linux-kernel - Re: [PATCH] perf/x86/intel/pt: Don't die on VMXON

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: <20160406085227.GO3448@twins.programming.kicks-ass.net>
Date:	Wed, 6 Apr 2016 10:52:27 +0200
From:	Peter Zijlstra <peterz@...radead.org>
To:	Alexander Shishkin <alexander.shishkin@...ux.intel.com>
Cc:	Gleb Natapov <gleb@...nel.org>,
	Paolo Bonzini <pbonzini@...hat.com>, x86@...nel.org,
	kvm@...r.kernel.org, Ingo Molnar <mingo@...hat.com>,
	linux-kernel@...r.kernel.org, tglx@...utronix.de, hpa@...or.com,
	Arnaldo Carvalho de Melo <acme@...radead.org>
Subject: Re: [PATCH] perf/x86/intel/pt: Don't die on VMXON

On Fri, Apr 01, 2016 at 07:24:14PM +0300, Alexander Shishkin wrote:
> +static void pt_config_stop(struct perf_event *event)
>  {
> +	u64 ctl = READ_ONCE(event->hw.config);
>  
> +	/* may be already stopped by a PMI*/
> +	if (!(ctl & RTIT_CTL_TRACEEN))
> +		return;
> +
> +	ctl ^= RTIT_CTL_TRACEEN;

Would that not be much less confusing when written like |= ?

>  	wrmsrl(MSR_IA32_RTIT_CTL, ctl);
>  
> +	WRITE_ONCE(event->hw.config, ctl);
> +
>  	/*
>  	 * A wrmsr that disables trace generation serializes other PT
>  	 * registers and causes all data packets to be written to memory,


> +void intel_pt_vmxon(int entry)
> +{
> +	struct pt *pt = this_cpu_ptr(&pt_ctx);
> +	struct perf_event *event;
> +	unsigned long flags;
> +
> +	/* PT plays nice with VMX, do nothing */
> +	if (pt_pmu.vmx)
> +		return;
> +
> +	/*
> +	 * VMX entry will clear RTIT_CTL.TraceEn; we need to make
> +	 * sure to not try to set it while VMX is on. Disable
> +	 * interrupts to avoid racing with pmu callbacks;
> +	 * concurrent PMI should be handled fine.
> +	 */
> +	local_irq_save(flags);
> +	WRITE_ONCE(pt->vmx_on, entry);

So you mix: "VMX is on" and "VMX entry", please pick one.

Since the function is called vmxon, I find .entry a very confusing
argument name.

> +
> +	if (entry) {
> +		/* prevent pt_config_stop() from writing RTIT_CTL */
> +		event = pt->handle.event;
> +		if (event)
> +			event->hw.config = 0;
> +	}
> +	local_irq_restore(flags);
> +}
> +EXPORT_SYMBOL_GPL(intel_pt_vmxon);