#19273 closed bug (fixed)
vfs: KDL when copying file in mounted BFS disk image
Reported by: | X512 | Owned by: | nobody |
---|---|---|---|
Priority: | normal | Milestone: | R1/beta6 |
Component: | System/Kernel | Version: | R1/Development |
Keywords: | Cc: | ||
Blocked By: | Blocking: | ||
Platform: | All |
Description (last modified by )
This is hrev58368.
It happens when copying ~300 KB file from NVMe BFS disk into 4 GB BFS disk image mounted with diskimage
tool. File is copied using Tracker GUI.
BFS disk image was formatted and filled by BeOS R5. Haiku checkfs
show no errors on it.
Attachments (1)
Change History (10)
by , 4 days ago
Attachment: | IMG_20241201_121811-2.jpg added |
---|
comment:1 by , 4 days ago
Milestone: | Unscheduled → R1/beta6 |
---|---|
Summary: | [regression] vfs: KDL when copying file in mounted BFS disk image → vfs: KDL when copying file in mounted BFS disk image |
comment:2 by , 4 days ago
<x512> In my case it always happens when copying ProcessController3.pkg into BeOS 4 GB disk image using Tracker.
comment:3 by , 4 days ago
Description: | modified (diff) |
---|
comment:4 by , 4 days ago
This probably only could happen if we wind up "iterating" more than once (i.e. the underlying file has more than one block_run), but the source iovec is larger than the first iteration's I/O.
comment:5 by , 4 days ago
Description: | modified (diff) |
---|
comment:6 by , 4 days ago
Description: | modified (diff) |
---|
comment:7 by , 37 hours ago
I found some disk images that reliably reproduce this, will investigate.
Note:
See TracTickets
for help on using tickets.
Probably not a regression but an uncovered bug, since the assertion was added recently. Likely the only problem here is a memory leak, but that's still significant of course.