[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Y70it59wuvsnKJK1@hirez.programming.kicks-ass.net>
Date: Tue, 10 Jan 2023 09:32:55 +0100
From: Peter Zijlstra <peterz@...radead.org>
To: Heiko Carstens <hca@...ux.ibm.com>
Cc: Thomas Richter <tmricht@...ux.ibm.com>,
torvalds@...ux-foundation.org, corbet@....net, will@...nel.org,
boqun.feng@...il.com, mark.rutland@....com,
catalin.marinas@....com, dennis@...nel.org, tj@...nel.org,
cl@...ux.com, gor@...ux.ibm.com, agordeev@...ux.ibm.com,
borntraeger@...ux.ibm.com, svens@...ux.ibm.com,
Herbert Xu <herbert@...dor.apana.org.au>, davem@...emloft.net,
tglx@...utronix.de, mingo@...hat.com, bp@...en8.de,
dave.hansen@...ux.intel.com, x86@...nel.org, hpa@...or.com,
joro@...tes.org, suravee.suthikulpanit@....com,
robin.murphy@....com, dwmw2@...radead.org,
baolu.lu@...ux.intel.com, Arnd Bergmann <arnd@...db.de>,
penberg@...nel.org, rientjes@...gle.com, iamjoonsoo.kim@....com,
Andrew Morton <akpm@...ux-foundation.org>, vbabka@...e.cz,
roman.gushchin@...ux.dev, 42.hyeyoo@...il.com,
linux-doc@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-mm@...ck.org, linux-s390@...r.kernel.org,
linux-crypto@...r.kernel.org, iommu@...ts.linux.dev,
linux-arch@...r.kernel.org
Subject: Re: [RFC][PATCH 08/12] s390: Replace cmpxchg_double() with
cmpxchg128()
On Tue, Jan 10, 2023 at 08:23:05AM +0100, Heiko Carstens wrote:
> So, Alexander Gordeev reported that this code was already prior to your
> changes potentially broken with respect to missing READ_ONCE() within the
> cmpxchg_double() loops.
Unless there's an early exit, that shouldn't matter. If you managed to
read garbage the cmpxchg itself will simply fail and the loop retries.
> @@ -1294,12 +1306,16 @@ static void hw_perf_event_update(struct perf_event *event, int flush_all)
> num_sdb++;
>
> /* Reset trailer (using compare-double-and-swap) */
> + /* READ_ONCE() 16 byte header */
> + prev.val = __cdsg(&te->header.val, 0, 0);
> do {
> + old.val = prev.val;
> + new.val = prev.val;
> + new.f = 0;
> + new.a = 1;
> + new.overflow = 0;
> + prev.val = __cdsg(&te->header.val, old.val, new.val);
> + } while (prev.val != old.val);
So this, and
> + /* READ_ONCE() 16 byte header */
> + prev.val = __cdsg(&te->header.val, 0, 0);
> do {
> + old.val = prev.val;
> + new.val = prev.val;
> + orig_overflow = old.overflow;
> + new.f = 0;
> + new.overflow = 0;
> if (idx == aux->alert_mark)
> + new.a = 1;
> else
> + new.a = 0;
> + prev.val = __cdsg(&te->header.val, old.val, new.val);
> + } while (prev.val != old.val);
this case are just silly and expensive. If that initial read is split
and manages to read gibberish the cmpxchg will fail and we retry anyway.
> + /* READ_ONCE() 16 byte header */
> + prev.val = __cdsg(&te->header.val, 0, 0);
> do {
> + old.val = prev.val;
> + new.val = prev.val;
> + *overflow = old.overflow;
> + if (old.f) {
> /*
> * SDB is already set by hardware.
> * Abort and try to set somewhere
> @@ -1490,10 +1509,10 @@ static bool aux_set_alert(struct aux_buffer *aux, unsigned long alert_index,
> */
> return false;
> }
> + new.a = 1;
> + new.overflow = 0;
> + prev.val = __cdsg(&te->header.val, old.val, new.val);
> + } while (prev.val != old.val);
And while this case has an early exit, it only cares about a single bit
(although you made it a full word) and so also shouldn't care. If
aux_reset_buffer() returns false, @overflow isn't consumed.
So I really don't see the point of this patch.
Powered by blists - more mailing lists