Skip to contentSkip to conversation

fs: jffs2: possible sleep-in-atomic-context bugs in jffs2_iget

BUG Jia-Ju Bai 1 participant 1 message Aug 13, 2018

lore
Share & tools
lore

Jia-Ju Bai

author
#0 lore raw

The kernel may sleep with holding a spinlock.

The function call paths (from bottom to top) in Linux-4.16 are:

[FUNC] schedule
fs/inode.c, 1916: schedule in __wait_on_freeing_inode fs/inode.c, 826: __wait_on_freeing_inode in find_inode_fast fs/inode.c, 1107: find_inode_fast in iget_locked fs/jffs2/fs.c, 263: iget_locked in jffs2_iget fs/jffs2/fs.c, 665: jffs2_iget in jffs2_gc_fetch_inode fs/jffs2/wbuf.c, 505: jffs2_gc_fetch_inode in jffs2_wbuf_recover fs/jffs2/wbuf.c, 462: spin_lock in jffs2_wbuf_recover

[FUNC] mutex_lock_nested
fs/jffs2/fs.c, 273: mutex_lock_nested in jffs2_iget fs/jffs2/fs.c, 665: jffs2_iget in jffs2_gc_fetch_inode fs/jffs2/wbuf.c, 505: jffs2_gc_fetch_inode in jffs2_wbuf_recover fs/jffs2/wbuf.c, 462: spin_lock in jffs2_wbuf_recover

I do not find a good way to fix, so I only report.
Maybe the spinlock should be released before calling jffs2_iget(), and then be acquired again. This is found by my static analysis tool (DSAC).

Thanks,
Jia-Ju Bai