| Message ID | 20190605161633.12245-1-mathieu.poirier@linaro.org (mailing list archive) |
|---|---|
| State | Mainlined |
| Commit | e45c48a9a4d20ebc7b639a62c3ef8f4b08007027 |
| Headers | show |
| Series | perf tools: Properly set the value of 'old' and 'head' in snapshot mode | expand |
Em Wed, Jun 05, 2019 at 10:16:33AM -0600, Mathieu Poirier escreveu: > This patch adds the necessay intelligence to properly compute the value > of 'old' and 'head' when operating in snapshot mode. That way we can get > the latest information in the AUX buffer and be compatible with the > generic AUX ring buffer mechanic. Leo, have you had the chance to test/review this one? Suzuki? I also changed the subject to: [PATCH] perf cs-etm: Properly set the value of 'old' and 'head' in snapshot mode So that when looking at a 'git log --oneline' one can have the proper context and know that its about cs-etm. - Arnaldo > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> > --- > tools/perf/arch/arm/util/cs-etm.c | 127 +++++++++++++++++++++++++++++- > 1 file changed, 123 insertions(+), 4 deletions(-) > > diff --git a/tools/perf/arch/arm/util/cs-etm.c b/tools/perf/arch/arm/util/cs-etm.c > index 911426721170..0a278bbcaba6 100644 > --- a/tools/perf/arch/arm/util/cs-etm.c > +++ b/tools/perf/arch/arm/util/cs-etm.c > @@ -31,6 +31,8 @@ struct cs_etm_recording { > struct auxtrace_record itr; > struct perf_pmu *cs_etm_pmu; > struct perf_evlist *evlist; > + int wrapped_cnt; > + bool *wrapped; > bool snapshot_mode; > size_t snapshot_size; > }; > @@ -536,16 +538,131 @@ static int cs_etm_info_fill(struct auxtrace_record *itr, > return 0; > } > > -static int cs_etm_find_snapshot(struct auxtrace_record *itr __maybe_unused, > +static int cs_etm_alloc_wrapped_array(struct cs_etm_recording *ptr, int idx) > +{ > + bool *wrapped; > + int cnt = ptr->wrapped_cnt; > + > + /* Make @ptr->wrapped as big as @idx */ > + while (cnt <= idx) > + cnt++; > + > + /* > + * Free'ed in cs_etm_recording_free(). Using realloc() to avoid > + * cross compilation problems where the host's system supports > + * reallocarray() but not the target. > + */ > + wrapped = realloc(ptr->wrapped, cnt * sizeof(bool)); > + if (!wrapped) > + return -ENOMEM; > + > + wrapped[cnt - 1] = false; > + ptr->wrapped_cnt = cnt; > + ptr->wrapped = wrapped; > + > + return 0; > +} > + > +static bool cs_etm_buffer_has_wrapped(unsigned char *buffer, > + size_t buffer_size, u64 head) > +{ > + u64 i, watermark; > + u64 *buf = (u64 *)buffer; > + size_t buf_size = buffer_size; > + > + /* > + * We want to look the very last 512 byte (chosen arbitrarily) in > + * the ring buffer. > + */ > + watermark = buf_size - 512; > + > + /* > + * @head is continuously increasing - if its value is equal or greater > + * than the size of the ring buffer, it has wrapped around. > + */ > + if (head >= buffer_size) > + return true; > + > + /* > + * The value of @head is somewhere within the size of the ring buffer. > + * This can be that there hasn't been enough data to fill the ring > + * buffer yet or the trace time was so long that @head has numerically > + * wrapped around. To find we need to check if we have data at the very > + * end of the ring buffer. We can reliably do this because mmap'ed > + * pages are zeroed out and there is a fresh mapping with every new > + * session. > + */ > + > + /* @head is less than 512 byte from the end of the ring buffer */ > + if (head > watermark) > + watermark = head; > + > + /* > + * Speed things up by using 64 bit transactions (see "u64 *buf" above) > + */ > + watermark >>= 3; > + buf_size >>= 3; > + > + /* > + * If we find trace data at the end of the ring buffer, @head has > + * been there and has numerically wrapped around at least once. > + */ > + for (i = watermark; i < buf_size; i++) > + if (buf[i]) > + return true; > + > + return false; > +} > + > +static int cs_etm_find_snapshot(struct auxtrace_record *itr, > int idx, struct auxtrace_mmap *mm, > - unsigned char *data __maybe_unused, > + unsigned char *data, > u64 *head, u64 *old) > { > + int err; > + bool wrapped; > + struct