterba wrote:
> > On Mon, Apr 16, 2018 at 09:55:45PM +0200, René Rebe wrote:
> >> On 04/16/2018 06:48 PM, David Sterba wrote:
> >>> The warnings are valid, there's unaligned access
The last update to readdir introduced a temporary buffer to store the
emitted readdir data, but as there are file names of variable length,
there's a lot of unaligned access.
This was observed on a sparc64 machine:
Kernel unaligned access at TPC[102f3080] btrfs_real_readdir+0x51c/0x718
On Tue, Apr 17, 2018 at 04:19:04PM -0700, Howard McLauchlan wrote:
> Presently, convert_free_space_to_extents() does a linear scan of the
> bitmap. We can speed this up with find_next_{bit,zero_bit}_le().
>
> This patch replaces the linear scan with find_next_{bit,zero_bit}_le().
> Testing shows
On 2018-04-18 11:10, Brendan Hide wrote:
Hi, all
I'm looking for some advice re compression with NVME. Compression helps
performance with a minor CPU hit - but is it still worth it with the far
higher throughputs offered by newer PCI and NVME-type SSDs?
I've ordered a PCIe-to-M.2 adapter
On 17.04.2018 19:08, Matthew Lai wrote:
> Hello!
>
> I am getting ENOSPC on my root filesystem with plenty of unallocated
> space according to "fi usage". Any idea why that may be? This is a root
> partition for Debian Stable. Not doing anything unusual that I'm aware
> of. No snapshots.
>
>
On Wed, Apr 18, 2018 at 6:17 AM, David Sterba wrote:
> The last update to readdir introduced a temporary buffer to store the
> emitted readdir data, but as there are file names of variable length,
> there's a lot of unaligned access.
>
> This was observed on a sparc64 machine:
>
On Wed, Apr 18, 2018 at 09:41:54AM +0300, Nikolay Borisov wrote:
> When the delayed refs for a head are all run, eventually
> cleanup_ref_head is called which (in case of deletion) obtains a
> reference for the relevant btrfs_space_info struct by querying the bg
> for the range. This is
le_bitmap_clear() was implemented as a parallel to le_bitmap_set() but
is actually not being used. This patch removes it.
Signed-off-by: Howard McLauchlan
---
fs/btrfs/extent_io.c | 20
fs/btrfs/extent_io.h | 1 -
2 files changed, 21 deletions(-)
diff
Presently, convert_free_space_to_extents() does a linear scan of the
bitmap. We can speed this up with find_next_{bit,zero_bit}_le().
This patch replaces the linear scan with find_next_{bit,zero_bit}_le().
Testing shows a 20-33% decrease in execution time for
convert_free_space_to_extents().
On Wed, Apr 18, 2018 at 06:14:07PM +0300, Nikolay Borisov wrote:
>
>
> On 18.04.2018 18:10, Brendan Hide wrote:
> > Hi, all
> >
> > I'm looking for some advice re compression with NVME. Compression helps
> > performance with a minor CPU hit - but is it still worth it with the far
> > higher
On 04/18/2018 08:50 AM, David Sterba wrote:
> On Tue, Apr 17, 2018 at 04:19:04PM -0700, Howard McLauchlan wrote:
>> Presently, convert_free_space_to_extents() does a linear scan of the
>> bitmap. We can speed this up with find_next_{bit,zero_bit}_le().
>>
>> This patch replaces the linear scan
On Wed, Apr 18, 2018 at 10:38 AM, Austin S. Hemmelgarn wrote:
> For reference, the zstd compression in BTRFS uses level 3 by default (as
> does zlib compression IIRC), though I'm not sure about lzop (I think it
> uses the lowest compression setting).
>
The user space
On Wed, Apr 18, 2018 at 11:32:57AM +0900, Misono Tomohiro wrote:
> cangelog:
> v4 -> v5 ... Merge 2nd and 3rd patches and update commit log
>No code change in total
> v3 -> v4 ... Reorganize patches and update commit log
>No code change in total
> v2 -> v3 ...
Thank you, all
Though the info is useful, there's not a clear consensus on what I
should expect. For interest's sake, I'll post benchmarks from the device
itself when it arrives.
