Skip to contentSkip to conversation

vfs fix

GIT PULL Al Viro 2 participants 4 messages Nov 22, 2011 VFS

lore
Share & tools
lore

Al Viro

author maintainer
#0 lore raw

Really embarrassing braino; kudos to jlayton for catching that one. Please, pull from
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git for-linus

Al Viro (1):
mount_subtree() pointless use-after-free

 fs/namespace.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)
#1 lore raw
Quoted 6 lines · Al Viro

Really embarrassing braino; kudos to jlayton for catching that one. Please, pull from
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git for-linus

Al Viro (1):
     mount_subtree() pointless use-after-free

Hmm. That's not at all what I get.

Or rather, that's just *one* of the commits I get. I also get these two:

microblaze: bury asm/namei.h
iio: fix a leak due to improper use of anon_inode_getfd()

which look like they may be valid, but for all I know you meant to leave them for the next merge window. So not pulled. Please check your tree,

Linus

author maintainer
#2 lore raw
Quoted 10 lines · Linus Torvalds
Quoted 6 lines · Al Viro

Really embarrassing braino; kudos to jlayton for catching that one. Please, pull from
git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs.git for-linus

Al Viro (1):
? ? ?mount_subtree() pointless use-after-free

Hmm. That's not at all what I get.

Quoted 8 lines

Or rather, that's just *one* of the commits I get. I also get these two:

microblaze: bury asm/namei.h
iio: fix a leak due to improper use of anon_inode_getfd()

which look like they may be valid, but for all I know you meant to leave them for the next merge window. So not pulled. Please check your tree,

Umm... OK, I see what happened - buggered script generating shortlog/diffstat plus me not noticing that it hadn't reported a couple of fixes I'd put there yesterday. All 3 are intended to be sent; iio is an outright bugfix, microblaze one could wait, in theory, but it's guaranteed to be safe (removal of header nothing includes, done on all architectures years ago, with microblaze tree apparently copying it from sparc and keeping all along) and the sooner such things are buried, the fewer places will copy that *again*...

Anyway, consider that a pull request for all 3 commits. Proper shortlog and diffstat follow.

Shortlog:
Al Viro (3):
microblaze: bury asm/namei.h
iio: fix a leak due to improper use of anon_inode_getfd() mount_subtree() pointless use-after-free

Diffstat:

 arch/microblaze/include/asm/namei.h     |   22 ----------------------
 drivers/staging/iio/industrialio-core.c |   10 +++++++++-
 fs/namespace.c                          |    6 ++++--
 3 files changed, 13 insertions(+), 25 deletions(-)

FWIW, the fixed form of script follows; if you see any potential problems in it, please tell:
echo Shortlog:
base=git merge-base origin "$1"
git log --pretty=short "$base..$1" | git shortlog
echo Diffstat:
git diff --stat -M "$base..$1"

#3 lore raw
Quoted 3 lines · Al Viro

FWIW, the fixed form of script follows; if you see any potential problems in it, please tell:

So not problems per se, just not very pretty.

Quoted 5 lines

echo Shortlog:
base=git merge-base origin "$1"
git log --pretty=short "$base..$1" | git shortlog
echo Diffstat:
git diff --stat -M "$base..$1"

I would suggest this instead:

  echo Shortlog:
  git shortlog "origin..$1"
  echo Diffstat:
  git diff --stat --summary -M "origin...$1"

and skip all the merge-base crap (git will do it for you: with a "git log" you don't need it, and with a "git diff" the three-dot version will do it for you). And just use "git shortlog" directly.

And the above also uses "--summary" for the "git diff" to show a summary of renames/deletes.

Linus