cs_etm_recording *ptr = > + container_of(itr, struct cs_etm_recording, itr); > + > + /* > + * Allocate memory to keep track of wrapping if this is the first > + * time we deal with this *mm. > + */ > + if (idx >= ptr->wrapped_cnt) { > + err = cs_etm_alloc_wrapped_array(ptr, idx); > + if (err) > + return err; > + } > + > + /* > + * Check to see if *head has wrapped around. If it hasn't only the > + * amount of data between *head and *old is snapshot'ed to avoid > + * bloating the perf.data file with zeros. But as soon as *head has > + * wrapped around the entire size of the AUX ring buffer it taken. > + */ > + wrapped = ptr->wrapped[idx]; > + if (!wrapped && cs_etm_buffer_has_wrapped(data, mm->len, *head)) { > + wrapped = true; > + ptr->wrapped[idx] = true; > + } > + > pr_debug3("%s: mmap index %d old head %zu new head %zu size %zu\n", > __func__, idx, (size_t)*old, (size_t)*head, mm->len); > > - *old = *head; > - *head += mm->len; > + /* No wrap has occurred, we can just use *head and *old. */ > + if (!wrapped) > + return 0; > + > + /* > + * *head has wrapped around - adjust *head and *old to pickup the > + * entire content of the AUX buffer. > + */ > + if (*head >= mm->len) { > + *old = *head - mm->len; > + } else { > + *head += mm->len; > + *old = *head - mm->len; > + } > > return 0; > } > @@ -586,6 +703,8 @@ static void cs_etm_recording_free(struct auxtrace_record *itr) > { > struct cs_etm_recording *ptr = > container_of(itr, struct cs_etm_recording, itr); > + > + zfree(&ptr->wrapped); > free(ptr); > } > > -- > 2.17.1
On Thu, Jun 06, 2019 at 05:10:56PM -0300, Arnaldo Carvalho de Melo wrote: > Em Wed, Jun 05, 2019 at 10:16:33AM -0600, Mathieu Poirier escreveu: > > This patch adds the necessay intelligence to properly compute the value > > of 'old' and 'head' when operating in snapshot mode. That way we can get > > the latest information in the AUX buffer and be compatible with the > > generic AUX ring buffer mechanic. > > Leo, have you had the chance to test/review this one? Suzuki? Sure. I applied this patch on the perf/core branch (with latest commit 3e4fbf36c1e3 'perf augmented_raw_syscalls: Move reading filename to the loop') and passed testing with below steps: # perf record -e cs_etm/@tmc_etr0/ -S -m,64 --per-thread ./sort & [1] 19097 Bubble sorting array of 30000 elements # kill -USR2 19097 # kill -USR2 19097 # kill -USR2 19097 [ perf record: Woken up 4 times to write data ] [ perf record: Captured and wrote 0.753 MB perf.data ] FWIW: Tested-by: Leo Yan <leo.yan@linaro.org> > I also changed the subject to: > > [PATCH] perf cs-etm: Properly set the value of 'old' and 'head' in snapshot mode > > So that when looking at a 'git log --oneline' one can have the proper > context and know that its about cs-etm. > > - Arnaldo > > > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> > > --- > > tools/perf/arch/arm/util/cs-etm.c | 127 +++++++++++++++++++++++++++++- > > 1 file changed, 123 insertions(+), 4 deletions(-) > > > > diff --git a/tools/perf/arch/arm/util/cs-etm.c b/tools/perf/arch/arm/util/cs-etm.c > > index 911426721170..0a278bbcaba6 100644 > > --- a/tools/perf/arch/arm/util/cs-etm.c > > +++ b/tools/perf/arch/arm/util/cs-etm.c > > @@ -31,6 +31,8 @@ struct cs_etm_recording { > > struct auxtrace_record itr; > > struct perf_pmu *cs_etm_pmu; > > struct perf_evlist *evlist; > > + int wrapped_cnt; > > + bool *wrapped; > > bool snapshot_mode; > > size_t snapshot_size; > > }; > > @@ -536,16 +538,131 @@ static int cs_etm_info_fill(struct auxtrace_record *itr, > > return 0; > > } > > > > -static int cs_etm_find_snapshot(struct auxtrace_record *itr __maybe_unused, > > +static int cs_etm_alloc_wrapped_array(struct cs_etm_recording *ptr, int idx) > > +{ > > + bool *wrapped; > > + int cnt = ptr->wrapped_cnt; > > + > > + /* Make @ptr->wrapped as big as @idx */ > > + while (cnt <= idx) > > + cnt++; > > + > > + /* > > + * Free'ed in cs_etm_recording_free(). Using realloc() to avoid > > + * cross compilation problems where the host's system supports > > + * reallocarray() but not the target. > > + */ > > + wrapped = realloc(ptr->wrapped, cnt * sizeof(bool)); > > + if (!wrapped) > > + return -ENOMEM; > > + > > + wrapped[cnt - 1] = false; > > + ptr->wrapped_cnt = cnt; > > + ptr->wrapped = wrapped; > > + > > + return 0; > > +} > > + > > +static bool cs_etm_buffer_has_wrapped(unsigned char *buffer, > > + size_t buffer_size, u64 head) > > +{ > > + u64 i, watermark; > > + u64 *buf = (u64 *)buffer; > > + size_t buf_size = buffer_size; > > + > > + /* > > + * We want to look the very last 512 byte (chosen arbitrarily) in > > + * the ring buffer. > > + */ > > + watermark = buf_size - 512; > > + > > + /* > > + * @head is continuously increasing - if its value is equal or greater > > + * than the size of the ring buffer, it has wrapped around. > > + */ > > + if (head >= buffer_size) > > + return true; > > + > > + /* > > + * The value of @head is somewhere within the size of the ring buffer. > > + * This can be that there hasn't been enough data to fill the ring > > + * buffer yet or the trace time was so long that @head has numerically > > + * wrapped around. To find we need to check if we have data at the very > > + * end of the ring buffer. We can reliably do this because mmap'ed > > + * pages are zeroed out and there is a fresh mapping with every new > > + * session. > > + */ > > + > > + /* @head is less than 512 byte from the end of the ring buffer */ > > + if (head > watermark) > > + watermark = head; > > + > > + /* > > + * Speed things up by using 64 bit transactions (see "u64 *buf" above) > > + */ > > + watermark >>= 3; > > + buf_size >>= 3; > > + > > + /* > > + * If we find trace data at the end of the ring buffer, @head has > > + * been there and has numerically wrapped around at least once. > > + */ > > + for (i = watermark; i < buf_size; i++) > > + if (buf[i]) > > + return true; > > + > > + return false; > > +} > > + > > +static int cs_etm_find_snapshot(struct auxtrace_record *itr, > > int idx, struct auxtrace_mmap *mm, > > - unsigned char *data __maybe_unused, > > + unsigned char *data, > > u64 *head, u64 *old) > > { > > + int err; > > + bool wrapped; > > + struct cs_etm_recording *ptr = > > + container_of(itr, struct cs_etm_recording, itr); > > + > > + /* > > + * Allocate memory to keep track of wrapping if this is the first > > + * time we deal with this *mm. > > + */ > > + if (idx >= ptr->wrapped_cnt) { > > + err = cs_etm_alloc_wrapped_array(ptr, idx); > > + if (err) > > + return err; > > + } > > + > > + /* > > + * Check to see if *head has wrapped around. If it hasn't only the > > + * amount of data between *head and *old is snapshot'ed to avoid > > + * bloating the perf.data file with zeros. But as soon as *head has > > + * wrapped around the entire size of the AUX ring buffer it taken. > > + */ > > + wrapped = ptr->wrapped[idx]; > > + if (!wrapped && cs_etm_buffer_has_wrapped(data, mm->len, *head)) { > > + wrapped = true; > > + ptr->wrapped[idx] = true; > > + } > > + > > pr_debug3("%s: mmap index %d old head %zu new head %zu size %zu\n", > > __func__, idx, (size_t)*old, (size_t)*head, mm->len); > > > > - *old = *head; > > - *head += mm->len; > > + /* No wrap has occurred, we can just use *head and *old. */ > > + if (!wrapped) > > + return 0; > > + > > + /* > > + * *head has wrapped around - adjust *head and *old to pickup the > > + * entire content of the AUX buffer. > > + */ > > + if (*head >= mm->len) { > > + *old = *head - mm->len; > > + } else { > > + *head += mm->len; > > + *old = *head - mm->len; > > + } > > > > return 0; > > } > > @@ -586,6 +703,8 @@ static void cs_etm_recording_free(struct auxtrace_record *itr) > > { > > struct cs_etm_recording *ptr = > > container_of(itr, struct cs_etm_recording, itr); > > + > > + zfree(&ptr->wrapped); > > free(ptr); > > } > > > > -- > > 2.17.1 > > -- > > - Arnaldo