I'm expecting at least that I'll be blown away :)
On 04/18/2018 09:23 PM, Chris Murphy wrote:
On Wed, Apr 18,
On 18.04.2018 21:28, David Sterba wrote:
> On Wed, Apr 18, 2018 at 06:14:07PM +0300, Nikolay Borisov wrote:
>>
>>
>> On 18.04.2018 18:10, Brendan Hide wrote:
>>> Hi, all
>>>
>>> I'm looking for some advice re compression with NVME. Compression helps
>>> performance with a minor CPU hit - but is
On Wed, Apr 18, 2018 at 09:08:10AM +0800, Qu Wenruo wrote:
> >>dst_rsv = _info->delayed_block_rsv;
> >>
> >>num_bytes = btrfs_calc_trans_metadata_size(fs_info, 1);
> >> +
> >> + /*
> >> + * Here we migrate space rsv from transaction rsv, since have
> >> + * already reserved space
On 18.04.2018 10:27, Nikolay Borisov wrote:
> do_chunk_alloc implements logic to detect whether there is currently
> pending chunk allocation (by means of space_info->chunk_alloc being
> set) and if so it loops around to the 'again' label. Additionally,
> based on the state of the space_info
On Wed, Apr 18, 2018 at 1:02 AM, Jayashree Mohan
wrote:
> Hi,
>
> A gentle reminder on the crash consistency bug that we found on btrfs:
Why do you call it a consistency bug?
The filesystem does not stay in inconsistent state. The link count
stays 1 and the dentry used
Make the uuid_mutex and device_list_mutex comments inline with
the changes.
Signed-off-by: Anand Jain
---
Hi David,
Can you kindly add this patch to the set: 'Review uuid_mutex usage'
Thanks, Anand
fs/btrfs/volumes.c | 21 +
1 file changed, 9
From: Anand Jain
btrfs inspect dump-tree picks the disk with the largest generation
to read the root tree by scanning for the required devices by default.
But in 2 or more disks RAID1/5/6 you may need to know what's in the
disks individually, so this option --noscan
When the delayed refs for a head are all run, eventually
cleanup_ref_head is called which (in case of deletion) obtains a
reference for the relevant btrfs_space_info struct by querying the bg
for the range. This is problematic because when the last extent of a
bg is deleted a race window emerges
Adds comments about BTRFS_FS_EXCL_OP to existing comments
about the device locks.
Signed-off-by: Anand Jain
---
v1->v2: add more clarity to the comments added.
fs/btrfs/volumes.c | 32
1 file changed, 32 insertions(+)
diff --git
From: Ting-Chang Hou
fstatat will return -1 with errno EBADF if path_name is relative path.
This caused an error of the return value of overwrite_ok().
When restoring the subvolume to destination with relative path,
it will overwrite the existing file rather than skip it.
On Wed, Apr 18, 2018 at 12:39 AM, Howard McLauchlan wrote:
> Presently btrfs send/receive does not propagate inode attribute flags;
> all chattr operations are effectively discarded upon transmission.
>
> This patch adds kernel support for inode attribute flags. Userspace
>
On Tue, Apr 17, 2018 at 04:39:35PM -0700, Howard McLauchlan wrote:
> Presently btrfs send/receive does not propagate inode attribute flags;
> all chattr operations are effectively discarded upon transmission.
>
> This patch adds kernel support for inode attribute flags. Userspace
> support can be
do_chunk_alloc implements logic to detect whether there is currently
pending chunk allocation (by means of space_info->chunk_alloc being
set) and if so it loops around to the 'again' label. Additionally,
based on the state of the space_info (e.g. whether it's full or not)
and the return value of
On 18.04.2018 23:47, David Sterba wrote:
> On Wed, Apr 18, 2018 at 09:41:54AM +0300, Nikolay Borisov wrote:
>> When the delayed refs for a head are all run, eventually
>> cleanup_ref_head is called which (in case of deletion) obtains a
>> reference for the relevant btrfs_space_info struct by
On 04/18/2018 03:26 PM, David Sterba wrote:
> On Wed, Apr 18, 2018 at 02:30:59PM -0700, Howard McLauchlan wrote:
>> Presently, convert_free_space_to_extents() does a linear scan of the
>> bitmap. We can speed this up with find_next_{bit,zero_bit}_le().
