[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aBUF6JduwlhHVR9E@arm.com>
Date: Fri, 2 May 2025 18:50:32 +0100
From: Catalin Marinas <catalin.marinas@....com>
To: Yeoreum Yun <yeoreum.yun@....com>
Cc: will@...nel.org, broonie@...nel.org, anshuman.khandual@....com,
joey.gouly@....com, maz@...nel.org, oliver.upton@...ux.dev,
frederic@...nel.org, james.morse@....com,
hardevsinh.palaniya@...iconsignals.io,
shameerali.kolothum.thodi@...wei.com, huangxiaojia2@...wei.com,
mark.rutland@....com, samuel.holland@...ive.com,
palmer@...osinc.com, charlie@...osinc.com,
thiago.bauermann@...aro.org, bgray@...ux.ibm.com,
tglx@...utronix.de, puranjay@...nel.org, david@...hat.com,
yang@...amperecomputing.com, mbenes@...e.cz,
joel.granados@...nel.org, linux-arm-kernel@...ts.infradead.org,
linux-kernel@...r.kernel.org, nd@....com,
Peter Collingbourne <pcc@...gle.com>
Subject: Re: [PATCH v3 3/4] arm64/kernel: support store-only mte tag check
On Thu, Apr 10, 2025 at 09:07:22AM +0100, Yeoreum Yun wrote:
> Introduce new flag -- MTE_CTRL_STORE_ONLY used to set store-only tag check.
> This flag isn't overrided by prefered tcf flag setting but set together
Nit: s/overrided/overridden/
> with prefered setting of way to report tag check fault.
The preferred mode set via sysfs is about whether we want synchronous or
asynchronous tag check faults for reads/writes (or asymmetric). The
store-only checking can be combined with sync/async, so they are
slightly complementary. The question is whether one wants some global
knob to turn on store-only in combination with sync/async. We could add
more strings for sysfs like "(a)sync+storeonly"
It would be good to hear Peter's opinion from an Android perspective.
--
Catalin
Powered by blists - more mailing lists