Re: [PATCH 1/5] ftrace: Support full glob matching
On Thu, 20 Oct 2016 09:43:20 -0400 Steven Rostedtwrote: > On Thu, 20 Oct 2016 11:07:52 +0900 > Masami Hiramatsu wrote: > > > > And OK, what kind of tests are failed on 4.9-rc1? I though a possible > > kernel freeze in 4.8 when I ran ftracetest, but it seemed disappeared > > in 4.9-rc1. > > The kernel crashed on enabling the mmiotracer, which disables CPUs. > That tracer is notorious for finding CPU hotplug bugs. There's a fix in > tip/urgent that I pulled in, which lets it pass. I rather wait till rc2 > to base changes on where I don't need to add commits to make my tests > pass. Ah, I think this should same as what I hit :) I've tested this series on tip/master, so it was fixed at that point. Thanks, -- Masami Hiramatsu
Re: [PATCH 1/5] ftrace: Support full glob matching
On Thu, 20 Oct 2016 09:43:20 -0400 Steven Rostedt wrote: > On Thu, 20 Oct 2016 11:07:52 +0900 > Masami Hiramatsu wrote: > > > > And OK, what kind of tests are failed on 4.9-rc1? I though a possible > > kernel freeze in 4.8 when I ran ftracetest, but it seemed disappeared > > in 4.9-rc1. > > The kernel crashed on enabling the mmiotracer, which disables CPUs. > That tracer is notorious for finding CPU hotplug bugs. There's a fix in > tip/urgent that I pulled in, which lets it pass. I rather wait till rc2 > to base changes on where I don't need to add commits to make my tests > pass. Ah, I think this should same as what I hit :) I've tested this series on tip/master, so it was fixed at that point. Thanks, -- Masami Hiramatsu
Re: [PATCH 1/5] ftrace: Support full glob matching
On Thu, 20 Oct 2016 11:07:52 +0900 Masami Hiramatsuwrote: > And OK, what kind of tests are failed on 4.9-rc1? I though a possible > kernel freeze in 4.8 when I ran ftracetest, but it seemed disappeared > in 4.9-rc1. The kernel crashed on enabling the mmiotracer, which disables CPUs. That tracer is notorious for finding CPU hotplug bugs. There's a fix in tip/urgent that I pulled in, which lets it pass. I rather wait till rc2 to base changes on where I don't need to add commits to make my tests pass. -- Steve
Re: [PATCH 1/5] ftrace: Support full glob matching
On Thu, 20 Oct 2016 11:07:52 +0900 Masami Hiramatsu wrote: > And OK, what kind of tests are failed on 4.9-rc1? I though a possible > kernel freeze in 4.8 when I ran ftracetest, but it seemed disappeared > in 4.9-rc1. The kernel crashed on enabling the mmiotracer, which disables CPUs. That tracer is notorious for finding CPU hotplug bugs. There's a fix in tip/urgent that I pulled in, which lets it pass. I rather wait till rc2 to base changes on where I don't need to add commits to make my tests pass. -- Steve
Re: [PATCH 1/5] ftrace: Support full glob matching
On Wed, 19 Oct 2016 09:20:57 -0400 Steven Rostedtwrote: > On Wed, 19 Oct 2016 13:53:36 +0900 > Masami Hiramatsu wrote: > > > Use glob_match() to support flexible glob wildcards (*,?) > > and character classes ([) for ftrace. > > Since the full glob matching is slower than the current > > partial matching routines(*pat, pat*, *pat*), this leaves > > those routines and just add MATCH_GLOB for complex glob > > expression. > > > > e.g. > > > > [root@localhost tracing]# echo 'sched*group' > set_ftrace_filter > > [root@localhost tracing]# cat set_ftrace_filter > > sched_free_group > > sched_change_group > > sched_create_group > > sched_online_group > > sched_destroy_group > > sched_offline_group > > [root@localhost tracing]# echo '[Ss]y[Ss]_*' > set_ftrace_filter > > [root@localhost tracing]# head set_ftrace_filter > > sys_arch_prctl > > sys_rt_sigreturn > > sys_ioperm > > SyS_iopl > > sys_modify_ldt > > SyS_mmap > > SyS_set_thread_area > > SyS_get_thread_area > > SyS_set_tid_address > > sys_fork > > > > Hi Masami, > > Is this any different than your last patch, as I already pulled it into > my queue. I haven't been able to test it much, as I found that 4.9-rc1 > is failing my tests, and I can't add new code to linux-next till 4.9 is > stable. This patch is completely same as the last patch. If you already queued it, please ignore this [1/5]. And OK, what kind of tests are failed on 4.9-rc1? I though a possible kernel freeze in 4.8 when I ran ftracetest, but it seemed disappeared in 4.9-rc1. Thanks, -- Masami Hiramatsu
