On Fri, 2024-10-11 at 09:04 -0400, Jeff Layton wrote:
On Thu, 2024-10-10 at 18:18 -0400, Olga Kornievskaia wrote:
There is a race between laundromat handling of revoked delegations and a client sending free_stateid operation. Laundromat thread finds that delegation has expired and needs to be revoked so it marks the delegation stid revoked and it puts it on a reaper list but then it unlock the state lock and the actual delegation revocation happens without the lock. Once the stid is marked revoked a racing free_stateid processing thread does the following (1) it calls list_del_init() which removes it from the reaper list and (2) frees the delegation stid structure. The laundromat thread ends up not calling the revoke_delegation() function for this particular delegation but that means it will no release the lock lease that exists on the file.
Now, a new open for this file comes in and ends up finding that lease list isn't empty and calls nfsd_breaker_owns_lease() which ends up trying to derefence a freed delegation stateid. Leading to the followint use-after-free KASAN warning:
kernel:
kernel: BUG: KASAN: slab-use-after-free in nfsd_breaker_owns_lease+0x140/0x160 [nfsd] kernel: Read of size 8 at addr ffff0000e73cd0c8 by task nfsd/6205 kernel: kernel: CPU: 2 UID: 0 PID: 6205 Comm: nfsd Kdump: loaded Not tainted 6.11.0-rc7+ #9 kernel: Hardware name: Apple Inc. Apple Virtualization Generic Platform, BIOS 2069.0.0.0.0 08/03/2024 kernel: Call trace: kernel: dump_backtrace+0x98/0x120 kernel: show_stack+0x1c/0x30 kernel: dump_stack_lvl+0x80/0xe8 kernel: print_address_description.constprop.0+0x84/0x390 kernel: print_report+0xa4/0x268 kernel: kasan_report+0xb4/0xf8 kernel: __asan_report_load8_noabort+0x1c/0x28 kernel: nfsd_breaker_owns_lease+0x140/0x160 [nfsd] kernel: leases_conflict+0x68/0x370 kernel: __break_lease+0x204/0xc38 kernel: nfsd_open_break_lease+0x8c/0xf0 [nfsd] kernel: nfsd_file_do_acquire+0xb3c/0x11d0 [nfsd] kernel: nfsd_file_acquire_opened+0x84/0x110 [nfsd] kernel: nfs4_get_vfs_file+0x634/0x958 [nfsd] kernel: nfsd4_process_open2+0xa40/0x1a40 [nfsd] kernel: nfsd4_open+0xa08/0xe80 [nfsd] kernel: nfsd4_proc_compound+0xb8c/0x2130 [nfsd] kernel: nfsd_dispatch+0x22c/0x718 [nfsd] kernel: svc_process_common+0x8e8/0x1960 [sunrpc] kernel: svc_process+0x3d4/0x7e0 [sunrpc] kernel: svc_handle_xprt+0x828/0xe10 [sunrpc] kernel: svc_recv+0x2cc/0x6a8 [sunrpc] kernel: nfsd+0x270/0x400 [nfsd] kernel: kthread+0x288/0x310 kernel: ret_from_fork+0x10/0x20
Proposing to have laundromat thread hold the state_lock over both marking thru revoking the delegation as well as making free_stateid acquire state_lock before accessing the list. Making sure that revoke_delegation() (ie kernel_setlease(unlock)) is called for every delegation that was revoked and added to the reaper list.
Nice detective work!
CC: stable@vger.kernel.org Signed-off-by: Olga Kornievskaia okorniev@redhat.com
--- I can't figure out the Fixes: tag. Laundromat's behaviour has been like that forever. But the free_stateid bits wont apply before the 1e3577a4521e ("SUNRPC: discard sv_refcnt, and svc_get/svc_put"). But we used that fixes tag already with a previous fix for a different problem.
fs/nfsd/nfs4state.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 9c2b1d251ab3..c97907d7fb38 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -6605,13 +6605,13 @@ nfs4_laundromat(struct nfsd_net *nn) unhash_delegation_locked(dp, SC_STATUS_REVOKED); list_add(&dp->dl_recall_lru, &reaplist); }
- spin_unlock(&state_lock);
while (!list_empty(&reaplist)) { dp = list_first_entry(&reaplist, struct nfs4_delegation, dl_recall_lru); list_del_init(&dp->dl_recall_lru); revoke_delegation(dp); }
- spin_unlock(&state_lock);
spin_lock(&nn->client_lock); while (!list_empty(&nn->close_lru)) { @@ -7213,7 +7213,9 @@ nfsd4_free_stateid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if (s->sc_status & SC_STATUS_REVOKED) { spin_unlock(&s->sc_lock); dp = delegstateid(s);
spin_lock(&state_lock);list_del_init(&dp->dl_recall_lru);
spin_unlock(&state_lock);spin_unlock(&cl->cl_lock);
nfs4_set_delegation() takes these locks in the opposite order, so as it stands, this patch introduces a potential ABBA deadlock. I suggest moving the state_lock so that it is taken before and released after the cl->cl_lock.
nfs4_put_stid(s); ret = nfs_ok;
I'm not thrilled with this patch, but it does seem like it would fix the problem. Long term, I think we need to get rid of the state_lock, but I don't have a real plan for that just yet as it's not 100% clear what it still protects.
As far as a Fixes tag, this bug is likely very old. I'd say it probably got introduced here:
2d4a532d385f nfsd: ensure that clp->cl_revoked list is protected by clp->cl_lock
In any case, let's go with this patch until we can come up with a better plan for delegation handling.
Reviewed-by: Jeff Layton jlayton@kernel.org