FAT: optimize FSINFO writeback
PATCH 1/1 FAT: optimize FSINFO writeback
Only write the FSINFO block back to disk when its contents change. This optimization can be important when the underlying physical media can wear out, i.e. Flash.
---
fs/fat/misc.c +19 -5
@@ -61,11 +61,25 @@ void fat_clusters_flush(struct super_blole32_to_cpu (fsinfo->signature2), sbi->fsinfo_sector); }else { -if (sbi->free_clusters !=-1 ) - fsinfo->free_clusters =cpu_to_le32 (sbi->free_clusters); -if (sbi->prev_free !=-1 ) - fsinfo->next_cluster =cpu_to_le32 (sbi->prev_free); -mark_buffer_dirty (bh); +char write_needed =0 ; +__le32 le_value; + +if (sbi->free_clusters !=-1 ) { + le_value =cpu_to_le32 (sbi->free_clusters); +if (fsinfo->free_clusters != le_value) { + fsinfo->free_clusters = le_value; + write_needed =1 ; + } + } +if (sbi->prev_free !=-1 ) { + le_value =cpu_to_le32 (sbi->prev_free); +if (fsinfo->next_cluster != le_value) { + fsinfo->next_cluster = le_value; + write_needed =1 ; + } + } +if (write_needed) +mark_buffer_dirty (bh); }brelse (bh); }
OGAWA Hirofumi
maintainerQuoted 3 lines · [email protected] writes:
Only write the FSINFO block back to disk when its contents change. This optimization can be important when the underlying physical media can wear out, i.e. Flash.
I have no objection to this though. Was this tested on recent version? Well, now, we are using sb->s_dirty for fsinfo, so I'm wondering why this happen frequently.
Thanks.
Quoted 37 lines
Signed-off-by Steven J. Magnani <[email protected]>---
diff -uprNa/fs/fat/misc.cb/fs/fat/misc.cfs/fat/misc.c +19
-5@@ -61,11 +61,25 @@ void fat_clusters_flush(struct super_blole32_to_cpu (fsinfo->signature2), sbi->fsinfo_sector); }else { -if (sbi->free_clusters !=-1 ) - fsinfo->free_clusters =cpu_to_le32 (sbi->free_clusters); -if (sbi->prev_free !=-1 ) - fsinfo->next_cluster =cpu_to_le32 (sbi->prev_free); -mark_buffer_dirty (bh); +char write_needed =0 ; +__le32 le_value; + +if (sbi->free_clusters !=-1 ) { + le_value =cpu_to_le32 (sbi->free_clusters); +if (fsinfo->free_clusters != le_value) { + fsinfo->free_clusters = le_value; + write_needed =1 ; + } + } +if (sbi->prev_free !=-1 ) { + le_value =cpu_to_le32 (sbi->prev_free); +if (fsinfo->next_cluster != le_value) { + fsinfo->next_cluster = le_value; + write_needed =1 ; + } + } +if (write_needed) +mark_buffer_dirty (bh); }brelse (bh); }
Signature · 1 line
OGAWA Hirofumi <[email protected]>
Quoted 7 lines · OGAWA Hirofumi
Quoted 3 lines · [email protected] writes:
Only write the FSINFO block back to disk when its contents change. This optimization can be important when the underlying physical media can wear out, i.e. Flash.
I have no objection to this though. Was this tested on recent version?
I tested it on 2.6.30.
Quoted 2 lines
Well, now, we are using
sb->s_dirtyfor fsinfo, so I'm wondering why this happen frequently.
My scenario was modifying a sector of an existing file and using fdatasync() to flush it. The FSINFO sector was being updated even though nothing about the FAT layout had changed.
Quoted 2 lines
Thanks.
Regards,
------------------------------------------------------------------------
Steven J. Magnani "I claim this network for MARS!
www.digidescorp.com Earthling, return my space modulator!"
#include <standard.disclaimer>
OGAWA Hirofumi
maintainerQuoted 6 lines · Steven J. Magnani
Quoted 2 lines
Well, now, we are using
sb->s_dirtyfor fsinfo, so I'm wondering why this happen frequently.My scenario was modifying a sector of an existing file and using
fdatasync()to flush it. The FSINFO sector was being updated even though nothing about the FAT layout had changed.
I see. Probably, I'm missing something, or handling of sb->s_dirt may be buggy, or something.
If it was fixed, is this patch still needed? I guess this patch would still be useful on some case though. If you can explain, it would be good.
Thanks.
Signature · 1 line
OGAWA Hirofumi <[email protected]>
Quoted 9 lines · OGAWA Hirofumi
Quoted 6 lines · Steven J. Magnani
Quoted 2 lines
Well, now, we are using
sb->s_dirtyfor fsinfo, so I'm wondering why this happen frequently.My scenario was modifying a sector of an existing file and using
fdatasync()to flush it. The FSINFO sector was being updated even though nothing about the FAT layout had changed.
Quoted 6 lines
I see. Probably, I'm missing something, or handling of
sb->s_dirtmay be buggy, or something.If it was fixed, is this patch still needed? I guess this patch would still be useful on some case though. If you can explain, it would be good.
It is I who am missing something. The patch originated against a 2.6.20 kernel, where it does indeed suppress unnecessary updates. I saw the same code in the 2.6.30 kernel and assumed the same issue was present, and tested that with the patch present there were no unnecessary updates. It appears that there are no unnecessary updates even _without_ the patch, so I withdraw it.
Quoted 2 lines
Thanks.
Thanks for looking at this so carefully.
Regards,
------------------------------------------------------------------------
Steven J. Magnani "I claim this network for MARS!
www.digidescorp.com Earthling, return my space modulator!"
#include <standard.disclaimer>