linux-kernel - Re: [PATCH 3/5] arm64: refactor delay() to enable polling for value

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Zys0TAHZzqbGst93@arm.com>
Date: Wed, 6 Nov 2024 09:18:04 +0000
From: Catalin Marinas <catalin.marinas@....com>
To: Haris Okanovic <harisokn@...zon.com>
Cc: ankur.a.arora@...cle.com, linux-pm@...r.kernel.org, kvm@...r.kernel.org,
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
	will@...nel.org, tglx@...utronix.de, mingo@...hat.com, bp@...en8.de,
	dave.hansen@...ux.intel.com, x86@...nel.org, hpa@...or.com,
	pbonzini@...hat.com, wanpengli@...cent.com, vkuznets@...hat.com,
	rafael@...nel.org, daniel.lezcano@...aro.org, peterz@...radead.org,
	arnd@...db.de, lenb@...nel.org, mark.rutland@....com,
	mtosatti@...hat.com, sudeep.holla@....com, cl@...two.org,
	misono.tomohiro@...itsu.com, maobibo@...ngson.cn,
	joao.m.martins@...cle.com, boris.ostrovsky@...cle.com,
	konrad.wilk@...cle.com
Subject: Re: [PATCH 3/5] arm64: refactor delay() to enable polling for value

On Tue, Nov 05, 2024 at 12:30:39PM -0600, Haris Okanovic wrote:
> +		do {
> +			cur = __READ_ONCE_EX(*addr);
> +			if ((cur & mask) == val) {
> +				break;
> +			}
>  			wfet(end);

Constructs like this need to be entirely in assembly. The compiler may
spill 'cur' onto the stack and the write could clear the exclusive
monitor which makes the wfet return immediately. That's highly CPU
implementation specific but it's the reason we have functions like
__cmpwait() in assembly (or whatever else deals with exclusives). IOW,
we can't have other memory accesses between the LDXR and whatever is
consuming the exclusive monitor armed state (typically STXR but WFE/WFET
can be another).

-- 
Catalin