Date: Fri, 3 Oct 2025 05:55:16 GMT From: Ryan Libby <rlibby@FreeBSD.org> To: src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-main@FreeBSD.org Subject: git: e7e958d83d85 - main - ddb show allchains: avoid traps and show pointers for lockmgr and sx Message-ID: <202510030555.5935tGss018026@gitrepo.freebsd.org>
next in thread | raw e-mail | index | archive | help
The branch main has been updated by rlibby: URL: https://cgit.FreeBSD.org/src/commit/?id=e7e958d83d8503c3b8714f68418d4509b8288ed9 commit e7e958d83d8503c3b8714f68418d4509b8288ed9 Author: Ryan Libby <rlibby@FreeBSD.org> AuthorDate: 2025-10-03 05:47:47 +0000 Commit: Ryan Libby <rlibby@FreeBSD.org> CommitDate: 2025-10-03 05:47:47 +0000 ddb show allchains: avoid traps and show pointers for lockmgr and sx When trying to walk lock chains, we have to deduce what a thread is blocked on. Checking LOCK_CLASS(td->td_wchan) is not a very reliable method by itself, as it just tests whether a particular 4 bits near where td_wchan points have a particular value. Misinterpreting wait channel pointers of other sleeps as lockmgr locks would frequently cause ddb show allchains (or show lockchain) to trap, or to produce incorrect output. Now, check the sleepq_type. When calling sleepq_add, we use SLEEPQ_LK for lockmgr locks and SLEEPQ_SX for sx locks. This is a more reliable indication that the td_wchan is actually a lock. While here, also make the output of lockmgr and sx locks consistent with the output for other locks (see print_lockchain). Outputting a pointer to the lock allows it to be inspected further with ddb show lock or other methods. Reviewed by: markj Sponsored by: Dell Inc. Differential Revision: https://reviews.freebsd.org/D52794 --- sys/kern/kern_lock.c | 6 ++++-- sys/kern/kern_sx.c | 11 +++++++---- 2 files changed, 11 insertions(+), 6 deletions(-) diff --git a/sys/kern/kern_lock.c b/sys/kern/kern_lock.c index 31bff6d2c1aa..76f68677e292 100644 --- a/sys/kern/kern_lock.c +++ b/sys/kern/kern_lock.c @@ -1780,9 +1780,11 @@ lockmgr_chain(struct thread *td, struct thread **ownerp) lk = td->td_wchan; - if (LOCK_CLASS(&lk->lock_object) != &lock_class_lockmgr) + if (!TD_ON_SLEEPQ(td) || sleepq_type(td->td_wchan) != SLEEPQ_LK || + LOCK_CLASS(&lk->lock_object) != &lock_class_lockmgr) return (0); - db_printf("blocked on lockmgr %s", lk->lock_object.lo_name); + db_printf("blocked on lock %p (%s) \"%s\" ", &lk->lock_object, + lock_class_lockmgr.lc_name, lk->lock_object.lo_name); if (lk->lk_lock & LK_SHARE) db_printf("SHARED (count %ju)\n", (uintmax_t)LK_SHARERS(lk->lk_lock)); diff --git a/sys/kern/kern_sx.c b/sys/kern/kern_sx.c index c005e112d3b9..249faf5b1ec4 100644 --- a/sys/kern/kern_sx.c +++ b/sys/kern/kern_sx.c @@ -1539,16 +1539,19 @@ sx_chain(struct thread *td, struct thread **ownerp) /* * Check to see if this thread is blocked on an sx lock. - * First, we check the lock class. If that is ok, then we - * compare the lock name against the wait message. + * The thread should be on a sleep queue with type SLEEPQ_SX, the + * purported lock should have the lock class index of sx, and the lock + * name should match the wait message. */ sx = td->td_wchan; - if (LOCK_CLASS(&sx->lock_object) != &lock_class_sx || + if (!TD_ON_SLEEPQ(td) || sleepq_type(td->td_wchan) != SLEEPQ_SX || + LOCK_CLASS(&sx->lock_object) != &lock_class_sx || sx->lock_object.lo_name != td->td_wmesg) return (0); /* We think we have an sx lock, so output some details. */ - db_printf("blocked on sx \"%s\" ", td->td_wmesg); + db_printf("blocked on lock %p (%s) \"%s\" ", &sx->lock_object, + lock_class_sx.lc_name, td->td_wmesg); *ownerp = sx_xholder(sx); if (sx->sx_lock & SX_LOCK_SHARED) db_printf("SLOCK (count %ju)\n",
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?202510030555.5935tGss018026>