Skip to content
  1. Jul 09, 2019
    • Richard Weinberger's avatar
      ubifs: Don't leak orphans on memory during commit · 8009ce95
      Richard Weinberger authored 
      

If an orphan has child orphans (xattrs), and due
to a commit the parent orpahn cannot get free()'ed immediately,
put also all child orphans on the erase list.
Otherwise UBIFS will free() them only upon unmount and we
waste memory.

Fixes: 988bec41 ("ubifs: orphan: Handle xattrs like files")
Signed-off-by: default avatarRichard Weinberger <richard@nod.at>
      8009ce95
    • Richard Weinberger's avatar
      ubifs: Check link count of inodes when killing orphans. · ee1438ce
      Richard Weinberger authored 
      

O_TMPFILE files can change their link count back to non-zero.
This corner case needs to get addressed in the orphans subsystem
too.

Fixes: 474b9370 ("ubifs: Implement O_TMPFILE")
Reported-by: default avatarLars Persson <lists@bofh.nu>
Signed-off-by: default avatarRichard Weinberger <richard@nod.at>
      ee1438ce
    • Michele Dionisio's avatar
      ubifs: Add support for zstd compression. · eeabb986
      Michele Dionisio authored 
      

zstd shows a good compression rate and is faster than lzo,
also on slow ARM cores.

Cc: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
Signed-off-by: default avatarMichele Dionisio <michele.dionisio@gmail.com>
[rw: rewrote commit message]
Signed-off-by: default avatarRichard Weinberger <richard@nod.at>
      eeabb986
    • Sascha Hauer's avatar
      ubifs: support offline signed images · 817aa094
      Sascha Hauer authored 
      

HMACs can only be generated on the system the UBIFS image is running on.
To support offline signed images we add a PKCS#7 signature to the UBIFS
image which can be created by mkfs.ubifs.

Both the master node and the superblock need to be authenticated, during
normal runtime both are protected with HMACs. For offline signature
support however only a single signature is desired. We add a signature
covering the superblock node directly behind it. To protect the master
node a hash of the master node is added to the superblock which is used
when the master node doesn't contain a HMAC.

Transition to a read/write filesystem is also supported. During
transition first the master node is rewritten with a HMAC (implicitly,
it is written anyway as the FS is marked dirty). Afterwards the
superblock is rewritten with a HMAC. Once after the image has been
mounted read/write it is HMAC only, the signature is no longer required
or even present on the filesystem.

In an offline signed image the master node is authenticated by the
superblock. In a transition to r/w we have to make sure that the master
node is rewritten before the superblock node. In this case the master
node gets a HMAC and its authenticity no longer depends on the
superblock node. There are some cases in which the current code first
writes the superblock node though, so with this patch writing of the
superblock node is delayed until the master node is written.

Signed-off-by: default avatarSascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: default avatarRichard Weinberger <richard@nod.at>
      817aa094
    • Liu Song's avatar
      ubifs: remove unnecessary check in ubifs_log_start_commit · 8ba0a2ab
      Liu Song authored 
      

In ubifs_log_start_commit, the value of c->lhead_offs is zero or set
to zero by code bellow.

	/* Switch to the next log LEB */
	if (c->lhead_offs) {
		c->lhead_lnum = ubifs_next_log_lnum(c, c->lhead_lnum);
		ubifs_assert(c->lhead_lnum != c->ltail_lnum);
		c->lhead_offs = 0;
	}

The value of 'len' can not exceed 'max_len' which assigned value by
code bellow.

	max_len = UBIFS_CS_NODE_SZ + c->jhead_cnt * UBIFS_REF_NODE_SZ;

The value of c->lhead_offs changed by code bellow and cannot exceed
'max_len'.

	c->lhead_offs += len;
	if (c->lhead_offs == c->leb_size) {
		c->lhead_lnum = ubifs_next_log_lnum(c, c->lhead_lnum);
		c->lhead_offs = 0;
	}

Usually, the size of PEB is between 64KB and 256KB. So the value of
c->lhead_offs is far less than c->leb_size. The check
'if (c->lhead_offs == c->leb_size)' could never to be true.

Signed-off-by: default avatarLiu Song <liu.song11@zte.com.cn>
Reviewed-by: default avatarJiang Biao <jiang.biao2@zte.com.cn>
Signed-off-by: default avatarRichard Weinberger <richard@nod.at>
      8ba0a2ab
    • Liu Song's avatar
      ubifs: Fix typo of output in get_cs_sqnum · 7d8c811b
      Liu Song authored 
      

"Not a CS node" makes more sense than "Node a CS node".

Signed-off-by: default avatarLiu Song <liu.song11@zte.com.cn>
Reviewed-by: default avatarJiang Biao <jiang.biao2@zte.com.cn>
Signed-off-by: default avatarRichard Weinberger <richard@nod.at>
      7d8c811b
    • Liu Song's avatar
      ubifs: Simplify redundant code · d5cf9473
      Liu Song authored 
      

cbuf's size can be simply assigned.

Signed-off-by: default avatarLiu Song <liu.song11@zte.com.cn>
Reviewed-by: default avatarJiang Biao <jiang.biao2@zte.com.cn>
Signed-off-by: default avatarRichard Weinberger <richard@nod.at>
      d5cf9473
    • Richard Weinberger's avatar
      ubifs: Correctly use tnc_next() in search_dh_cookie() · bacfa94b
      Richard Weinberger authored 
      

Commit c877154d fixed an uninitialized variable and optimized
the function to not call tnc_next() in the first iteration of the
loop. While this seemed perfectly legit and wise, it turned out to
be illegal.
If the lookup function does not find an exact match it will rewind
the cursor by 1.
The rewinded cursor will not match the name hash we are looking for
and this results in a spurious -ENOENT.
So we need to move to the next entry in case of an non-exact match,
but not if the match was exact.

While we are here, update the documentation to avoid further confusion.

Cc: Hyunchul Lee <hyc.lee@gmail.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Fixes: c877154d ("ubifs: Fix uninitialized variable in search_dh_cookie()")
Fixes: 781f675e ("ubifs: Fix unlink code wrt. double hash lookups")
Signed-off-by: default avatarRichard Weinberger <richard@nod.at>
      bacfa94b
  3. Jun 30, 2019
  5. Jun 29, 2019