On Thu, 6 Jun 2019 at 14:11, Arnaldo Carvalho de Melo <arnaldo.melo@gmail.com> wrote: > > Em Wed, Jun 05, 2019 at 10:16:33AM -0600, Mathieu Poirier escreveu: > > This patch adds the necessay intelligence to properly compute the value > > of 'old' and 'head' when operating in snapshot mode. That way we can get > > the latest information in the AUX buffer and be compatible with the > > generic AUX ring buffer mechanic. > > Leo, have you had the chance to test/review this one? Suzuki? Leo did test this before and added his Tested-by on the Coresight mailing list. I did not carried it here because I changed the call to reallocarray() to realloc() in order to avoid cross compilation problems. I think it is safe enough but other people's opinion may differ so I played it safe. Leo, please test this again if/when you have the time. > > I also changed the subject to: > > [PATCH] perf cs-etm: Properly set the value of 'old' and 'head' in snapshot mode > > So that when looking at a 'git log --oneline' one can have the proper > context and know that its about cs-etm. Very well. Mathieu > > - Arnaldo > > > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> > > --- > > tools/perf/arch/arm/util/cs-etm.c | 127 +++++++++++++++++++++++++++++- > > 1 file changed, 123 insertions(+), 4 deletions(-) > > > > diff --git a/tools/perf/arch/arm/util/cs-etm.c b/tools/perf/arch/arm/util/cs-etm.c > > index 911426721170..0a278bbcaba6 100644 > > --- a/tools/perf/arch/arm/util/cs-etm.c > > +++ b/tools/perf/arch/arm/util/cs-etm.c > > @@ -31,6 +31,8 @@ struct cs_etm_recording { > > struct auxtrace_record itr; > > struct perf_pmu *cs_etm_pmu; > > struct perf_evlist *evlist; > > + int wrapped_cnt; > > + bool *wrapped; > > bool snapshot_mode; > > size_t snapshot_size; > > }; > > @@ -536,16 +538,131 @@ static int cs_etm_info_fill(struct auxtrace_record *itr, > > return 0; > > } > > > > -static int cs_etm_find_snapshot(struct auxtrace_record *itr __maybe_unused, > > +static int cs_etm_alloc_wrapped_array(struct cs_etm_recording *ptr, int idx) > > +{ > > + bool *wrapped; > > + int cnt = ptr->wrapped_cnt; > > + > > + /* Make @ptr->wrapped as big as @idx */ > > + while (cnt <= idx) > > + cnt++; > > + > > + /* > > + * Free'ed in cs_etm_recording_free(). Using realloc() to avoid > > + * cross compilation problems where the host's system supports > > + * reallocarray() but not the target. > > + */ > > + wrapped = realloc(ptr->wrapped, cnt * sizeof(bool)); > > + if (!wrapped) > > + return -ENOMEM; > > + > > + wrapped[cnt - 1] = false; > > + ptr->wrapped_cnt = cnt; > > + ptr->wrapped = wrapped; > > + > > + return 0; > > +} > > + > > +static bool cs_etm_buffer_has_wrapped(unsigned char *buffer, > > + size_t buffer_size, u64 head) > > +{ > > + u64 i, watermark; > > + u64 *buf = (u64 *)buffer; > > + size_t buf_size = buffer_size; > > + > > + /* > > + * We want to look the very last 512 byte (chosen arbitrarily) in > > + * the ring buffer. > > + */ > > + watermark = buf_size - 512; > > + > > + /* > > + * @head is continuously increasing - if its value is equal or greater > > + * than the size of the ring buffer, it has wrapped around. > > + */ > > + if (head >= buffer_size) > > + return true; > > + > > + /* > > + * The value of @head is somewhere within the size of the ring buffer. > > + * This can be that there hasn't been enough data to fill the ring > > + * buffer yet or the trace time was so long that @head has numerically > > + * wrapped around. To find we need to check if we have data at the very > > + * end of the ring buffer. We can reliably do this because mmap'ed > > + * pages are zeroed out and there is a fresh mapping with every new > > + * session. > > + */ > > + > > + /* @head is less than 512 byte from the end of the ring buffer */ > > + if (head > watermark) > > + watermark = head; > > + > > + /* > > + * Speed things up by using 64 bit transactions (see "u64 *buf" above) > > + */ > > + watermark >>= 3; > > + buf_size >>= 3; > > + > > + /* > > + * If we find trace data at the end of the ring buffer, @head has > > + * been there and has numerically wrapped around at least once. > > + */ > > + for (i = watermark; i < buf_size; i++) > > + if (buf[i]) > > + return true; > > + > > + return false; > > +} > > + > > +static int cs_etm_find_snapshot(struct auxtrace_record *itr, > > int idx, struct auxtrace_mmap *mm, > > - unsigned char *data __maybe_unused, > > + unsigned char *data, > > u64 *head, u64 *old) > > { > > + int err; > > + bool wrapped; > > + struct cs_etm_recording *ptr = > > + container_of(itr, struct cs_etm_recording, itr); > > + > > + /* > > + * Allocate memory to keep track of wrapping if this is the first > > + * time we deal with this *mm. > > + */ > > + if (idx >= ptr->wrapped_cnt) { > > + err = cs_etm_alloc_wrapped_array(ptr, idx); > > + if (err) > > + return err; > > + } > > + > > + /* > > + * Check to see if *head has wrapped around. If it hasn't only the > > + * amount of data between *head and *old is snapshot'ed to avoid > > + * bloating the perf.data file with zeros. But as soon as *head has > > + * wrapped around the entire size of the AUX ring buffer it taken. > > + */ > > + wrapped = ptr->wrapped[idx]; > > + if (!wrapped && cs_etm_buffer_has_wrapped(data, mm->len, *head)) { > > + wrapped = true; > > + ptr->wrapped[idx] = true; > > + } > > + > > pr_debug3("%s: mmap index %d old head %zu new head %zu size %zu\n", > > __func__, idx, (size_t)*old, (size_t)*head, mm->len); > > > > - *old = *head; > > - *head += mm->len; > > + /* No wrap has occurred, we can just use *head and *old. */ > > + if (!wrapped) > > + return 0; > > + > > + /* > > + * *head has wrapped around - adjust *head and *old to pickup the > > + * entire content of the AUX buffer. > > + */ > > + if (*head >= mm->len) { > > + *old = *head - mm->len; > > + } else { > > + *head += mm->len; > > + *old = *head - mm->len; > > + } > > > > return 0; > > } > > @@ -586,6 +703,8 @@ static void cs_etm_recording_free(struct auxtrace_record *itr) > > { > > struct cs_etm_recording *ptr = > > container_of(itr, struct cs_etm_recording, itr); > > + > > + zfree(&ptr->wrapped); > > free(ptr); > > } > > > > -- > > 2.17.1 > > -- > > - Arnaldo
Em Fri, Jun 07, 2019 at 02:44:25PM +0800, Leo Yan escreveu: > On Thu, Jun 06, 2019 at 05:10:56PM -0300, Arnaldo Carvalho de Melo wrote: > > Em Wed, Jun 05, 2019 at 10:16:33AM -0600, Mathieu Poirier escreveu: > > > This patch adds the necessay intelligence to properly compute the value > > > of 'old' and 'head' when operating in snapshot mode. That way we can get > > > the latest information in the AUX buffer and be compatible with the > > > generic AUX ring buffer mechanic. > > > > Leo, have you had the chance to test/review this one? Suzuki? > > Sure. I applied this patch on the perf/core branch (with latest > commit 3e4fbf36c1e3 'perf augmented_raw_syscalls: Move reading > filename to the loop') and passed testing with below steps: > > # perf record -e cs_etm/@tmc_etr0/ -S -m,64 --per-thread ./sort & > [1] 19097 > Bubble sorting array of 30000 elements > > # kill -USR2 19097 > # kill -USR2 19097 > # kill -USR2 19097 > [ perf record: Woken up 4 times to write data ] > [ perf record: Captured and wrote 0.753 MB perf.data ] > > FWIW: > > Tested-by: Leo Yan <leo.yan@linaro.org> Thanks a lot, I've added your "Tester notes:" and also your Tested-by:. As I don't have hardware (yet) to test these patches, tests by people who can test on real hardware is always super appreciated. Any suggestions for a SBC that I could buy to be able to do so? Regards, - Arnaldo > > I also changed the subject to: > > > > [PATCH] perf cs-etm: Properly set the value of 'old' and 'head' in snapshot mode > > > > So that when looking at a 'git log --oneline' one can have the proper > > context and know that its about cs-etm. > > > > - Arnaldo > > > > > Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> > > > --- > > > tools/perf/arch/arm/util/cs-etm.c | 127 +++++++++++++++++++++++++++++- > > > 1 file changed, 123 insertions(+), 4 deletions(-) > > > > > > diff --git a/tools/perf/arch/arm/util/cs-etm.c b/tools/perf/arch/arm/util/cs-etm.c > > > index 911426721170..0a278bbcaba6 100644 > > > --- a/tools/perf/arch/arm/util/cs-etm.c > > > +++ b/tools/perf/arch/arm/util/cs-etm.c > > > @@ -31,6 +31,8 @@ struct cs_etm_recording { > > > struct auxtrace_record itr; > > > struct perf_pmu *cs_etm_pmu; > > > struct perf_evlist *evlist; > > > + int wrapped_cnt; > > > + bool *wrapped; > > > bool snapshot_mode; > > > size_t snapshot_size; > > > }; > > > @@ -536,16 +538,131 @@ static int cs_etm_info_fill(struct auxtrace_record *itr, > > > return 0; > > > } > > > > > > -static int cs_etm_find_snapshot(struct auxtrace_record *itr __maybe_unused, > > > +static int cs_etm_alloc_wrapped_array(struct cs_etm_recording *ptr, int idx) > > > +{ > > > + bool *wrapped; > > > + int cnt = ptr->wrapped_cnt; > > > + > > > + /* Make @ptr->wrapped as big as @idx */ > > > + while (cnt <= idx) > > > + cnt++; > > > + > > > + /* > > > + * Free'ed in cs_etm_recording_free(). Using realloc() to avoid > > > + * cross compilation problems where the host's system supports > > > + * reallocarray() but not the target. > > > + */ > > > + wrapped = realloc(ptr->wrapped, cnt * sizeof(bool)); > > > + if (!wrapped) > > > + return -ENOMEM; > > > + > > > + wrapped[cnt - 1] = false; > > > + ptr->wrapped_cnt = cnt; > > > + ptr->wrapped = wrapped; > > > + > > > + return 0; > > > +} > > > + > > > +static bool cs_etm_buffer_has_wrapped(unsigned char *buffer, > > > + size_t buffer_size, u64 head) > > > +{ > > > + u64 i, watermark; > > > + u64 *buf = (u64 *)buffer; > > > + size_t buf_size = buffer_size; > > > + > > > + /* > > > + * We want to look the very last 512 byte (chosen arbitrarily) in > > > + * the ring buffer. > > > + */ > > > + watermark = buf_size - 512; > > > + > > > + /* > > > + * @head is continuously increasing - if its value is equal or greater > > > + * than the size of the ring buffer, it has wrapped around. > > > + */ > > > + if (head >= buffer_size) > > > + return true; > > > + > > > + /* > > > + * The value of @head is somewhere within the size of the ring buffer. > > > + * This can be that there hasn't been enough data to fill the ring > > > + * buffer yet or the trace time was so long that @head has numerically > > > + * wrapped around. To find we need to check if we have data at the very > > > + * end of the ring buffer. We can reliably do this because mmap'ed > > > + * pages are zeroed out and there is a fresh mapping with every new > > > + * session. > > > + */ > > > + > > > + /* @head is less than 512 byte from the end of the ring buffer */ > > > + if (head > watermark) > > > + watermark = head; > > > + > > > + /* > > > + * Speed things up by using 64 bit transactions (see "u64 *buf" above) > > > + */ > > > + watermark >>= 3; > > > + buf_size >>= 3; > > > + > > > + /* > > > + * If we find trace data at the end of the ring buffer, @head has > > > + * been there and has numerically wrapped around at least once. > > > + */ > > > + for (i = watermark; i < buf_size; i++) > > > + if (buf[i]) > > > + return true; > > > + > > > + return false; > > > +} > > > + > > > +static int cs_etm_find_snapshot(struct auxtrace_record *itr, > > > int idx, struct auxtrace_mmap *mm, > > > - unsigned char *data __maybe_unused, > > > + unsigned char *data, > > > u64 *head, u64 *old) > > > { > > > + int err; > > > + bool wrapped; > > > + struct cs_etm_recording *ptr = > > > + container_of(itr, struct cs_etm_recording, itr); > > > + > > > + /* > > > + * Allocate memory to keep track of wrapping if this is the first > > > + * time we deal with this *mm. > > > + */ > > > + if (idx >= ptr->wrapped_cnt) { > > > + err = cs_etm_alloc_wrapped_array(ptr, idx); > > > + if (err) > > > + return err; > > > + } > > > + > > > + /* > > > + * Check to see if *head has wrapped around. If it hasn't only the > > > + * amount of data between *head and *old is snapshot'ed to avoid > > > + * bloating the perf.data file with zeros. But as soon as *head has > > > + * wrapped around the entire size of the AUX ring buffer it taken. > > > + */ > > > + wrapped = ptr->wrapped[idx]; > > > + if (!wrapped && cs_etm_buffer_has_wrapped(data, mm->len, *head)) { > > > + wrapped = true; > > > + ptr->wrapped[idx] = true; > > > + } > > > + > > > pr_debug3("%s: mmap index %d old head %zu new head %zu size %zu\n", > > > __func__, idx, (size_t)*old, (size_t)*head, mm->len); > > > > > > - *old = *head; > > > - *head += mm->len; > > > + /* No wrap has occurred, we can just use *head and *old. */ > > > + if (!wrapped) > > > + return 0; > > > + > > > + /* > > > + * *head has wrapped around - adjust *head and *old to pickup the > > > + * entire content of the AUX buffer. > > > + */ > > > + if (*head >= mm->len) { > > > + *old = *head - mm->len; > > > + } else { > > > + *head += mm->len; > > > + *old = *head - mm->len; > > > + } > > > > > > return 0; > > > } > > > @@ -586,6 +703,8 @@ static void cs_etm_recording_free(struct auxtrace_record *itr) > > > { > > > struct cs_etm_recording *ptr = > > > container_of(itr, struct cs_etm_recording, itr); > > > + > > > + zfree(&ptr->wrapped); > > > free(ptr); > > > } > > > > > > -- > > > 2.17.1 > > > > -- > > > > - Arnaldo
Hi Arnaldo, On Fri, Jun 07, 2019 at 03:23:25PM -0300, Arnaldo Carvalho de Melo wrote: > Em Fri, Jun 07, 2019 at 02:44:25PM +0800, Leo Yan escreveu: > > On Thu, Jun 06, 2019 at 05:10:56PM -0300, Arnaldo Carvalho de Melo wrote: > > > Em Wed, Jun 05, 2019 at 10:16:33AM -0600, Mathieu Poirier escreveu: > > > > This patch adds the necessay intelligence to properly compute the value > > > > of 'old' and 'head' when operating in snapshot mode. That way we can get > > > > the latest information in the AUX buffer and be compatible with the > > > > generic AUX ring buffer mechanic. > > > > > > Leo, have you had the chance to test/review this one? Suzuki? > > > > Sure. I applied this patch on the perf/core branch (with latest > > commit 3e4fbf36c1e3 'perf augmented_raw_syscalls: Move reading > > filename to the loop') and passed testing with below steps: > > > > # perf record -e cs_etm/@tmc_etr0/ -S -m,64 --per-thread ./sort & > > [1] 19097 > > Bubble sorting array of 30000 elements > > > > # kill -USR2 19097 > > # kill -USR2 19097 > > # kill -USR2 19097 > > [ perf record: Woken up 4 times to write data ] > > [ perf record: Captured and wrote 0.753 MB perf.data ] > > > > FWIW: > > > > Tested-by: Leo Yan <leo.yan@linaro.org> > > Thanks a lot, I've added your "Tester notes:" and also your Tested-by:. > > As I don't have hardware (yet) to test these patches, tests by people > who can test on real hardware is always super appreciated. You are very welcome and it's my pleasure :) > Any suggestions for a SBC that I could buy to be able to do so? Below are several Arm development boards for referrence: - DB410c [1]: This board is the first choice for myself, since this board provides Debian (and Fedora :) support and it supports the mainline kernel pretty well; the CoreSight also is well supported. This board is about 80 USD so the cost is not expensive; on the other hand, please note one cons is the SDRAM is only 1GB, this will be impossible if you build some big projects (e.g. LLVM/Clang and BCC); but it's sufficient for perf related development and verification. - There have other several boards are in my mind: Raspberry Pi3 [2] and Hikey960 [3]. Raspberry Pi3 misses some features in the mainline kernel [4] and it has not enabled CoreSight hardware tracing feature; Hikey960 also have some patches are out of the mainline kenrel. Except you have special requirement (e.g. you want to use the board to build LLVM/Clang/BCC with big DDR size, etc), these two boards can be secondary choices. Please feel free let me know if you have questions for boards. Thanks, Leo Yan [1] https://www.96boards.org/product/dragonboard410c/ [2] https://www.raspberrypi.org/products/raspberry-pi-3-model-b-plus/ [3] https://www.96boards.org/product/hikey960/ [4] https://www.raspberrypi.org/forums/viewtopic.php?t=236568