>>
>> This patch replaces the linear scan
On Wed, Apr 18, 2018 at 05:00:29PM -0700, Eric Biggers wrote:
> Hi Chris and other btrfs folks,
>
> btrfs_mkdir() calls d_instantiate() before unlock_new_inode(), which is wrong
> because it exposes the inode to lookups before it's been fully initialized.
Huh? It *is* fully initialized by that
On Thu, Apr 19, 2018 at 01:06:13AM +0100, Al Viro wrote:
> On Wed, Apr 18, 2018 at 05:00:29PM -0700, Eric Biggers wrote:
> > Hi Chris and other btrfs folks,
> >
> > btrfs_mkdir() calls d_instantiate() before unlock_new_inode(), which is
> > wrong
> > because it exposes the inode to lookups
On Tue, Apr 17, 2018 at 09:47:19AM +0800, Qu Wenruo wrote:
> @@ -2680,7 +2681,7 @@ int open_ctree(struct super_block *sb,
>
> memcpy(fs_info->fsid, fs_info->super_copy->fsid, BTRFS_FSID_SIZE);
>
> - ret = btrfs_check_super_valid(fs_info);
> + ret =
On Wed, Apr 18, 2018 at 02:30:58PM -0700, Howard McLauchlan wrote:
> le_bitmap_clear() was implemented as a parallel to le_bitmap_set() but
> is actually not being used. This patch removes it.
>
> Signed-off-by: Howard McLauchlan
Reviewed-by: David Sterba
On 2018年04月19日 06:04, David Sterba wrote:
> On Tue, Apr 17, 2018 at 09:47:19AM +0800, Qu Wenruo wrote:
>> @@ -2680,7 +2681,7 @@ int open_ctree(struct super_block *sb,
>>
>> memcpy(fs_info->fsid, fs_info->super_copy->fsid, BTRFS_FSID_SIZE);
>>
>> -ret =
Hi Chris and other btrfs folks,
btrfs_mkdir() calls d_instantiate() before unlock_new_inode(), which is wrong
because it exposes the inode to lookups before it's been fully initialized.
Most filesystems get it right, but f2fs and btrfs don't. I sent a f2fs patch
On Wed, Apr 18, 2018 at 02:30:59PM -0700, Howard McLauchlan wrote:
> Presently, convert_free_space_to_extents() does a linear scan of the
> bitmap. We can speed this up with find_next_{bit,zero_bit}_le().
>
> This patch replaces the linear scan with find_next_{bit,zero_bit}_le().
> Testing shows
On Thu, Apr 19, 2018 at 01:15:59AM +0100, Al Viro wrote:
> On Thu, Apr 19, 2018 at 01:06:13AM +0100, Al Viro wrote:
> > On Wed, Apr 18, 2018 at 05:00:29PM -0700, Eric Biggers wrote:
> > > Hi Chris and other btrfs folks,
> > >
> > > btrfs_mkdir() calls d_instantiate() before unlock_new_inode(),
With commit b18253ec57c0 ("btrfs: optimize free space tree bitmap
conversion"), there are no more callers to le_test_bit(). This patch
removes le_test_bit().
Signed-off-by: Howard McLauchlan
---
fs/btrfs/extent_io.h | 5 -
1 file changed, 5 deletions(-)
diff --git
Presently, convert_free_space_to_extents() does a linear scan of the
bitmap. We can speed this up with find_next_{bit,zero_bit}_le().
This patch replaces the linear scan with find_next_{bit,zero_bit}_le().
Testing shows a 20-33% decrease in execution time for
convert_free_space_to_extents().
le_bitmap_set() is only used by free-space-tree, so move it there and
make it static. le_bitmap_clear() is not used, so remove it.
Signed-off-by: Howard McLauchlan
---
V1->V2: Also move le_bitmap_set()
based on 4.17-rc1
fs/btrfs/extent_io.c | 40
42 matches
Mail list logo