From: hejunhao <hejunhao3@huawei.com>
To: Jonathan Cameron <Jonathan.Cameron@Huawei.com>,
Yicong Yang <yangyicong@huawei.com>
Cc: <will@kernel.org>, <mark.rutland@arm.com>,
<linux-arm-kernel@lists.infradead.org>,
<linux-kernel@vger.kernel.org>, <yangyicong@hisilicon.com>,
<linuxarm@huawei.com>, <prime.zeng@hisilicon.com>,
<fanghao11@huawei.com>
Subject: Re: Re: [PATCH 6/7] drivers/perf: hisi_pcie: Merge find_related_event() and get_event_idx()
Date: Thu, 22 Feb 2024 11:00:44 +0800 [thread overview]
Message-ID: <1c14d50d-fc51-cc34-aa97-358466c19fe6@huawei.com> (raw)
In-Reply-To: <20240208123946.00006618@Huawei.com>
Hi, Jonathan
On 2024/2/8 20:39, Jonathan Cameron wrote:
> On Sun, 4 Feb 2024 15:45:26 +0800
> Yicong Yang <yangyicong@huawei.com> wrote:
>
>> From: Junhao He <hejunhao3@huawei.com>
>>
>> The function xxx_find_related_event() scan all working events to find
>> related events. During this process, we also can find the idle counters.
>> If not found related events, return the first idle counter to simplify
>> the code.
>>
>> Signed-off-by: Junhao He <hejunhao3@huawei.com>
>> Signed-off-by: Yicong Yang <yangyicong@hisilicon.com>
> A suggestion inline to avoid the magic HISI_PCIE_MAX_COUNTER value
> being used outside of the function.
Thanks for the comments, will fix in next version.
>
>> ---
>> drivers/perf/hisilicon/hisi_pcie_pmu.c | 55 ++++++++++----------------
>> 1 file changed, 21 insertions(+), 34 deletions(-)
>>
>> diff --git a/drivers/perf/hisilicon/hisi_pcie_pmu.c b/drivers/perf/hisilicon/hisi_pcie_pmu.c
>> index 1b45aeb82859..2edde66675e7 100644
>> --- a/drivers/perf/hisilicon/hisi_pcie_pmu.c
>> +++ b/drivers/perf/hisilicon/hisi_pcie_pmu.c
>> @@ -397,16 +397,24 @@ static u64 hisi_pcie_pmu_read_counter(struct perf_event *event)
>> return hisi_pcie_pmu_readq(pcie_pmu, event->hw.event_base, idx);
>> }
>>
>> -static int hisi_pcie_pmu_find_related_event(struct hisi_pcie_pmu *pcie_pmu,
>> - struct perf_event *event)
>> +/*
>> + * Check all work events, if a relevant event is found then we return it
>> + * first, otherwise return the first idle counter (need to reset).
>> + */
>> +static int hisi_pcie_pmu_get_event_idx(struct hisi_pcie_pmu *pcie_pmu,
>> + struct perf_event *event)
>> {
>> + int first_idle = HISI_PCIE_MAX_COUNTERS;
> int first_idle = -EAGAIN;
Yes, I will do that.
>> struct perf_event *sibling;
>> int idx;
>>
>> for (idx = 0; idx < HISI_PCIE_MAX_COUNTERS; idx++) {
>> sibling = pcie_pmu->hw_events[idx];
>> - if (!sibling)
>> + if (!sibling) {
>> + if (first_idle == HISI_PCIE_MAX_COUNTERS)
> if (first_idle == -EAGAIN)
Ok, will fix it.
>
>> + first_idle = idx;
>> continue;
>> + }
>>
>> /* Related events must be used in group */
>> if (hisi_pcie_pmu_cmp_event(sibling, event) &&
>> @@ -414,19 +422,7 @@ static int hisi_pcie_pmu_find_related_event(struct hisi_pcie_pmu *pcie_pmu,
>> return idx;
>> }
>>
>> - return idx;
>> -}
>> -
>> -static int hisi_pcie_pmu_get_event_idx(struct hisi_pcie_pmu *pcie_pmu)
>> -{
>> - int idx;
>> -
>> - for (idx = 0; idx < HISI_PCIE_MAX_COUNTERS; idx++) {
>> - if (!pcie_pmu->hw_events[idx])
>> - return idx;
>> - }
>> -
>> - return -EINVAL;
>> + return first_idle;
> Then this will return -EAGAIN;
Okay
>> }
>>
>> static void hisi_pcie_pmu_event_update(struct perf_event *event)
>> @@ -552,27 +548,18 @@ static int hisi_pcie_pmu_add(struct perf_event *event, int flags)
>>
>> hwc->state = PERF_HES_STOPPED | PERF_HES_UPTODATE;
>>
>> - /* Check all working events to find a related event. */
>> - idx = hisi_pcie_pmu_find_related_event(pcie_pmu, event);
>> - if (idx < 0)
>> - return idx;
>> -
>> - /* Current event shares an enabled counter with the related event */
>> - if (idx < HISI_PCIE_MAX_COUNTERS) {
>> - hwc->idx = idx;
>> - goto start_count;
>> - }
>> -
>> - idx = hisi_pcie_pmu_get_event_idx(pcie_pmu);
>> - if (idx < 0)
>> - return idx;
>> + idx = hisi_pcie_pmu_get_event_idx(pcie_pmu, event);
>> + if (idx == HISI_PCIE_MAX_COUNTERS)
>> + return -EAGAIN;
> Perhaps simpler to handle first_idle == HISI_PCIE_MAX_COUNTERS as
> an error return in hisi_pcie_pmu_get_event_idx - see above.
>
> if (idx < 0)
> return idx;
Sure.
>
>>
>> hwc->idx = idx;
>> - pcie_pmu->hw_events[idx] = event;
>> - /* Reset Counter to avoid previous statistic interference. */
>> - hisi_pcie_pmu_reset_counter(pcie_pmu, idx);
>>
>> -start_count:
>> + /* No enabled counter found with related event, reset it */
>> + if (!pcie_pmu->hw_events[idx]) {
>> + hisi_pcie_pmu_reset_counter(pcie_pmu, idx);
>> + pcie_pmu->hw_events[idx] = event;
>> + }
>> +
>> if (flags & PERF_EF_START)
>> hisi_pcie_pmu_start(event, PERF_EF_RELOAD);
>>
>
next prev parent reply other threads:[~2024-02-22 3:00 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-02-04 7:45 [PATCH 0/7] Several updates for HiSilicon PCIe PMU driver Yicong Yang
2024-02-04 7:45 ` [PATCH 1/7] drivers/perf: hisi_pcie: Introduce hisi_pcie_pmu_get_filter() Yicong Yang
2024-02-08 12:06 ` Jonathan Cameron
2024-02-08 12:18 ` Jonathan Cameron
2024-02-21 9:39 ` Yicong Yang
2024-02-04 7:45 ` [PATCH 2/7] drivers/perf: hisi_pcie: Fix incorrect counting under metric mode Yicong Yang
2024-02-08 12:19 ` Jonathan Cameron
2024-02-04 7:45 ` [PATCH 3/7] drivers/perf: hisi_pcie: Add more events for counting TLP bandwidth Yicong Yang
2024-02-08 12:20 ` Jonathan Cameron
2024-02-21 9:40 ` Yicong Yang
2024-02-04 7:45 ` [PATCH 4/7] drivers/perf: hisi_pcie: Check the target filter properly Yicong Yang
2024-02-08 12:29 ` Jonathan Cameron
2024-02-21 9:46 ` Yicong Yang
2024-02-22 1:21 ` hejunhao
2024-02-22 1:29 ` hejunhao
2024-02-04 7:45 ` [PATCH 5/7] drivers/perf: hisi_pcie: Relax the check on related events Yicong Yang
2024-02-04 7:45 ` [PATCH 6/7] drivers/perf: hisi_pcie: Merge find_related_event() and get_event_idx() Yicong Yang
2024-02-08 12:39 ` Jonathan Cameron
2024-02-22 3:00 ` hejunhao [this message]
2024-02-04 7:45 ` [PATCH 7/7] docs: perf: Update usage for target filter of hisi-pcie-pmu 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=1c14d50d-fc51-cc34-aa97-358466c19fe6@huawei.com \
--to=hejunhao3@huawei.com \
--cc=Jonathan.Cameron@Huawei.com \
--cc=fanghao11@huawei.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxarm@huawei.com \
--cc=mark.rutland@arm.com \
--cc=prime.zeng@hisilicon.com \
--cc=will@kernel.org \
--cc=yangyicong@hisilicon.com \
--cc=yangyicong@huawei.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
all inboxes | Powered by JetHome®