diff --git a/tools/perf/arch/arm/util/cs-etm.c b/tools/perf/arch/arm/util/cs-etm.c index 911426721170..0a278bbcaba6 100644 --- a/tools/perf/arch/arm/util/cs-etm.c +++ b/tools/perf/arch/arm/util/cs-etm.c @@ -31,6 +31,8 @@ struct cs_etm_recording { struct auxtrace_record itr; struct perf_pmu *cs_etm_pmu; struct perf_evlist *evlist; + int wrapped_cnt; + bool *wrapped; bool snapshot_mode; size_t snapshot_size; }; @@ -536,16 +538,131 @@ static int cs_etm_info_fill(struct auxtrace_record *itr, return 0; } -static int cs_etm_find_snapshot(struct auxtrace_record *itr __maybe_unused, +static int cs_etm_alloc_wrapped_array(struct cs_etm_recording *ptr, int idx) +{ + bool *wrapped; + int cnt = ptr->wrapped_cnt; + + /* Make @ptr->wrapped as big as @idx */ + while (cnt <= idx) + cnt++; + + /* + * Free'ed in cs_etm_recording_free(). Using realloc() to avoid + * cross compilation problems where the host's system supports + * reallocarray() but not the target. + */ + wrapped = realloc(ptr->wrapped, cnt * sizeof(bool)); + if (!wrapped) + return -ENOMEM; + + wrapped[cnt - 1] = false; + ptr->wrapped_cnt = cnt; + ptr->wrapped = wrapped; + + return 0; +} + +static bool cs_etm_buffer_has_wrapped(unsigned char *buffer, + size_t buffer_size, u64 head) +{ + u64 i, watermark; + u64 *buf = (u64 *)buffer; + size_t buf_size = buffer_size; + + /* + * We want to look the very last 512 byte (chosen arbitrarily) in + * the ring buffer. + */ + watermark = buf_size - 512; + + /* + * @head is continuously increasing - if its value is equal or greater + * than the size of the ring buffer, it has wrapped around. + */ + if (head >= buffer_size) + return true; + + /* + * The value of @head is somewhere within the size of the ring buffer. + * This can be that there hasn't been enough data to fill the ring + * buffer yet or the trace time was so long that @head has numerically + * wrapped around. To find we need to check if we have data at the very + * end of the ring buffer. We can reliably do this because mmap'ed + * pages are zeroed out and there is a fresh mapping with every new + * session. + */ + + /* @head is less than 512 byte from the end of the ring buffer */ + if (head > watermark) + watermark = head; + + /* + * Speed things up by using 64 bit transactions (see "u64 *buf" above) + */ + watermark >>= 3; + buf_size >>= 3; + + /* + * If we find trace data at the end of the ring buffer, @head has + * been there and has numerically wrapped around at least once. + */ + for (i = watermark; i < buf_size; i++) + if (buf[i]) + return true; + + return false; +} + +static int cs_etm_find_snapshot(struct auxtrace_record *itr, int idx, struct auxtrace_mmap *mm, - unsigned char *data __maybe_unused, + unsigned char *data, u64 *head, u64 *old) { + int err; + bool wrapped; + struct cs_etm_recording *ptr = + container_of(itr, struct cs_etm_recording, itr); + + /* + * Allocate memory to keep track of wrapping if this is the first + * time we deal with this *mm. + */ + if (idx >= ptr->wrapped_cnt) { + err = cs_etm_alloc_wrapped_array(ptr, idx); + if (err) + return err; + } + + /* + * Check to see if *head has wrapped around. If it hasn't only the + * amount of data between *head and *old is snapshot'ed to avoid + * bloating the perf.data file with zeros. But as soon as *head has + * wrapped around the entire size of the AUX ring buffer it taken. + */ + wrapped = ptr->wrapped[idx]; + if (!wrapped && cs_etm_buffer_has_wrapped(data, mm->len, *head)) { + wrapped = true; + ptr->wrapped[idx] = true; + } + pr_debug3("%s: mmap index %d old head %zu new head %zu size %zu\n", __func__, idx, (size_t)*old, (size_t)*head, mm->len); - *old = *head; - *head += mm->len; + /* No wrap has occurred, we can just use *head and *old. */ + if (!wrapped) + return 0; + + /* + * *head has wrapped around - adjust *head and *old to pickup the + * entire content of the AUX buffer. + */ + if (*head >= mm->len) { + *old = *head - mm->len; + } else { + *head += mm->len; + *old = *head - mm->len; + } return 0; } @@ -586,6 +703,8 @@ static void cs_etm_recording_free(struct auxtrace_record *itr) { struct cs_etm_recording *ptr = container_of(itr, struct cs_etm_recording, itr); + + zfree(&ptr->wrapped); free(ptr); }
This patch adds the necessay intelligence to properly compute the value of 'old' and 'head' when operating in snapshot mode. That way we can get the latest information in the AUX buffer and be compatible with the generic AUX ring buffer mechanic. Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org> --- tools/perf/arch/arm/util/cs-etm.c | 127 +++++++++++++++++++++++++++++- 1 file changed, 123 insertions(+), 4 deletions(-)