#322 closed defect (fixed)
`cp` seems to produce a corrupted copy on a non-root fat
| Reported by: | Jakub Jermář | Owned by: | Jakub Jermář |
|---|---|---|---|
| Priority: | major | Milestone: | 0.4.3 |
| Component: | helenos/app/bdsh | Version: | |
| Keywords: | Cc: | ||
| Blocker for: | Depends on: | ||
| See also: |
Description ¶
Mounting a non-root FAT file system over TMPFS root and doing:
/tmp # cp /readme /data/a /tmp # ls /data/a /tmp # cat /data/a
will result in the corrupted contents of /data/a. The reported file size will be correct, but the last few tens of bytes will look like duplicated from the beginning of the file.
Change History (4)
comment:1 by , 14 years ago
comment:2 by , 14 years ago
| Component: | fs/fat → app/bdsh |
|---|
This is actually a problem of the bdsh implementation of cp. If the write() needs to be done multiple times during the copy, it is essential to move the buffer address according to the bytes written. Already fixed in lp:~jakub/helenos/fs.
comment:3 by , 14 years ago
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
Fix merged in changeset:mainline,890.
comment:4 by , 14 years ago
| Milestone: | 0.5.0 → 0.4.3 |
|---|
Note:
See TracTickets
for help on using tickets.
Not reproducible if the root FS is FAT too.