On Mon, 22 Nov 2010, Greg Freemyer wrote: > On Mon, Nov 22, 2010 at 6:29 AM, Lukas Czerner <lczerner@xxxxxxxxxx> wrote: > > It takes fstrim_range structure as an argument. fstrim_range is definec in > > the include/linux/fs.h. > > > > After the FITRIM is done, the number of actually discarded Bytes is stored > > in fstrim_range.len to give the user better insight on how much storage > > space has been really released for wear-leveling. > > > > Signed-off-by: Lukas Czerner <lczerner@xxxxxxxxxx> > > That's a misleading description in my opinion, from my understanding > this is more accurate: > > === > After the FITRIM is done, the number of bytes passed from the > filesystem down the block stack to the device for potential discard is > stored in fstrim_range.len. This number is a maximum discard amount > from the storage device's perspective, because FITRIM called repeated > will keep sending the same sectors for discard repeatedly. > fstrim_range.len will report the same potential discard bytes each > time, but only sectors which had been written to between the discards > would actually be discarded by the storage device. Further, the > kernel block layer reserves the right to adjust the discard ranges to > fit raid stripe geometry, non-trim capable devices in a LVM setup, > etc. These reductions would not be reflected in fstrim_range.len. > > As 2.6.37, the kernel block layer does not fully support discard and > as such will simply ignore all discard requests sent to volumes > created by device mapper or mdraid. This is done in a silent way, so > these failures to discard are also not reflected in fstrim_range.len. I think this is not entirely true, because we have discard support for dm linear and stripe: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=5ae89a8720c28caf35c4e53711d77df2856c404e http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=7b76ec11fec40203836b488496d2df082d5b2022 (Adding Mike Snitzer into cc) > > Thus fstrim_range.len can give the user better insight on how much > storage space has potentially been released for wear-leveling, but it > needs to be one of only one criteria the userspace tools take into > account when trying to optimize calls to FITRIM. > === > > Obviously, I'd like to also see that also in API documentation for > FITRIM. (And correct me if I'm wrong about device mapper / mdraid. > I'd love to be wrong about that statement..) > > Greg > Greg, thank you for this, aside the dm thing, this is definitely a lot better explanation and it would be nice to include this into commit message (it is too late for ext4:). Jan, do you want me to repost this with new commit message or you can add it yourself ? Thanks! -Lukas -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
