| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Thu, 23 Jul 2015 17:14:35 +0900
From: Chanwoo Choi <cw00.choi@...sung.com>
To: Krzysztof Kozlowski <k.kozlowski@...sung.com>
Cc: myungjoo.ham@...sung.com, kyungmin.park@...sung.com,
kgene@...nel.org, dan.carpenter@...cle.com,
linux-pm@...r.kernel.org, linux-samsung-soc@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/2] PM / devfreq: exynos-ppmu: Add the support of PPMUv2
for Exynos5433
Hi Krzysztof,
On 07/23/2015 04:28 PM, Krzysztof Kozlowski wrote:
> On 23.07.2015 10:57, Chanwoo Choi wrote:
>> This patch adds the support for PPMU (Platform Performance Monitoring Unit)
>> version 2.0 for Exynos5433 SoC. Exynos5433 SoC must need PPMUv2 which is
>> quite different from PPMUv1.1. The exynos-ppmu.c driver supports both PPMUv1.1
>> and PPMUv2.
>>
>> Cc: MyungJoo Ham <myungjoo.ham@...sung.com>
>> Cc: Kyungmin Park <kyungmin.park@...sung.com>
>> Signed-off-by: Chanwoo Choi <cw00.choi@...sung.com>
>
> Hi,
>
> Few comments at the end.
>
>> ---
>> drivers/devfreq/event/exynos-ppmu.c | 168 ++++++++++++++++++++++++++++++++++--
>> drivers/devfreq/event/exynos-ppmu.h | 70 +++++++++++++++
>> 2 files changed, 231 insertions(+), 7 deletions(-)
>>
>> diff --git a/drivers/devfreq/event/exynos-ppmu.c b/drivers/devfreq/event/exynos-ppmu.c
>> index 7d99d13bacd8..6066dc18fc73 100644
>> --- a/drivers/devfreq/event/exynos-ppmu.c
>> +++ b/drivers/devfreq/event/exynos-ppmu.c
>> @@ -1,7 +1,7 @@
>> /*
>> * exynos_ppmu.c - EXYNOS PPMU (Platform Performance Monitoring Unit) support
>> *
>> - * Copyright (c) 2014 Samsung Electronics Co., Ltd.
>> + * Copyright (c) 2014-2015 Samsung Electronics Co., Ltd.
>> * Author : Chanwoo Choi <cw00.choi@...sung.com>
>> *
>> * This program is free software; you can redistribute it and/or modify
>> @@ -82,6 +82,15 @@ struct __exynos_ppmu_events {
>> PPMU_EVENT(mscl),
>> PPMU_EVENT(fimd0x),
>> PPMU_EVENT(fimd1x),
(snip)
>> +static int exynos_ppmu_v2_get_event(struct devfreq_event_dev *edev,
>> + struct devfreq_event_data *edata)
>> +{
>> + struct exynos_ppmu *info = devfreq_event_get_drvdata(edev);
>> + int id = exynos_ppmu_find_ppmu_id(edev);
>> + u32 pmnc, cntenc;
>> + u32 pmcnt_high, pmcnt_low;
>> + u64 load_count = 0;
>> +
>> + /* Disable PPMU */
>> + pmnc = __raw_readl(info->ppmu.base + PPMUv2_PMNC);
>> + pmnc &= ~PPMU_PMNC_ENABLE_MASK;
>> + __raw_writel(pmnc, info->ppmu.base + PPMUv2_PMNC);
>> +
>> + /* Read cycle count and performance count */
>> + edata->total_count = __raw_readl(info->ppmu.base + PPMUv2_CCNT);
>> +
>> + switch (id) {
>> + case PPMU_PMNCNT0:
>> + case PPMU_PMNCNT1:
>> + case PPMU_PMNCNT2:
>> + load_count = __raw_readl(info->ppmu.base + PPMUv2_PMNCT(id));
>> + break;
>> + case PPMU_PMNCNT3:
>> + pmcnt_high = __raw_readl(info->ppmu.base + PPMUv2_PMCNT3_HIGH);
>> + pmcnt_low = __raw_readl(info->ppmu.base + PPMUv2_PMCNT3_LOW);
>> + load_count = (u64)((pmcnt_high & 0xff) << 32) + (u64)pmcnt_low;
>> + break;
>> + }
>> + edata->load_count = load_count;
>> +
>> + /* Disable all counters */
>> + cntenc = __raw_readl(info->ppmu.base + PPMUv2_CNTENC);
>> + cntenc |= (PPMU_CCNT_MASK | (PPMU_ENABLE << id));
>> + __raw_writel(cntenc, info->ppmu.base + PPMUv2_CNTENC);
>> +
>> + dev_dbg(&edev->dev, "%25s (load: %ld / %ld)\n", edev->desc->name,
>> + edata->load_count, edata->total_count);
>> + return 0;
>> +}
>> +
>> +static struct devfreq_event_ops exynos_ppmu_v2_ops = {
>
> This can be const.
