From: Alexander Shishkin <alexander.shishkin@linux.intel.com> To: Mathieu Poirier <mathieu.poirier@linaro.org>, gregkh@linuxfoundation.org Cc: zhang.chunyan@linaro.org, mike.leach@arm.com, tor@ti.com, al.grant@arm.com, pawel.moll@arm.com, fainelli@broadcom.com, linux-arm-kernel@lists.infradead.org, linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org, Mathieu Poirier <mathieu.poirier@linaro.org> Subject: Re: [PATCH V5 21/26] coresight: etm-perf: new PMU driver for ETM tracers Date: Tue, 01 Dec 2015 01:23:01 +0200 [thread overview] Message-ID: <87poyrysve.fsf@ashishki-desk.ger.corp.intel.com> (raw) In-Reply-To: <1448849687-5724-22-git-send-email-mathieu.poirier@linaro.org> Mathieu Poirier <mathieu.poirier@linaro.org> writes: > +static void etm_event_destroy(struct perf_event *event) {} > + > +static int etm_event_init(struct perf_event *event) > +{ > + if (event->attr.type != etm_pmu.type) > + return -ENOENT; > + > + if (event->cpu >= nr_cpu_ids) > + return -EINVAL; > + > + event->destroy = etm_event_destroy; You don't have to do this if it's a nop, event::destroy can be NULL. > + > + return 0; > +} > +static void *alloc_event_data(int cpu) > +{ > + int lcpu, size; > + cpumask_t *mask; > + struct etm_cpu_data *cpu_data; > + struct etm_event_data *event_data; > + > + /* First get memory for the session's data */ > + event_data = kzalloc(sizeof(struct etm_event_data), GFP_KERNEL); > + if (!event_data) Looks like a whitespace mixup. > + return NULL; > + > + /* Make sure nothing disappears under us */ > + get_online_cpus(); > + size = num_online_cpus(); > + > + mask = &event_data->mask; > + if (cpu != -1) > + cpumask_set_cpu(cpu, mask); > + else > + cpumask_copy(mask, cpu_online_mask); It would be nice to have a comment somewhere here explaining that you have to set up tracer on each cpu in case of per-thread counter and why. We must have discussed this, but I forgot already. Btw, do you want to also set 'size' to 1 for cpu != -1 case? > + put_online_cpus(); > + > + /* Allocate an array of cpu_data to work with */ > + event_data->cpu_data = kcalloc(size, > + sizeof(struct etm_cpu_data *), > + GFP_KERNEL); > + if (!event_data->cpu_data) > + goto free_event_data; > + > + /* Allocate a cpu_data for each CPU this event is dealing with */ > + for_each_cpu(lcpu, mask) { > + cpu_data = kzalloc(sizeof(struct etm_cpu_data), GFP_KERNEL); > + if (!cpu_data) > + goto free_event_data; > + > + event_data->cpu_data[lcpu] = cpu_data; > + } Wouldn't it be easier to allocate the whole thing with one event_data->cpu_data = kcalloc(size, sizeof(struct etm_cpu_data), GFP_KERNEL); ? Regards, -- Alex
WARNING: multiple messages have this Message-ID (diff)
From: alexander.shishkin@linux.intel.com (Alexander Shishkin) To: linux-arm-kernel@lists.infradead.org Subject: [PATCH V5 21/26] coresight: etm-perf: new PMU driver for ETM tracers Date: Tue, 01 Dec 2015 01:23:01 +0200 [thread overview] Message-ID: <87poyrysve.fsf@ashishki-desk.ger.corp.intel.com> (raw) In-Reply-To: <1448849687-5724-22-git-send-email-mathieu.poirier@linaro.org> Mathieu Poirier <mathieu.poirier@linaro.org> writes: > +static void etm_event_destroy(struct perf_event *event) {} > + > +static int etm_event_init(struct perf_event *event) > +{ > + if (event->attr.type != etm_pmu.type) > + return -ENOENT; > + > + if (event->cpu >= nr_cpu_ids) > + return -EINVAL; > + > + event->destroy = etm_event_destroy; You don't have to do this if it's a nop, event::destroy can be NULL. > + > + return 0; > +} > +static void *alloc_event_data(int cpu) > +{ > + int lcpu, size; > + cpumask_t *mask; > + struct etm_cpu_data *cpu_data; > + struct etm_event_data *event_data; > + > + /* First get memory for the session's data */ > + event_data = kzalloc(sizeof(struct etm_event_data), GFP_KERNEL); > + if (!event_data) Looks like a whitespace mixup. > + return NULL; > + > + /* Make sure nothing disappears under us */ > + get_online_cpus(); > + size = num_online_cpus(); > + > + mask = &event_data->mask; > + if (cpu != -1) > + cpumask_set_cpu(cpu, mask); > + else > + cpumask_copy(mask, cpu_online_mask); It would be nice to have a comment somewhere here explaining that you have to set up tracer on each cpu in case of per-thread counter and why. We must have discussed this, but I forgot already. Btw, do you want to also set 'size' to 1 for cpu != -1 case? > + put_online_cpus(); > + > + /* Allocate an array of cpu_data to work with */ > + event_data->cpu_data = kcalloc(size, > + sizeof(struct etm_cpu_data *), > + GFP_KERNEL); > + if (!event_data->cpu_data) > + goto free_event_data; > + > + /* Allocate a cpu_data for each CPU this event is dealing with */ > + for_each_cpu(lcpu, mask) { > + cpu_data = kzalloc(sizeof(struct etm_cpu_data), GFP_KERNEL); > + if (!cpu_data) > + goto free_event_data; > + > + event_data->cpu_data[lcpu] = cpu_data; > + } Wouldn't it be easier to allocate the whole thing with one event_data->cpu_data = kcalloc(size, sizeof(struct etm_cpu_data), GFP_KERNEL); ? Regards, -- Alex
