Created attachment 1103385 [details]
proposed patch

Description of problem:

With raid456 enabled and raid456.devices_handle_discard_safely=Y, discard_zeroes_data is set to 0 in /sys for md device.

Version-Release number of selected component (if applicable):

kernel-2.6.32-573.8.1.el6

How reproducible:

[root@host ~]# grep raid /boot/grub/grub.conf 
	kernel /vmlinuz-2.6.32-573.8.1.el6.x86_64 ro root=/dev/mapper/vg_host-LogVol01root rd_NO_LUKS LANG=en_US.UTF-8 rd_LVM_LV=vg_host/LogVol01root rd_NO_MD SYSFONT=latarcyrheb-sun16 crashkernel=auto rd_LVM_LV=vg_host/LogVol00swap KEYBOARDTYPE=pc KEYTABLE=us rd_NO_DM printk.time=1 raid456.devices_handle_discard_safely=Y

[root@host ~]# lsmod | grep raid456
raid456                84075  0 
async_raid6_recov       6491  1 raid456
async_pq                4638  2 raid456,async_raid6_recov
async_xor               3522  3 raid456,async_raid6_recov,async_pq
async_memcpy            2172  2 raid456,async_raid6_recov
async_tx                2995  5 raid456,async_raid6_recov,async_pq,async_xor,async_memcpy

[root@host ~]# mdadm --create /dev/md1 --level=raid6 --raid-devices=4 /dev/sdb /dev/sdc /dev/sdd /dev/sde
mdadm: Defaulting to version 1.2 metadata
mdadm: array /dev/md1 started.

[root@host ~]# cat /proc/mdstat 
Personalities : [raid6] [raid5] [raid4] 
md1 : active raid6 sde[3] sdd[2] sdc[1] sdb[0]
      234309632 blocks super 1.2 level 6, 512k chunk, algorithm 2 [4/4] [UUUU]
      bitmap: 0/1 pages [0KB], 65536KB chunk

unused devices: <none>

[root@host ~]# for d in {discard_granularity,discard_max_bytes,discard_zeroes_data}
> do for i in $(ls -l /sys/block | egrep 'sdb|sdc|sdd|sde|md' | awk '{ print $9}')
> do echo $i, $d: `cat /sys/block/$i/queue/$d`
> done
> done
md1, discard_granularity: 33553920
sdb, discard_granularity: 33553920
sdc, discard_granularity: 33553920
sdd, discard_granularity: 33553920
sde, discard_granularity: 33553920
md1, discard_max_bytes: 2199023255040
sdb, discard_max_bytes: 2199023255040
sdc, discard_max_bytes: 2199023255040
sdd, discard_max_bytes: 2199023255040
sde, discard_max_bytes: 2199023255040
md1, discard_zeroes_data: 0           <====== believe this should be 1
sdb, discard_zeroes_data: 1
sdc, discard_zeroes_data: 1
sdd, discard_zeroes_data: 1
sde, discard_zeroes_data: 1

[root@host ~]# pvcreate /dev/md1
  Physical volume "/dev/md1" successfully created

[root@host ~]# vgcreate vg2 /dev/md1
  Volume group "vg2" successfully created

[root@host ~]# lvcreate -n lv_data1 -l 100%FREE vg2
  Logical volume "lv_data1" created.

[root@host ~]# mkfs.ext4 /dev/mapper/vg2-lv_data1 
mke2fs 1.41.12 (17-May-2010)
Discarding device blocks: done                            
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
Stride=128 blocks, Stripe width=256 blocks
14647296 inodes, 58576896 blocks
2928844 blocks (5.00%) reserved for the super user
First data block=0
Maximum filesystem blocks=4294967296
1788 block groups
32768 blocks per group, 32768 fragments per group
8192 inodes per group
Superblock backups stored on blocks: 
	32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 
	4096000, 7962624, 11239424, 20480000, 23887872

Writing inode tables: done                            
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 33 mounts or
180 days, whichever comes first.  Use tune2fs -c or -i to override.

[root@host ~]# mkdir /test
[root@host ~]# mount /dev/vg2/lv_data1 /test
[root@host ~]# touch /test/testfile
[root@host ~]# rm /test/testfile
rm: remove regular empty file `/test/testfile'? y
[root@host ~]# fstrim -v /test/
/test/: 235969359872 bytes were trimmed

Actual results:

discard_zeroes_data set to 0

Expected results:

discard_zeroes_data set to 1 at discard_supported and raid456.devices_handle_discard_safely set to Y

* proposed patch created by maintenance engineering attached