OK.
>
>> + .disable = exynos_ppmu_v2_disable,
>> + .set_event = exynos_ppmu_v2_set_event,
>> + .get_event = exynos_ppmu_v2_get_event,
>> +};
>> +
>> +static struct of_device_id exynos_ppmu_id_match[] = {
>
> In the previous patch this was not present but now it can be made 'const'.
OK for adding the 'const'.
The original 'exynos_ppmu_id_match' is located on below of this patch.
>
>> + {
>> + .compatible = "samsung,exynos-ppmu",
>> + .data = (void *)&exynos_ppmu_ops,
>> + }, {
>> + .compatible = "samsung,exynos-ppmu-v2",
>> + .data = (void *)&exynos_ppmu_v2_ops,
>> + },
>> + { /* sentinel */ },
>> +};
>> +
>> +static struct devfreq_event_ops *exynos_bus_get_ops(struct device_node *np)
>> +{
>> + const struct of_device_id *match;
>> +
>> + match = of_match_node(exynos_ppmu_id_match, np);
>> + return (struct devfreq_event_ops *)match->data;
>> +}
>> +
>> static int of_get_devfreq_events(struct device_node *np,
>> struct exynos_ppmu *info)
>> {
>> @@ -238,7 +397,7 @@ static int of_get_devfreq_events(struct device_node *np,
>> continue;
>> }
>>
>> - desc[j].ops = &exynos_ppmu_ops;
>> + desc[j].ops = exynos_bus_get_ops(np);
>
> So for each ops callback (get/set/disable) you will have another layer
> of indirection where you will be matching the device each time?
>
> That seems like a waste of CPU time. Just match it once here and use
> either old ops or new ops_v2.
OK. I'll rework to reduce the unneeded operation.
>
>
>> desc[j].driver_data = info;
>>
>> of_property_read_string(node, "event-name", &desc[j].name);
>> @@ -354,11 +513,6 @@ static int exynos_ppmu_remove(struct platform_device *pdev)
>> return 0;
>> }
>>
>> -static struct of_device_id exynos_ppmu_id_match[] = {
>> - { .compatible = "samsung,exynos-ppmu", },
>> - { /* sentinel */ },
>> -};
>> -
>> static struct platform_driver exynos_ppmu_driver = {
>> .probe = exynos_ppmu_probe,
>> .remove = exynos_ppmu_remove,
>> diff --git a/drivers/devfreq/event/exynos-ppmu.h b/drivers/devfreq/event/exynos-ppmu.h
>> index 4e831d48c138..9a7cf6394f37 100644
>> --- a/drivers/devfreq/event/exynos-ppmu.h
>> +++ b/drivers/devfreq/event/exynos-ppmu.h
>> @@ -26,6 +26,9 @@ enum ppmu_counter {
>> PPMU_PMNCNT_MAX,
>> };
>>
>> +/***
>> + * PPMUv1.1 Definitions
>> + */
>> enum ppmu_event_type {
>> PPMU_RO_BUSY_CYCLE_CNT = 0x0,
>> PPMU_WO_BUSY_CYCLE_CNT = 0x1,
>> @@ -90,4 +93,71 @@ enum ppmu_reg {
>> #define PPMU_PMNCT(x) (PPMU_PMCNT0 + (0x10 * x))
>> #define PPMU_BEVTxSEL(x) (PPMU_BEVT0SEL + (0x100 * x))
>>
>> +/***
>> + * PPMUv2.0 definitions
>> + */
>> +enum ppmuv2_mode {
>> + PPMUv2_MODE_MANUAL = 0,
>> + PPMUv2_MODE_AUTO = 1,
>> + PPMUv2_MODE_CIG = 2, /* CIG (Conditional Interrupt Generation) */
>
> Mixing lower-upper case looks odd to me. How about:
> PPMU2_MODE_MANUAL = 0,
> or
> PPMU_V2_MODE_MANUAL = 0,
> ?
>
> (here and everywhere below)
As you know, after deciding the compatible name about using '-v2',
we can decide the correct register name.
Best Regards,
Chanwoo Choi
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists