This is a note to let you know that I've just added the patch titled UBIFS: fix a horrid bug to the 3.4-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: ubifs-fix-a-horrid-bug.patch and it can be found in the queue-3.4 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. >From 605c912bb843c024b1ed173dc427cd5c08e5d54d Mon Sep 17 00:00:00 2001 From: Artem Bityutskiy <artem.bityutskiy@xxxxxxxxxxxxxxx> Date: Fri, 28 Jun 2013 14:15:15 +0300 Subject: UBIFS: fix a horrid bug From: Artem Bityutskiy <artem.bityutskiy@xxxxxxxxxxxxxxx> commit 605c912bb843c024b1ed173dc427cd5c08e5d54d upstream. Al Viro pointed me to the fact that '->readdir()' and '->llseek()' have no mutual exclusion, which means the 'ubifs_dir_llseek()' can be run while we are in the middle of 'ubifs_readdir()'. This means that 'file->private_data' can be freed while 'ubifs_readdir()' uses it, and this is a very bad bug: not only 'ubifs_readdir()' can return garbage, but this may corrupt memory and lead to all kinds of problems like crashes an security holes. This patch fixes the problem by using the 'file->f_version' field, which '->llseek()' always unconditionally sets to zero. We set it to 1 in 'ubifs_readdir()' and whenever we detect that it became 0, we know there was a seek and it is time to clear the state saved in 'file->private_data'. I tested this patch by writing a user-space program which runds readdir and seek in parallell. I could easily crash the kernel without these patches, but could not crash it with these patches. Reported-by: Al Viro <viro@xxxxxxxxxxxxxxxxxx> Tested-by: Artem Bityutskiy <artem.bityutskiy@xxxxxxxxxxxxxxx> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@xxxxxxxxxxxxxxx> Signed-off-by: Al Viro <viro@xxxxxxxxxxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- fs/ubifs/dir.c | 30 +++++++++++++++++++++++++++--- 1 file changed, 27 insertions(+), 3 deletions(-) --- a/fs/ubifs/dir.c +++ b/fs/ubifs/dir.c @@ -373,6 +373,24 @@ static int ubifs_readdir(struct file *fi */ return 0; + if (file->f_version == 0) { + /* + * The file was seek'ed, which means that @file->private_data + * is now invalid. This may also be just the first + * 'ubifs_readdir()' invocation, in which case + * @file->private_data is NULL, and the below code is + * basically a no-op. + */ + kfree(file->private_data); + file->private_data = NULL; + } + + /* + * 'generic_file_llseek()' unconditionally sets @file->f_version to + * zero, and we use this for detecting whether the file was seek'ed. + */ + file->f_version = 1; + /* File positions 0 and 1 correspond to "." and ".." */ if (pos == 0) { ubifs_assert(!file->private_data); @@ -446,6 +464,14 @@ static int ubifs_readdir(struct file *fi file->f_pos = pos = key_hash_flash(c, &dent->key); file->private_data = dent; cond_resched(); + + if (file->f_version == 0) + /* + * The file was seek'ed meanwhile, lets return and start + * reading direntries from the new position on the next + * invocation. + */ + return 0; } out: @@ -456,15 +482,13 @@ out: kfree(file->private_data); file->private_data = NULL; + /* 2 is a special value indicating that there are no more direntries */ file->f_pos = 2; return 0; } -/* If a directory is seeked, we have to free saved readdir() state */ static loff_t ubifs_dir_llseek(struct file *file, loff_t offset, int origin) { - kfree(file->private_data); - file->private_data = NULL; return generic_file_llseek(file, offset, origin); } Patches currently in stable-queue which might be from artem.bityutskiy@xxxxxxxxxxxxxxx are queue-3.4/ubifs-prepare-to-fix-a-horrid-bug.patch queue-3.4/ubifs-fix-a-horrid-bug.patch -- To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html