fs/9p: avoid accessing utsname after namespace has been torn down
During trinity fuzzing in a kvmtool guest, I stumbled across the
following:
Unable to handle kernel NULL pointer dereference at virtual address 00000004
PC is at v9fs_file_do_lock+0xc8/0x1a0
LR is at v9fs_file_do_lock+0x48/0x1a0
[<c01e2ed0>] (v9fs_file_do_lock+0xc8/0x1a0) from [<c0119154>] (locks_remove_flock+0x8c/0x124)
[<c0119154>] (locks_remove_flock+0x8c/0x124) from [<c00d9bf0>] (__fput+0x58/0x1e4)
[<c00d9bf0>] (__fput+0x58/0x1e4) from [<c0044340>] (task_work_run+0xac/0xe8)
[<c0044340>] (task_work_run+0xac/0xe8) from [<c002e36c>] (do_exit+0x6bc/0x8d8)
[<c002e36c>] (do_exit+0x6bc/0x8d8) from [<c002e674>] (do_group_exit+0x3c/0xb0)
[<c002e674>] (do_group_exit+0x3c/0xb0) from [<c002e6f8>] (__wake_up_parent+0x0/0x18)
I believe this is due to an attempt to access utsname()->nodename, after
exit_task_namespaces() has been called, leaving current->nsproxy->uts_ns
as NULL and causing the above dereference.
A similar issue was fixed for lockd in 9a1b6bf818e7 ("LOCKD: Don't call
utsname()->nodename from nlmclnt_setlockargs"), so this patch attempts
something similar for 9pfs.
Cc: Eric Van Hensbergen <ericvh@gmail.com>
Cc: Ron Minnich <rminnich@sandia.gov>
Cc: Latchesar Ionkov <lucho@ionkov.net>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c
index d384a8b..aa5ecf4 100644
--- a/fs/9p/vfs_file.c
+++ b/fs/9p/vfs_file.c
@@ -183,7 +183,7 @@
else
flock.length = fl->fl_end - fl->fl_start + 1;
flock.proc_id = fl->fl_pid;
- flock.client_id = utsname()->nodename;
+ flock.client_id = fid->clnt->name;
if (IS_SETLKW(cmd))
flock.flags = P9_LOCK_FLAGS_BLOCK;
@@ -260,7 +260,7 @@
else
glock.length = fl->fl_end - fl->fl_start + 1;
glock.proc_id = fl->fl_pid;
- glock.client_id = utsname()->nodename;
+ glock.client_id = fid->clnt->name;
res = p9_client_getlock_dotl(fid, &glock);
if (res < 0)