| Message ID | 20180904124201.3287-1-lufq.fnst@cn.fujitsu.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | btrfs-progs: calibrate extent_end when found a gap | expand |
On Tue, Sep 04, 2018 at 08:42:01PM +0800, Lu Fengqi wrote: > The extent_end will be used to check whether there is gap between this > extent and next extent. If it is not calibrated, check_file_extent will Do you mean 'synchronized' or 'matching'. > mistake that there are gaps between the remaining extents. If this is a bugfix, do you have a testcase? Thanks.
On Tue, Sep 11, 2018 at 04:41:21PM +0200, David Sterba wrote: >On Tue, Sep 04, 2018 at 08:42:01PM +0800, Lu Fengqi wrote: >> The extent_end will be used to check whether there is gap between this >> extent and next extent. If it is not calibrated, check_file_extent will > >Do you mean 'synchronized' or 'matching'. I apologize for this incomprehensible commit message, and I have updated the commit message. [PATCH] btrfs-progs: lowmem: fix false alert about the existence of gaps in the check_file_extent > >> mistake that there are gaps between the remaining extents. > >If this is a bugfix, do you have a testcase? Thanks. > The testcase requires some check repair's fixes (including originl and lowmem) that my colleagues are working on. After they get it, I will send the testcase. The attached is the image which can trigger the false alert. Without the patch mentioned before, lowmem check will false alert that expect the hole extent [257 EXTENT_DATA 8192]. ERROR: root 5 EXTENT_DATA[257 12288] gap exists, expected: EXTENT_DATA[257 8192]
On Thu, Sep 13, 2018 at 04:30:28PM +0800, Lu Fengqi wrote: >On Tue, Sep 11, 2018 at 04:41:21PM +0200, David Sterba wrote: >>On Tue, Sep 04, 2018 at 08:42:01PM +0800, Lu Fengqi wrote: >>> The extent_end will be used to check whether there is gap between this >>> extent and next extent. If it is not calibrated, check_file_extent will >> >>Do you mean 'synchronized' or 'matching'. > >I apologize for this incomprehensible commit message, and I have updated >the commit message. > >[PATCH] btrfs-progs: lowmem: fix false alert about the existence of gaps in the check_file_extent > >> >>> mistake that there are gaps between the remaining extents. >> >>If this is a bugfix, do you have a testcase? Thanks. >> > >The testcase requires some check repair's fixes (including originl and lowmem) >that my colleagues are working on. After they get it, I will send the >testcase. > >The attached is the image which can trigger the false alert. Sorry, I miss the attached.
diff --git a/check/mode-lowmem.c b/check/mode-lowmem.c index 1bce44f5658a..0f14a4968e84 100644 --- a/check/mode-lowmem.c +++ b/check/mode-lowmem.c @@ -1972,6 +1972,7 @@ static int check_file_extent(struct btrfs_root *root, struct btrfs_path *path, root->objectid, fkey.objectid, fkey.offset, fkey.objectid, *end); } + *end = fkey.offset; } *end += extent_num_bytes;
The extent_end will be used to check whether there is gap between this extent and next extent. If it is not calibrated, check_file_extent will mistake that there are gaps between the remaining extents. Signed-off-by: Lu Fengqi <lufq.fnst@cn.fujitsu.com> --- check/mode-lowmem.c | 1 + 1 file changed, 1 insertion(+)