Re: [PATCH 1/5] ftrace: Support full glob matching
On Wed, 19 Oct 2016 09:20:57 -0400 Steven Rostedt wrote: > On Wed, 19 Oct 2016 13:53:36 +0900 > Masami Hiramatsu wrote: > > > Use glob_match() to support flexible glob wildcards (*,?) > > and character classes ([) for ftrace. > > Since the full glob matching is slower than the current > > partial matching routines(*pat, pat*, *pat*), this leaves > > those routines and just add MATCH_GLOB for complex glob > > expression. > > > > e.g. > > > > [root@localhost tracing]# echo 'sched*group' > set_ftrace_filter > > [root@localhost tracing]# cat set_ftrace_filter > > sched_free_group > > sched_change_group > > sched_create_group > > sched_online_group > > sched_destroy_group > > sched_offline_group > > [root@localhost tracing]# echo '[Ss]y[Ss]_*' > set_ftrace_filter > > [root@localhost tracing]# head set_ftrace_filter > > sys_arch_prctl > > sys_rt_sigreturn > > sys_ioperm > > SyS_iopl > > sys_modify_ldt > > SyS_mmap > > SyS_set_thread_area > > SyS_get_thread_area > > SyS_set_tid_address > > sys_fork > > > > Hi Masami, > > Is this any different than your last patch, as I already pulled it into > my queue. I haven't been able to test it much, as I found that 4.9-rc1 > is failing my tests, and I can't add new code to linux-next till 4.9 is > stable. This patch is completely same as the last patch. If you already queued it, please ignore this [1/5]. And OK, what kind of tests are failed on 4.9-rc1? I though a possible kernel freeze in 4.8 when I ran ftracetest, but it seemed disappeared in 4.9-rc1. Thanks, -- Masami Hiramatsu
Re: [PATCH 1/5] ftrace: Support full glob matching
On Wed, 19 Oct 2016 13:53:36 +0900 Masami Hiramatsuwrote: > Use glob_match() to support flexible glob wildcards (*,?) > and character classes ([) for ftrace. > Since the full glob matching is slower than the current > partial matching routines(*pat, pat*, *pat*), this leaves > those routines and just add MATCH_GLOB for complex glob > expression. > > e.g. > > [root@localhost tracing]# echo 'sched*group' > set_ftrace_filter > [root@localhost tracing]# cat set_ftrace_filter > sched_free_group > sched_change_group > sched_create_group > sched_online_group > sched_destroy_group > sched_offline_group > [root@localhost tracing]# echo '[Ss]y[Ss]_*' > set_ftrace_filter > [root@localhost tracing]# head set_ftrace_filter > sys_arch_prctl > sys_rt_sigreturn > sys_ioperm > SyS_iopl > sys_modify_ldt > SyS_mmap > SyS_set_thread_area > SyS_get_thread_area > SyS_set_tid_address > sys_fork > Hi Masami, Is this any different than your last patch, as I already pulled it into my queue. I haven't been able to test it much, as I found that 4.9-rc1 is failing my tests, and I can't add new code to linux-next till 4.9 is stable. -- Steve
Re: [PATCH 1/5] ftrace: Support full glob matching
On Wed, 19 Oct 2016 13:53:36 +0900 Masami Hiramatsu wrote: > Use glob_match() to support flexible glob wildcards (*,?) > and character classes ([) for ftrace. > Since the full glob matching is slower than the current > partial matching routines(*pat, pat*, *pat*), this leaves > those routines and just add MATCH_GLOB for complex glob > expression. > > e.g. > > [root@localhost tracing]# echo 'sched*group' > set_ftrace_filter > [root@localhost tracing]# cat set_ftrace_filter > sched_free_group > sched_change_group > sched_create_group > sched_online_group > sched_destroy_group > sched_offline_group > [root@localhost tracing]# echo '[Ss]y[Ss]_*' > set_ftrace_filter > [root@localhost tracing]# head set_ftrace_filter > sys_arch_prctl > sys_rt_sigreturn > sys_ioperm > SyS_iopl > sys_modify_ldt > SyS_mmap > SyS_set_thread_area > SyS_get_thread_area > SyS_set_tid_address > sys_fork > Hi Masami, Is this any different than your last patch, as I already pulled it into my queue. I haven't been able to test it much, as I found that 4.9-rc1 is failing my tests, and I can't add new code to linux-next till 4.9 is stable. -- Steve