From: Yicong Yang <yangyicong@huawei.com>
To: Will Deacon <will@kernel.org>
Cc: <yangyicong@hisilicon.com>, <mark.rutland@arm.com>,
<catalin.marinas@arm.com>, <broonie@kernel.org>,
<james.morse@arm.com>, <anshuman.khandual@arm.com>,
<linux-arm-kernel@lists.infradead.org>,
<jonathan.cameron@huawei.com>,
<shameerali.kolothum.thodi@huawei.com>,
<prime.zeng@hisilicon.com>, <fanghao11@huawei.com>,
<linuxarm@huawei.com>, <maz@kernel.org>
Subject: Re: [PATCH v2 3/3] perf: arm_spe: Enable the profiling of EL0&1 translation regime
Date: Fri, 12 Apr 2024 17:22:28 +0800 [thread overview]
Message-ID: <f4642e89-d0f0-da20-66ef-85e83be9083c@huawei.com> (raw)
In-Reply-To: <20240411142829.GA26681@willie-the-truck>
On 2024/4/11 22:28, Will Deacon wrote:
> On Thu, Nov 30, 2023 at 03:46:09PM +0800, Yicong Yang wrote:
>> From: Yicong Yang <yangyicong@hisilicon.com>
>>
>> On a VHE enabled host, the PMSCR_EL1 will be redirect to PMSCR_EL2
>> and we're actually enabling E0SPE and E2SPE in the driver. This means
>> the data from EL0&1 translation regime of a VM will not be profiled.
>> So this patch tries to add the support of profiling EL0 and EL1 of
>> a VM. Users can filter data of different exception level by using
>> the perf's exclude_* attributes. The exclude_* decision is referred
>> to Documentation/arch/arm64/perf.rst and the implementation of
>> arm_pmuv3.
>>
>> Signed-off-by: Yicong Yang <yangyicong@hisilicon.com>
>> ---
>> drivers/perf/arm_spe_pmu.c | 37 ++++++++++++++++++++++++++++++-------
>> 1 file changed, 30 insertions(+), 7 deletions(-)
>>
>> diff --git a/drivers/perf/arm_spe_pmu.c b/drivers/perf/arm_spe_pmu.c
>> index 09570d4d63cd..a647d625f359 100644
>> --- a/drivers/perf/arm_spe_pmu.c
>> +++ b/drivers/perf/arm_spe_pmu.c
>> @@ -316,21 +316,44 @@ static u64 arm_spe_event_to_pmscr(struct perf_event *event)
>> static void arm_spe_pmu_set_pmscr(struct perf_event *event)
>> {
>> struct perf_event_attr *attr = &event->attr;
>> - u64 reg = 0;
>> + u64 pmscr_el1, pmscr_el12;
>>
>> - reg = arm_spe_event_to_pmscr(event);
>> - if (!attr->exclude_user)
>> - reg |= PMSCR_EL1x_E0SPE;
>> + pmscr_el1 = pmscr_el12 = arm_spe_event_to_pmscr(event);
>> +
>> + /*
>> + * Map the exclude_* descision to ELx according to
>> + * Documentation/arch/arm64/perf.rst.
>> + */
>> + if (is_kernel_in_hyp_mode()) {
>> + if (!attr->exclude_kernel && !attr->exclude_host)
>> + pmscr_el1 |= PMSCR_EL1x_E1SPE;
>>
>> - if (!attr->exclude_kernel)
>> - reg |= PMSCR_EL1x_E1SPE;
>> + if (!attr->exclude_kernel && !attr->exclude_guest)
>> + pmscr_el12 |= PMSCR_EL1x_E1SPE;
>> +
>> + if (!attr->exclude_user && !attr->exclude_host) {
>> + pmscr_el1 |= PMSCR_EL1x_E0SPE;
>> + pmscr_el12 |= PMSCR_EL1x_E0SPE;
>> + }
>
> Hmm, I don't understand this part. Doesn't this mean that setting
> 'exclude_host' to true will also exclude userspace (EL0) profiling for
> the guest?
>
I may misunderstand 'exclude_host' in the doc. Yes we won't include EL0 here
in the driver but we should handle it on guest enter/exit, which is missed
in this patch. Will see how to handle it properly on guest enter/exit and
respin a v3.
Thanks.
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
next prev parent reply other threads:[~2024-04-12 9:22 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-11-30 7:46 [PATCH v2 0/3] Enable the profiling of EL0&1 translation regime of ARM SPE Yicong Yang
2023-11-30 7:46 ` [PATCH v2 1/3] arm64/sysreg: Add PMSCR_EL12 and factor out the common fields Yicong Yang
2023-11-30 7:46 ` [PATCH v2 2/3] perf: arm_spe: Factor out PMSCR set/clear operations Yicong Yang
2023-11-30 7:46 ` [PATCH v2 3/3] perf: arm_spe: Enable the profiling of EL0&1 translation regime Yicong Yang
2024-04-11 14:28 ` Will Deacon
2024-04-12 9:22 ` Yicong Yang [this message]
2024-04-12 9:59 ` Marc Zyngier
2024-04-12 10:12 ` Yicong Yang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=f4642e89-d0f0-da20-66ef-85e83be9083c@huawei.com \
--to=yangyicong@huawei.com \
--cc=anshuman.khandual@arm.com \
--cc=broonie@kernel.org \
--cc=catalin.marinas@arm.com \
--cc=fanghao11@huawei.com \
--cc=james.morse@arm.com \
--cc=jonathan.cameron@huawei.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linuxarm@huawei.com \
--cc=mark.rutland@arm.com \
--cc=maz@kernel.org \
--cc=prime.zeng@hisilicon.com \
--cc=shameerali.kolothum.thodi@huawei.com \
--cc=will@kernel.org \
--cc=yangyicong@hisilicon.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox