From: Mao Han <han_mao@c-sky.com>
To: linux-kernel@vger.kernel.org
Cc: Mao Han <han_mao@c-sky.com>,
linux-csky@vger.kernel.org, Guo Ren <guoren@kernel.org>
Subject: [PATCH V5 2/6] csky: Add count-width property for csky pmu
Date: Tue, 4 Jun 2019 18:54:45 +0800 [thread overview]
Message-ID: <93df91a6bdd2fe46dcf71fdd996fa1e164a8adf4.1559644961.git.han_mao@c-sky.com> (raw)
In-Reply-To: <cover.1559644961.git.han_mao@c-sky.com>
In-Reply-To: <cover.1559644961.git.han_mao@c-sky.com>
The csky pmu counter may have different io width. When the counter is
smaller then 64 bits and counter value is smaller than the old value, it
will result to a extremely large delta value. So the sampled value should
be extend to 64 bits to avoid this, the extension bits base on the
count-width property from dts.
Signed-off-by: Mao Han <han_mao@c-sky.com>
Cc: Guo Ren <guoren@kernel.org>
---
arch/csky/kernel/perf_event.c | 15 ++++++++++++++-
1 file changed, 14 insertions(+), 1 deletion(-)
diff --git a/arch/csky/kernel/perf_event.c b/arch/csky/kernel/perf_event.c
index 2282554..a15b397 100644
--- a/arch/csky/kernel/perf_event.c
+++ b/arch/csky/kernel/perf_event.c
@@ -9,6 +9,7 @@
#include <linux/platform_device.h>
#define CSKY_PMU_MAX_EVENTS 32
+#define DEFAULT_COUNT_WIDTH 48
#define HPCR "<0, 0x0>" /* PMU Control reg */
#define HPCNTENR "<0, 0x4>" /* Count Enable reg */
@@ -18,6 +19,7 @@ static void (*hw_raw_write_mapping[CSKY_PMU_MAX_EVENTS])(uint64_t val);
struct csky_pmu_t {
struct pmu pmu;
+ uint32_t count_width;
uint32_t hpcr;
} csky_pmu;
@@ -804,7 +806,12 @@ static void csky_perf_event_update(struct perf_event *event,
struct hw_perf_event *hwc)
{
uint64_t prev_raw_count = local64_read(&hwc->prev_count);
- uint64_t new_raw_count = hw_raw_read_mapping[hwc->idx]();
+ /*
+ * Sign extend count value to 64bit, otherwise delta calculation
+ * would be incorrect when overflow occurs.
+ */
+ uint64_t new_raw_count = sign_extend64(
+ hw_raw_read_mapping[hwc->idx](), csky_pmu.count_width - 1);
int64_t delta = new_raw_count - prev_raw_count;
/*
@@ -1032,6 +1039,7 @@ int init_hw_perf_events(void)
int csky_pmu_device_probe(struct platform_device *pdev,
const struct of_device_id *of_table)
{
+ struct device_node *node = pdev->dev.of_node;
int ret;
ret = init_hw_perf_events();
@@ -1040,6 +1048,11 @@ int csky_pmu_device_probe(struct platform_device *pdev,
return ret;
}
+ if (of_property_read_u32(node, "count-width",
+ &csky_pmu.count_width)) {
+ csky_pmu.count_width = DEFAULT_COUNT_WIDTH;
+ }
+
return ret;
}
--
2.7.4
next prev parent reply other threads:[~2019-06-04 10:56 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-04 10:54 [PATCH V5 0/6] csky: Add pmu hardware sampling support Mao Han
2019-06-04 10:54 ` [PATCH V5 1/6] csky: Init pmu as a device Mao Han
2019-06-04 10:54 ` Mao Han [this message]
2019-06-04 10:54 ` [PATCH V5 3/6] csky: Add pmu interrupt support Mao Han
2019-06-04 10:54 ` [PATCH V5 4/6] dt-bindings: csky: Add csky PMU bindings Mao Han
2019-06-04 10:54 ` [PATCH V5 5/6] csky: Fixup some error count in 810 & 860 Mao Han
2019-06-04 10:54 ` [PATCH V5 6/6] csky: Fix perf record in kernel/user space Mao Han
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=93df91a6bdd2fe46dcf71fdd996fa1e164a8adf4.1559644961.git.han_mao@c-sky.com \
--to=han_mao@c-sky.com \
--cc=guoren@kernel.org \
--cc=linux-csky@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
/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;
as well as URLs for NNTP newsgroup(s).