From: Peter Zijlstra <peterz@infradead.org> To: Byungchul Park <byungchul.park@lge.com> Cc: mingo@kernel.org, tglx@linutronix.de, walken@google.com, boqun.feng@gmail.com, kirill@shutemov.name, linux-kernel@vger.kernel.org, linux-mm@kvack.org, iamjoonsoo.kim@lge.com, akpm@linux-foundation.org, willy@infradead.org, npiggin@gmail.com, kernel-team@lge.com Subject: Re: [PATCH v6 05/15] lockdep: Implement crossrelease feature Date: Mon, 24 Apr 2017 11:30:51 +0200 [thread overview] Message-ID: <20170424093051.imizyhpifqf4t6bc@hirez.programming.kicks-ass.net> (raw) In-Reply-To: <20170424030412.GG21430@X58A-UD3R> On Mon, Apr 24, 2017 at 12:04:12PM +0900, Byungchul Park wrote: > On Wed, Apr 19, 2017 at 07:19:54PM +0200, Peter Zijlstra wrote: > > > +/* > > > + * For crosslock. > > > + */ > > > +static int add_xlock(struct held_lock *hlock) > > > +{ > > > + struct cross_lock *xlock; > > > + unsigned int gen_id; > > > + > > > + if (!graph_lock()) > > > + return 0; > > > + > > > + xlock = &((struct lockdep_map_cross *)hlock->instance)->xlock; > > > + > > > + gen_id = (unsigned int)atomic_inc_return(&cross_gen_id); > > > + xlock->hlock = *hlock; > > > + xlock->hlock.gen_id = gen_id; > > > + graph_unlock(); > > > > What does graph_lock protect here? > > Modifying xlock(not xhlock) instance should be protected with graph_lock. > Don't you think so? Ah, right you are. I think I got confused between our xhlock (local) array and the xlock instance thing. The latter needs protection to serialize concurrent acquires. > > > +static int commit_xhlocks(struct cross_lock *xlock) > > > +{ > > > + unsigned int cur = current->xhlock_idx; > > > + unsigned int i; > > > + > > > + if (!graph_lock()) > > > + return 0; > > > + > > > + for (i = cur - 1; !xhlock_same(i, cur); i--) { > > > + struct hist_lock *xhlock = &xhlock(i); > > > > *blink*, you mean this? > > > > for (i = 0; i < MAX_XHLOCKS_NR; i++) { > > struct hist_lock *xhlock = &xhlock(cur - i); > > I will change the loop to this form. > > > Except you seem to skip over the most recent element (@cur), why? > > Currently 'cur' points to the next *free* slot. Well, there's no such thing has a 'free' slot, its a _ring_ buffer. But: +static void add_xhlock(struct held_lock *hlock) +{ + unsigned int idx = current->xhlock_idx++; + struct hist_lock *xhlock = &xhlock(idx); Yes, I misread that. Then '0' has the oldest entry, which is slightly weird. Should we change that? > > > + > > > + if (!xhlock_used(xhlock)) > > > + break; > > > + > > > + if (before(xhlock->hlock.gen_id, xlock->hlock.gen_id)) > > > + break; > > > + > > > + if (same_context_xhlock(xhlock) && > > > + !commit_xhlock(xlock, xhlock)) > > > > return with graph_lock held? > > No. When commit_xhlock() returns 0, the lock was already unlocked. Please add a comment, because I completely missed that. That's at least 2 functions deeper.
WARNING: multiple messages have this Message-ID (diff)
From: Peter Zijlstra <peterz@infradead.org> To: Byungchul Park <byungchul.park@lge.com> Cc: mingo@kernel.org, tglx@linutronix.de, walken@google.com, boqun.feng@gmail.com, kirill@shutemov.name, linux-kernel@vger.kernel.org, linux-mm@kvack.org, iamjoonsoo.kim@lge.com, akpm@linux-foundation.org, willy@infradead.org, npiggin@gmail.com, kernel-team@lge.com Subject: Re: [PATCH v6 05/15] lockdep: Implement crossrelease feature Date: Mon, 24 Apr 2017 11:30:51 +0200 [thread overview] Message-ID: <20170424093051.imizyhpifqf4t6bc@hirez.programming.kicks-ass.net> (raw) In-Reply-To: <20170424030412.GG21430@X58A-UD3R> On Mon, Apr 24, 2017 at 12:04:12PM +0900, Byungchul Park wrote: > On Wed, Apr 19, 2017 at 07:19:54PM +0200, Peter Zijlstra wrote: > > > +/* > > > + * For crosslock. > > > + */ > > > +static int add_xlock(struct held_lock *hlock) > > > +{ > > > + struct cross_lock *xlock; > > > + unsigned int gen_id; > > > + > > > + if (!graph_lock()) > > > + return 0; > > > + > > > + xlock = &((struct lockdep_map_cross *)hlock->instance)->xlock; > > > + > > > + gen_id = (unsigned int)atomic_inc_return(&cross_gen_id); > > > + xlock->hlock = *hlock; > > > + xlock->hlock.gen_id = gen_id; > > > + graph_unlock(); > > > > What does graph_lock protect here? > > Modifying xlock(not xhlock) instance should be protected with graph_lock. > Don't you think so? Ah, right you are. I think I got confused between our xhlock (local) array and the xlock instance thing. The latter needs protection to serialize concurrent acquires. > > > +static int commit_xhlocks(struct cross_lock *xlock) > > > +{ > > > + unsigned int cur = current->xhlock_idx; > > > + unsigned int i; > > > + > > > + if (!graph_lock()) > > > + return 0; > > > + > > > + for (i = cur - 1; !xhlock_same(i, cur); i--) { > > > + struct hist_lock *xhlock = &xhlock(i); > > > > *blink*, you mean this? > > > > for (i = 0; i < MAX_XHLOCKS_NR; i++) { > > struct hist_lock *xhlock = &xhlock(cur - i); > > I will change the loop to this form. > > > Except you seem to skip over the most recent element (@cur), why? > > Currently 'cur' points to the next *free* slot. Well, there's no such thing has a 'free' slot, its a _ring_ buffer. But: +static void add_xhlock(struct held_lock *hlock) +{ + unsigned int idx = current->xhlock_idx++; + struct hist_lock *xhlock = &xhlock(idx); Yes, I misread that. Then '0' has the oldest entry, which is slightly weird. Should we change that? > > > + > > > + if (!xhlock_used(xhlock)) > > > + break; > > > + > > > + if (before(xhlock->hlock.gen_id, xlock->hlock.gen_id)) > > > + break; > > > + > > > + if (same_context_xhlock(xhlock) && > > > + !commit_xhlock(xlock, xhlock)) > > > > return with graph_lock held? > > No. When commit_xhlock() returns 0, the lock was already unlocked. Please add a comment, because I completely missed that. That's at least 2 functions deeper. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2017-04-24 9:31 UTC|newest] Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top 2017-03-14 8:18 [PATCH v6 00/15] lockdep: Implement crossrelease feature Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 01/15] lockdep: Refactor lookup_chain_cache() Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 02/15] lockdep: Add a function building a chain between two classes Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 03/15] lockdep: Change the meaning of check_prev_add()'s return value Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 04/15] lockdep: Make check_prev_add() able to handle external stack_trace Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 05/15] lockdep: Implement crossrelease feature Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-04-19 14:25 ` Peter Zijlstra 2017-04-19 14:25 ` Peter Zijlstra 2017-04-24 5:11 ` Byungchul Park 2017-04-24 5:11 ` Byungchul Park 2017-04-24 10:17 ` Peter Zijlstra 2017-04-24 10:17 ` Peter Zijlstra 2017-04-25 5:40 ` Byungchul Park 2017-04-25 5:40 ` Byungchul Park 2017-05-16 14:18 ` Peter Zijlstra 2017-05-16 14:18 ` Peter Zijlstra 2017-05-18 6:22 ` Byungchul Park 2017-05-18 6:22 ` Byungchul Park 2017-04-19 15:08 ` Peter Zijlstra 2017-04-19 15:08 ` Peter Zijlstra 2017-04-24 4:36 ` Byungchul Park 2017-04-24 4:36 ` Byungchul Park 2017-04-19 17:19 ` Peter Zijlstra 2017-04-19 17:19 ` Peter Zijlstra 2017-04-24 3:04 ` Byungchul Park 2017-04-24 3:04 ` Byungchul Park 2017-04-24 9:30 ` Peter Zijlstra [this message] 2017-04-24 9:30 ` Peter Zijlstra 2017-04-25 6:59 ` Byungchul Park 2017-04-25 6:59 ` Byungchul Park 2017-04-19 17:20 ` Peter Zijlstra 2017-04-19 17:20 ` Peter Zijlstra 2017-04-24 3:13 ` Byungchul Park 2017-04-24 3:13 ` Byungchul Park 2017-05-19 8:07 ` Byungchul Park 2017-05-19 8:07 ` Byungchul Park 2017-05-19 10:30 ` Peter Zijlstra 2017-05-19 10:30 ` Peter Zijlstra 2017-05-19 10:56 ` Byungchul Park 2017-05-19 10:56 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 06/15] lockdep: Handle non(or multi)-acquisition of a crosslock Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 07/15] lockdep: Avoid adding redundant direct links of crosslocks Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 08/15] lockdep: Fix incorrect condition to print bug msgs for MAX_LOCKDEP_CHAIN_HLOCKS Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 09/15] lockdep: Make print_circular_bug() aware of crossrelease Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 10/15] lockdep: Apply crossrelease to completions Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 11/15] pagemap.h: Remove trailing white space Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:18 ` [PATCH v6 12/15] lockdep: Apply crossrelease to PG_locked locks Byungchul Park 2017-03-14 8:18 ` Byungchul Park 2017-03-14 8:19 ` [PATCH v6 13/15] lockdep: Apply lock_acquire(release) on __Set(__Clear)PageLocked Byungchul Park 2017-03-14 8:19 ` Byungchul Park 2017-03-14 8:19 ` [PATCH v6 14/15] lockdep: Move data of CONFIG_LOCKDEP_PAGELOCK from page to page_ext Byungchul Park 2017-03-14 8:19 ` Byungchul Park 2017-03-14 8:19 ` [PATCH v6 15/15] lockdep: Crossrelease feature documentation Byungchul Park 2017-03-14 8:19 ` Byungchul Park
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20170424093051.imizyhpifqf4t6bc@hirez.programming.kicks-ass.net \ --to=peterz@infradead.org \ --cc=akpm@linux-foundation.org \ --cc=boqun.feng@gmail.com \ --cc=byungchul.park@lge.com \ --cc=iamjoonsoo.kim@lge.com \ --cc=kernel-team@lge.com \ --cc=kirill@shutemov.name \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-mm@kvack.org \ --cc=mingo@kernel.org \ --cc=npiggin@gmail.com \ --cc=tglx@linutronix.de \ --cc=walken@google.com \ --cc=willy@infradead.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.