erofs: Escape overlayfs features #175
Conversation
|
We should probably also add an escaped xattr to one of the test images. |
cgwalters
added
enhancement
Yeah makes sense; moved to draft for that reason |
This escapes overlayfs xattrs and whiteouts according to the support in this patch series: https://lore.kernel.org/linux-unionfs/[email protected]/ Signed-off-by: Alexander Larsson <[email protected]>
alexlarsson
force-pushed
the
escape-overlayfs
branch
from
13b27b6
to
83123d6
Compare
|
Patch series 4 is more or less acked, so lets get this in |
|
|
||
| if (str_has_prefix(name, "trusted.overlay.")) { | ||
| cleanup_free char *renamed = | ||
| str_join("trusted.overlay.overlay.", |
There was a problem hiding this comment.
Is it really worth carrying str_join over just asprintf ("trusted.overlay.overlay.%s", name + strlen("trusted.overlay.")) ?
|
|
||
| lcfs_node_set_mode(node, | ||
| S_IFREG | (lcfs_node_get_mode(node) & ~S_IFMT)); | ||
| ret = lcfs_node_set_xattr(node, "trusted.overlay.overlay.whiteout", |
There was a problem hiding this comment.
Optional future cleanup: we could have a linux-overlayfs-internals.h header that includes #defines that link to the upstream source code.
There was a problem hiding this comment.
I might look at extracting this. Although we can't use the kernel headers here, the way the xattr defines are structured in the kernel is a bit weird.
|
Yeah, I plan to land all the outstanding stuff, and then do work on adding more tests in general. |
This escapes overlayfs xattrs and whiteouts according to the support in this patch series:
https://lore.kernel.org/linux-unionfs/[email protected]/
Note: We should probably wait with merging this until we have more feedback on the kernel list about it.
Fixes #172