Re: [PATCH v6 2/7] arm64: perf: Avoid PMXEV* indirection

From: Will Deacon <will@kernel.org>
To: Alexandru Elisei <alexandru.elisei@arm.com>
Cc: mark.rutland@arm.com, sumit.garg@linaro.org,
	Julien Thierry <julien.thierry@arm.com>,
	Peter Zijlstra <peterz@infradead.org>,
	maz@kernel.org, Will Deacon <will.deacon@arm.com>,
	linux-kernel@vger.kernel.org,
	Arnaldo Carvalho de Melo <acme@kernel.org>,
	swboyd@chromium.org,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Ingo Molnar <mingo@redhat.com>,
	Julien Thierry <julien.thierry.kdev@gmail.com>,
	catalin.marinas@arm.com, Namhyung Kim <namhyung@kernel.org>,
	Jiri Olsa <jolsa@redhat.com>,
	linux-arm-kernel@lists.infradead.org
Subject: Re: [PATCH v6 2/7] arm64: perf: Avoid PMXEV* indirection
Date: Mon, 21 Sep 2020 14:43:58 +0100	[thread overview]
Message-ID: <20200921134357.GL2139@willie-the-truck> (raw)
In-Reply-To: <20200819133419.526889-3-alexandru.elisei@arm.com>

On Wed, Aug 19, 2020 at 02:34:14PM +0100, Alexandru Elisei wrote:
> From: Mark Rutland <mark.rutland@arm.com>
> 
> Currently we access the counter registers and their respective type
> registers indirectly. This requires us to write to PMSELR, issue an ISB,
> then access the relevant PMXEV* registers.
> 
> This is unfortunate, because:
> 
> * Under virtualization, accessing one register requires two traps to
>   the hypervisor, even though we could access the register directly with
>   a single trap.
> 
> * We have to issue an ISB which we could otherwise avoid the cost of.
> 
> * When we use NMIs, the NMI handler will have to save/restore the select
>   register in case the code it preempted was attempting to access a
>   counter or its type register.
> 
> We can avoid these issues by directly accessing the relevant registers.
> This patch adds helpers to do so.
> 
> In armv8pmu_enable_event() we still need the ISB to prevent the PE from
> reordering the write to PMINTENSET_EL1 register. If the interrupt is
> enabled before we disable the counter and the new event is configured,
> we might get an interrupt triggered by the previously programmed event
> overflowing, but which we wrongly attribute to the event that we are
> enabling.
> 
> In the process, remove the comment that refers to the ARMv7 PMU.
> 
> Cc: Julien Thierry <julien.thierry.kdev@gmail.com>
> Cc: Will Deacon <will.deacon@arm.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Ingo Molnar <mingo@redhat.com>
> Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
> Cc: Jiri Olsa <jolsa@redhat.com>
> Cc: Namhyung Kim <namhyung@kernel.org>
> Cc: Catalin Marinas <catalin.marinas@arm.com>
> Signed-off-by: Mark Rutland <mark.rutland@arm.com>
> [Julien T.: Don't inline read/write functions to avoid big code-size
> 	increase, remove unused read_pmevtypern function,
> 	fix counter index issue.]
> Signed-off-by: Julien Thierry <julien.thierry@arm.com>
> [Removed comment, removed trailing semicolons in macros, added ISB]

nit: but it's customary to prefix these with your name, so it's easy to
figure out who made changes (like Julien did above).

(similar comment for other patches in this series)

> @@ -620,9 +686,14 @@ static void armv8pmu_enable_event(struct perf_event *event)
>  	 * Disable counter
>  	 */
>  	armv8pmu_disable_event_counter(event);
> +	/*
> +	 * Make sure the effects of disabling the counter are visible before we
> +	 * start configuring the event.
> +	 */
> +	isb();

With the isb() added by patch 1, why don't we just make these implicit
in armv8_{enable,disable}_event_counter() ?

Will

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel