| Message ID | 20190813163757.473-1-i.maximets@samsung.com |
|---|---|
| State | Accepted |
| Headers | show |
| Series | [ovs-dev] dpif-netdev: Add core id in the PMD thread name. | expand |
On 13 Aug 2019, at 18:37, Ilya Maximets wrote: > This is highly useful to see on which core PMD is running by > only looking at the thread name. Thread Id still allows to > distinguish different threads running on the same core over the time: > > |dpif_netdev(pmd-c10/id:53)|DBG|Creating 2. subtable <...> > |dpif_netdev(pmd-c10/id:53)|DBG|flow_add: <...>, actions:2 > |dpif_netdev(pmd-c09/id:70)|DBG|Core 9 processing port <..> > > In gdb, top or any other utility it's useful to quickly catch up > needed thread without parsing logs, memory or matching threads by port > names they're handling. > > Signed-off-by: Ilya Maximets <i.maximets@samsung.com> > --- > lib/dpif-netdev.c | 9 ++++++++- > 1 file changed, 8 insertions(+), 1 deletion(-) > > diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c > index d0a1c58ad..34ba03836 100644 > --- a/lib/dpif-netdev.c > +++ b/lib/dpif-netdev.c > @@ -4735,9 +4735,16 @@ reconfigure_pmd_threads(struct dp_netdev *dp) > FOR_EACH_CORE_ON_DUMP(core, pmd_cores) { > pmd = dp_netdev_get_pmd(dp, core->core_id); > if (!pmd) { > + struct ds name = DS_EMPTY_INITIALIZER; > + > pmd = xzalloc(sizeof *pmd); > dp_netdev_configure_pmd(pmd, dp, core->core_id, > core->numa_id); > - pmd->thread = ovs_thread_create("pmd", pmd_thread_main, > pmd); > + > + ds_put_format(&name, "pmd-c%02d/id:", core->core_id); This is a really good idea :) One remark should we make it %03d? > + pmd->thread = ovs_thread_create(ds_cstr(&name), > + pmd_thread_main, pmd); > + ds_destroy(&name); > + > VLOG_INFO("PMD thread on numa_id: %d, core id: %2d > created.", > pmd->numa_id, pmd->core_id); > changed = true; > -- > 2.17.1
On 13.08.2019 19:46, Eelco Chaudron wrote: > > > On 13 Aug 2019, at 18:37, Ilya Maximets wrote: > >> This is highly useful to see on which core PMD is running by >> only looking at the thread name. Thread Id still allows to >> distinguish different threads running on the same core over the time: >> >> |dpif_netdev(pmd-c10/id:53)|DBG|Creating 2. subtable <...> >> |dpif_netdev(pmd-c10/id:53)|DBG|flow_add: <...>, actions:2 >> |dpif_netdev(pmd-c09/id:70)|DBG|Core 9 processing port <..> >> >> In gdb, top or any other utility it's useful to quickly catch up >> needed thread without parsing logs, memory or matching threads by port >> names they're handling. >> >> Signed-off-by: Ilya Maximets <i.maximets@samsung.com> >> --- >> lib/dpif-netdev.c | 9 ++++++++- >> 1 file changed, 8 insertions(+), 1 deletion(-) >> >> diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c >> index d0a1c58ad..34ba03836 100644 >> --- a/lib/dpif-netdev.c >> +++ b/lib/dpif-netdev.c >> @@ -4735,9 +4735,16 @@ reconfigure_pmd_threads(struct dp_netdev *dp) >> FOR_EACH_CORE_ON_DUMP(core, pmd_cores) { >> pmd = dp_netdev_get_pmd(dp, core->core_id); >> if (!pmd) { >> + struct ds name = DS_EMPTY_INITIALIZER; >> + >> pmd = xzalloc(sizeof *pmd); >> dp_netdev_configure_pmd(pmd, dp, core->core_id, core->numa_id); >> - pmd->thread = ovs_thread_create("pmd", pmd_thread_main, pmd); >> + >> + ds_put_format(&name, "pmd-c%02d/id:", core->core_id); > > This is a really good idea :) One remark should we make it %03d? There is a hard limit for the thread name. It's 15 meaningful chars excluding the terminating null byte. 'pmd-c02/id:' is 11 bytes wide keeping 4 bytes for the thread id. 'pmd-c002/id:' is 12 bytes wide with only 3 bytes remaining for id. Thread ids could easily become big ( > 1000) for a long running process, that is why %02d was chosen, to save some space. BTW, even on a systems with 100+ CPUs I'm usually placing OVS threads on a lower cores. Anyway, this is only the matter of a bit more visual beauty in the logs. What do you think? > >> + pmd->thread = ovs_thread_create(ds_cstr(&name), >> + pmd_thread_main, pmd); >> + ds_destroy(&name); >> + >> VLOG_INFO("PMD thread on numa_id: %d, core id: %2d created.", >> pmd->numa_id, pmd->core_id); >> changed = true; >> -- >> 2.17.1 > >