next prev parent reply other threads:[~2015-11-30 23:23 UTC|newest] Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top 2015-11-30 2:14 [PATCH V5 00/26] Coresight integration with perf Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 01/26] coresight: etm3x: moving etm_readl/writel to header file Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 02/26] coresight: etm3x: moving sysFS entries to dedicated file Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 03/26] coresight: etm3x: unlocking tracers in default arch init Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 04/26] coresight: etm3x: splitting struct etm_drvdata Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 6:54 ` kbuild test robot 2015-11-30 6:54 ` kbuild test robot 2015-11-30 2:14 ` [PATCH V5 05/26] coresight: etm3x: implementing 'cpu_id()' API Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 06/26] coresight: associating path with session rather than tracer Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 07/26] coresight: moving PM runtime operations to core framework Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 08/26] coresight: etm3x: adding operation mode for etm_enable() Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 09/26] coresight: add API to get sink from path Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 10/26] coresight: etm3x: set progbit to stop trace collection Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 11/26] coresight: etm3x: changing default trace configuration Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 12/26] coresight: etm3x: consolidating initial config Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 13/26] coresight: etm3x: implementing user/kernel mode tracing Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 14/26] coresight: etm3x: adding perf_get/set_config() API Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 15/26] coresight: etm3x: implementing perf_enable/disable() API Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 16/26] coresight: etb10: moving to local atomic operations Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 17/26] coresight: adding operation mode for sink->enable() Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 18/26] coresight: etb10: implementing AUX space API Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 19/26] coresight: updating documentation to reflect integration with perf Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 20/26] perf: changing pmu::setup_aux() parameter to include event Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 21/26] coresight: etm-perf: new PMU driver for ETM tracers Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 23:23 ` Alexander Shishkin [this message] 2015-11-30 23:23 ` Alexander Shishkin 2015-12-01 17:25 ` Mathieu Poirier 2015-12-01 17:25 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 22/26] coresight: introducing a global trace ID function Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 23/26] perf tools: making function set_max_cpu_num() non static Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 24/26] perf tools: adding perf_session to *info_prive_size() Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 16:15 ` Arnaldo Carvalho de Melo 2015-11-30 16:15 ` Arnaldo Carvalho de Melo 2015-11-30 2:14 ` [PATCH V5 25/26] perf tools: making coresight PMU listable Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier 2015-11-30 2:14 ` [PATCH V5 26/26] perf tools: adding coresight etm PMU record capabilities Mathieu Poirier 2015-11-30 2:14 ` Mathieu Poirier
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=87poyrysve.fsf@ashishki-desk.ger.corp.intel.com \ --to=alexander.shishkin@linux.intel.com \ --cc=al.grant@arm.com \ --cc=fainelli@broadcom.com \ --cc=gregkh@linuxfoundation.org \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-doc@vger.kernel.org \ --cc=linux-kernel@vger.kernel.org \ --cc=mathieu.poirier@linaro.org \ --cc=mike.leach@arm.com \ --cc=pawel.moll@arm.com \ --cc=tor@ti.com \ --cc=zhang.chunyan@linaro.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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.