Re: [PATCH v8 1/3] perf: cavium: Support memory controller PMU counters

From	Suzuki K Poulose <Suzuki.Poulose@arm.com>
Newsgroups	linux.kernel
Subject	Re: [PATCH v8 1/3] perf: cavium: Support memory controller PMU counters
Date	2017-08-07 11:40 +0200
Message-ID	<ubMFX-6qF-3@gated-at.bofh.it> (permalink)
References	(6 earlier) <u7AN4-4TN-13@gated-at.bofh.it> <u7AN4-4TN-23@gated-at.bofh.it> <u7BJ7-5vg-9@gated-at.bofh.it> <u7Jnk-1WL-17@gated-at.bofh.it> <u7QRQ-6zj-15@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

On 27/07/17 10:08, Jan Glauber wrote:
> On Thu, Jul 27, 2017 at 07:11:57AM +0200, Borislav Petkov wrote:
>> On Wed, Jul 26, 2017 at 02:02:42PM -0700, David Daney wrote:
>>> Also, if a given configuration disables CONFIG_EDAC there is some hackery
>>> needed to get the perf portion of the driver included.
>>
>> Yes, and we don't do performance counters in EDAC.
>>
>> So you could add a small memory controller driver which does the
>> arbitration or fix the firmware.
>
> OK. As fixing the firmware will take quite some time I'll go for the memory
> controller driver that starts EDAC / PMU depending on their CONFIG_.

Please could you also rename CAVIUM_PMU to something like CAVIUM_MMC_PMU or even
CAVIUM_UNCOR_PMU ? The symbol could easily be mistaken for a "cavium" CPU PMU.

Suzuki


>
> What would be the proper location for the multiplexer?
> drivers/soc/cavium or drivers/misc?
>
> --Jan
>

Back to linux.kernel | Previous | Next | Find similar | Unroll thread

Thread

Re: [PATCH v8 1/3] perf: cavium: Support memory controller PMU  counters Suzuki K Poulose <Suzuki.Poulose@arm.com> - 2017-08-07 11:40 +0200

csiph-web