[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <5bc35f89-1b77-4c0c-b6ea-fe9b2cb74870@citrix.com>
Date: Wed, 29 Oct 2025 01:42:42 +0000
From: Andrew Cooper <andrew.cooper3@...rix.com>
To: torvalds@...ux-foundation.org
Cc: alex@...ti.fr, aou@...s.berkeley.edu, axboe@...nel.dk, bp@...en8.de,
brauner@...nel.org, catalin.marinas@....com, christophe.leroy@...roup.eu,
dave.hansen@...el.com, dave.hansen@...ux.intel.com,
david.laight.linux@...il.com, edumazet@...gle.com, hpa@...or.com,
kuni1840@...il.com, kuniyu@...gle.com, linux-arm-kernel@...ts.infradead.org,
linux-kernel@...r.kernel.org, linux-riscv@...ts.infradead.org,
linuxppc-dev@...ts.ozlabs.org, maddy@...ux.ibm.com, mingo@...hat.com,
mpe@...erman.id.au, npiggin@...il.com, palmer@...belt.com, pjw@...nel.org,
tglx@...utronix.de, will@...nel.org, x86@...nel.org
Subject: Re: [PATCH v1 2/2] epoll: Use __user_write_access_begin() and
unsafe_put_user() in epoll_put_uevent().
> Yeah, clac/stac are horrenously expensive on Zen 2. I think they are
> microcoded - and not the fast kind - so effectively serializing.
AIUI, Zen5 is the first uarch capable of renaming the AC flag.
This make STAC/CLAC "lfence-like" in practice on older CPUs, because of
an implementation constraint of not being able to execute speculatively.
~Andrew
Powered by blists - more mailing lists