grub-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH 3/5] zfs com.delphix:hole_birth feature support


From: Andrei Borzenkov
Subject: Re: [PATCH 3/5] zfs com.delphix:hole_birth feature support
Date: Sun, 3 May 2015 18:15:05 +0300

В Sun, 19 Apr 2015 17:57:04 +0300
Toomas Soome <address@hidden> пишет:

> 
> > On 19.04.2015, at 17:51, Andrei Borzenkov <address@hidden> wrote:
> > 
> > В Sun, 19 Apr 2015 17:40:16 +0300
> > Toomas Soome <address@hidden> пишет:
> > 
> >> 
> >> the features in openzfs have different effects, some affect only writes 
> >> and are therefore read only compatible - such feature does not need any 
> >> changes for reads, this feature is not read only compatible and therefore, 
> >> once this feature is enabled, reader code must be changed accordingly.
> >> 
> >> so, what they did with hole_birth was they started to insert block birth 
> >> timestamps for blocks being released to make it possible to track such 
> >> blocks while doing zfs send (when block has birth time, you know its place 
> >> on timeline of snapshots). as old reader code was relying on blk_birth == 
> >> 0 to detect an hole, after hole_birth is enabled, hole blk_birth is not 0 
> >> any more, so the solution is to check if DVA pointers are zero instead. 
> >> and thats exactly what its about.  
> > 
> > And if hole_birth is *not* enabled on a filesystem we are reading?
> > Should not old code be used in this case?
> 
> 
> not really, as for holes, DVA pointers are  zero anyhow. I guess the original 
> implementation used blk_birth to save extra compare - so the new read code is 
> compatible with old data.

Committed with explanatory text for non-insiders.

> 
> 
> > 
> >> 
> >> reference:
> >> https://github.com/illumos/illumos-gate/commit/43466aae47bfcd2ad9bf501faec8e75c08095e4f
> >>  
> >> 
> >> 
> >> 
> >>> On 19.04.2015, at 17:11, Andrei Borzenkov <address@hidden> wrote:
> >>> 
> >>> В Thu, 16 Apr 2015 08:22:08 +0300
> >>> Toomas Soome <address@hidden> пишет:
> >>> 
> >>> This really needs better explanation. Otherwise this looks like either
> >>> old code was broken to start with and it is a bug fix or new code needs
> >>> some conditionals on new feature.
> >>> 
> >>>> 
> >>>> ---
> >>>> grub-core/fs/zfs/zfs.c |    6 ++++--
> >>>> include/grub/zfs/spa.h |    4 +++-
> >>>> 2 files changed, 7 insertions(+), 3 deletions(-)
> >>>> 
> >>>> diff --git a/grub-core/fs/zfs/zfs.c b/grub-core/fs/zfs/zfs.c
> >>>> index 2689986..a731c3d 100644
> >>>> --- a/grub-core/fs/zfs/zfs.c
> >>>> +++ b/grub-core/fs/zfs/zfs.c
> >>>> @@ -280,7 +280,9 @@ grub_crypto_cipher_handle_t (*grub_zfs_load_key) 
> >>>> (const struct grub_zfs_key *key
> >>>> */
> >>>> #define MAX_SUPPORTED_FEATURE_STRLEN 50
> >>>> static const char *spa_feature_names[] = {
> >>>> -  "org.illumos:lz4_compress",NULL
> >>>> +  "org.illumos:lz4_compress",
> >>>> +  "com.delphix:hole_birth",
> >>>> +  NULL
> >>>> };
> >>>> 
> >>>> static int
> >>>> @@ -1751,7 +1753,7 @@ zio_read_gang (blkptr_t * bp, grub_zfs_endian_t 
> >>>> endian, dva_t * dva, void *buf,
> >>>> 
> >>>>  for (i = 0; i < SPA_GBH_NBLKPTRS; i++)
> >>>>    {
> >>>> -      if (zio_gb->zg_blkptr[i].blk_birth == 0)
> >>>> +      if (BP_IS_HOLE(&zio_gb->zg_blkptr[i]))
> >>>>  continue;
> >>>> 
> >>>>      err = zio_read_data (&zio_gb->zg_blkptr[i], endian, buf, data);
> >>>> diff --git a/include/grub/zfs/spa.h b/include/grub/zfs/spa.h
> >>>> index 7edb8ab..df43b6b 100644
> >>>> --- a/include/grub/zfs/spa.h
> >>>> +++ b/include/grub/zfs/spa.h
> >>>> @@ -279,7 +279,9 @@ typedef struct blkptr {
> >>>> 
> >>>> #define  BP_IDENTITY(bp)         (&(bp)->blk_dva[0])
> >>>> #define  BP_IS_GANG(bp)          DVA_GET_GANG(BP_IDENTITY(bp))
> >>>> -#define BP_IS_HOLE(bp)          ((bp)->blk_birth == 0)
> >>>> +#define DVA_IS_EMPTY(dva)       ((dva)->dva_word[0] == 0ULL && \
> >>>> +                                (dva)->dva_word[1] == 0ULL)
> >>>> +#define BP_IS_HOLE(bp)          DVA_IS_EMPTY(BP_IDENTITY(bp))
> >>>> 
> >>>> /* BP_IS_RAIDZ(bp) assumes no block compression */
> >>>> #define  BP_IS_RAIDZ(bp)         (DVA_GET_ASIZE(&(bp)->blk_dva[0]) > \
> >>> 
> >>> 
> >>> _______________________________________________
> >>> Grub-devel mailing list
> >>> address@hidden
> >>> https://lists.gnu.org/mailman/listinfo/grub-devel
> >> 
> >> 
> >> _______________________________________________
> >> Grub-devel mailing list
> >> address@hidden
> >> https://lists.gnu.org/mailman/listinfo/grub-devel
> 




reply via email to

[Prev in Thread] Current Thread [Next in Thread]