On 14 Aug 2019, at 9:45, Ilya Maximets wrote: > On 13.08.2019 19:46, Eelco Chaudron wrote: >> >> >> On 13 Aug 2019, at 18:37, Ilya Maximets wrote: >> >>> This is highly useful to see on which core PMD is running by >>> only looking at the thread name. Thread Id still allows to >>> distinguish different threads running on the same core over the >>> time: >>> >>> |dpif_netdev(pmd-c10/id:53)|DBG|Creating 2. subtable <...> >>> |dpif_netdev(pmd-c10/id:53)|DBG|flow_add: <...>, actions:2 >>> |dpif_netdev(pmd-c09/id:70)|DBG|Core 9 processing port <..> >>> >>> In gdb, top or any other utility it's useful to quickly catch up >>> needed thread without parsing logs, memory or matching threads by >>> port >>> names they're handling. >>> >>> Signed-off-by: Ilya Maximets <i.maximets@samsung.com> >>> --- >>> lib/dpif-netdev.c | 9 ++++++++- >>> 1 file changed, 8 insertions(+), 1 deletion(-) >>> >>> diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c >>> index d0a1c58ad..34ba03836 100644 >>> --- a/lib/dpif-netdev.c >>> +++ b/lib/dpif-netdev.c >>> @@ -4735,9 +4735,16 @@ reconfigure_pmd_threads(struct dp_netdev *dp) >>> FOR_EACH_CORE_ON_DUMP(core, pmd_cores) { >>> pmd = dp_netdev_get_pmd(dp, core->core_id); >>> if (!pmd) { >>> + struct ds name = DS_EMPTY_INITIALIZER; >>> + >>> pmd = xzalloc(sizeof *pmd); >>> dp_netdev_configure_pmd(pmd, dp, >>> core->core_id, core->numa_id); >>> - pmd->thread = ovs_thread_create("pmd", >>> pmd_thread_main, pmd); >>> + >>> + ds_put_format(&name, "pmd-c%02d/id:", >>> core->core_id); >> >> This is a really good idea :) One remark should we make it %03d? > > There is a hard limit for the thread name. It's 15 meaningful chars > excluding the > terminating null byte. 'pmd-c02/id:' is 11 bytes wide keeping 4 bytes > for the > thread id. 'pmd-c002/id:' is 12 bytes wide with only 3 bytes remaining > for id. > Thread ids could easily become big ( > 1000) for a long running > process, that is > why %02d was chosen, to save some space. > > BTW, even on a systems with 100+ CPUs I'm usually placing OVS threads > on a lower > cores. Anyway, this is only the matter of a bit more visual beauty in > the logs. > > What do you think? Makes sense Acked-by: Eelco Chaudron <echaudro@redhat.com> > >> >>> + pmd->thread = >>> ovs_thread_create(ds_cstr(&name), >>> + >>> pmd_thread_main, pmd); >>> + ds_destroy(&name); >>> + >>> VLOG_INFO("PMD thread on numa_id: %d, core >>> id: %2d created.", >>> pmd->numa_id, >>> pmd->core_id); >>> changed = true; >>> -- >>> 2.17.1 >> >>
On Wed, Aug 14, 2019 at 9:45 AM Ilya Maximets <i.maximets@samsung.com> wrote: > > On 13.08.2019 19:46, Eelco Chaudron wrote: > > > > > > On 13 Aug 2019, at 18:37, Ilya Maximets wrote: > > > >> This is highly useful to see on which core PMD is running by > >> only looking at the thread name. Thread Id still allows to > >> distinguish different threads running on the same core over the time: > >> > >> |dpif_netdev(pmd-c10/id:53)|DBG|Creating 2. subtable <...> > >> |dpif_netdev(pmd-c10/id:53)|DBG|flow_add: <...>, actions:2 > >> |dpif_netdev(pmd-c09/id:70)|DBG|Core 9 processing port <..> > >> > >> In gdb, top or any other utility it's useful to quickly catch up > >> needed thread without parsing logs, memory or matching threads by port > >> names they're handling. > >> > >> Signed-off-by: Ilya Maximets <i.maximets@samsung.com> > >> --- > >> lib/dpif-netdev.c | 9 ++++++++- > >> 1 file changed, 8 insertions(+), 1 deletion(-) > >> > >> diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c > >> index d0a1c58ad..34ba03836 100644 > >> --- a/lib/dpif-netdev.c > >> +++ b/lib/dpif-netdev.c > >> @@ -4735,9 +4735,16 @@ reconfigure_pmd_threads(struct dp_netdev *dp) > >> FOR_EACH_CORE_ON_DUMP(core, pmd_cores) { > >> pmd = dp_netdev_get_pmd(dp, core->core_id); > >> if (!pmd) { > >> + struct ds name = DS_EMPTY_INITIALIZER; > >> + > >> pmd = xzalloc(sizeof *pmd); > >> dp_netdev_configure_pmd(pmd, dp, core->core_id, core->numa_id); > >> - pmd->thread = ovs_thread_create("pmd", pmd_thread_main, pmd); > >> + > >> + ds_put_format(&name, "pmd-c%02d/id:", core->core_id); > > > > This is a really good idea :) One remark should we make it %03d? > > There is a hard limit for the thread name. It's 15 meaningful chars excluding the > terminating null byte. 'pmd-c02/id:' is 11 bytes wide keeping 4 bytes for the > thread id. 'pmd-c002/id:' is 12 bytes wide with only 3 bytes remaining for id. > Thread ids could easily become big ( > 1000) for a long running process, that is > why %02d was chosen, to save some space. Do we really need the /id: part? A c%03d/pmd prefix would keep the existing pmd%d pattern and save 3 characters. Otherwise, this is a good idea yes.
On 20.08.2019 13:00, David Marchand wrote: > On Wed, Aug 14, 2019 at 9:45 AM Ilya Maximets <i.maximets@samsung.com> wrote: >> >> On 13.08.2019 19:46, Eelco Chaudron wrote: >>> >>> >>> On 13 Aug 2019, at 18:37, Ilya Maximets wrote: >>> >>>> This is highly useful to see on which core PMD is running by >>>> only looking at the thread name. Thread Id still allows to >>>> distinguish different threads running on the same core over the time: >>>> >>>> |dpif_netdev(pmd-c10/id:53)|DBG|Creating 2. subtable <...> >>>> |dpif_netdev(pmd-c10/id:53)|DBG|flow_add: <...>, actions:2 >>>> |dpif_netdev(pmd-c09/id:70)|DBG|Core 9 processing port <..> >>>> >>>> In gdb, top or any other utility it's useful to quickly catch up >>>> needed thread without parsing logs, memory or matching threads by port >>>> names they're handling. >>>> >>>> Signed-off-by: Ilya Maximets <i.maximets@samsung.com> >>>> --- >>>> lib/dpif-netdev.c | 9 ++++++++- >>>> 1 file changed, 8 insertions(+), 1 deletion(-) >>>> >>>> diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c >>>> index d0a1c58ad..34ba03836 100644 >>>> --- a/lib/dpif-netdev.c >>>> +++ b/lib/dpif-netdev.c >>>> @@ -4735,9 +4735,16 @@ reconfigure_pmd_threads(struct dp_netdev *dp) >>>> FOR_EACH_CORE_ON_DUMP(core, pmd_cores) { >>>> pmd = dp_netdev_get_pmd(dp, core->core_id); >>>> if (!pmd) { >>>> + struct ds name = DS_EMPTY_INITIALIZER; >>>> + >>>> pmd = xzalloc(sizeof *pmd); >>>> dp_netdev_configure_pmd(pmd, dp, core->core_id, core->numa_id); >>>> - pmd->thread = ovs_thread_create("pmd", pmd_thread_main, pmd); >>>> + >>>> + ds_put_format(&name, "pmd-c%02d/id:", core->core_id); >>> >>> This is a really good idea :) One remark should we make it %03d? >> >> There is a hard limit for the thread name. It's 15 meaningful chars excluding the >> terminating null byte. 'pmd-c02/id:' is 11 bytes wide keeping 4 bytes for the >> thread id. 'pmd-c002/id:' is 12 bytes wide with only 3 bytes remaining for id. >> Thread ids could easily become big ( > 1000) for a long running process, that is >> why %02d was chosen, to save some space. > > Do we really need the /id: part? > A c%03d/pmd prefix would keep the existing pmd%d pattern and save 3 characters. This will break lib/ovs-thread.c:thread_is_pmd() function. > > Otherwise, this is a good idea yes.
On Tue, Aug 20, 2019 at 12:06 PM Ilya Maximets <i.maximets@samsung.com> wrote: > > On 20.08.2019 13:00, David Marchand wrote: > > On Wed, Aug 14, 2019 at 9:45 AM Ilya Maximets <i.maximets@samsung.com> wrote: > >> > >> On 13.08.2019 19:46, Eelco Chaudron wrote: > >>> This is a really good idea :) One remark should we make it %03d? > >> > >> There is a hard limit for the thread name. It's 15 meaningful chars excluding the > >> terminating null byte. 'pmd-c02/id:' is 11 bytes wide keeping 4 bytes for the > >> thread id. 'pmd-c002/id:' is 12 bytes wide with only 3 bytes remaining for id. > >> Thread ids could easily become big ( > 1000) for a long running process, that is > >> why %02d was chosen, to save some space. > > > > Do we really need the /id: part? > > A c%03d/pmd prefix would keep the existing pmd%d pattern and save 3 characters. > > This will break lib/ovs-thread.c:thread_is_pmd() function. Ok, not really a problem to change from my pov. I don't really like the /id: in the thread name, as usually userspace thread names contains simple alphanumeric characters. But I can see no problem with this so: Reviewed-by: David Marchand <david.marchand@redhat.com> -- David Marchand
On 13.08.2019 19:37, Ilya Maximets wrote: > This is highly useful to see on which core PMD is running by > only looking at the thread name. Thread Id still allows to > distinguish different threads running on the same core over the time: > > |dpif_netdev(pmd-c10/id:53)|DBG|Creating 2. subtable <...> > |dpif_netdev(pmd-c10/id:53)|DBG|flow_add: <...>, actions:2 > |dpif_netdev(pmd-c09/id:70)|DBG|Core 9 processing port <..> > > In gdb, top or any other utility it's useful to quickly catch up > needed thread without parsing logs, memory or matching threads by port > names they're handling. > > Signed-off-by: Ilya Maximets <i.maximets@samsung.com> > --- Thanks, Eelco and David! Applied to master. Best regards, Ilya Maximets.
diff --git a/lib/dpif-netdev.c b/lib/dpif-netdev.c index d0a1c58ad..34ba03836 100644 --- a/lib/dpif-netdev.c +++ b/lib/dpif-netdev.c @@ -4735,9 +4735,16 @@ reconfigure_pmd_threads(struct dp_netdev *dp) FOR_EACH_CORE_ON_DUMP(core, pmd_cores) { pmd = dp_netdev_get_pmd(dp, core->core_id); if (!pmd) { + struct ds name = DS_EMPTY_INITIALIZER; + pmd = xzalloc(sizeof *pmd); dp_netdev_configure_pmd(pmd, dp, core->core_id, core->numa_id); - pmd->thread = ovs_thread_create("pmd", pmd_thread_main, pmd); + + ds_put_format(&name, "pmd-c%02d/id:", core->core_id); + pmd->thread = ovs_thread_create(ds_cstr(&name), + pmd_thread_main, pmd); + ds_destroy(&name); + VLOG_INFO("PMD thread on numa_id: %d, core id: %2d created.", pmd->numa_id, pmd->core_id); changed = true;
This is highly useful to see on which core PMD is running by only looking at the thread name. Thread Id still allows to distinguish different threads running on the same core over the time: |dpif_netdev(pmd-c10/id:53)|DBG|Creating 2. subtable <...> |dpif_netdev(pmd-c10/id:53)|DBG|flow_add: <...>, actions:2 |dpif_netdev(pmd-c09/id:70)|DBG|Core 9 processing port <..> In gdb, top or any other utility it's useful to quickly catch up needed thread without parsing logs, memory or matching threads by port names they're handling. Signed-off-by: Ilya Maximets <i.maximets@samsung.com> --- lib/dpif-netdev.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)