| Message ID | 20241226191238.2559880-1-robert_joslyn@selinc.com |
|---|---|
| State | New |
| Headers | show |
| Series | powerpc: Allow scheduling of IRQ with PREEMPT_RT | expand |
On 2024-12-26 11:12:38 [-0800], Robert Joslyn wrote: > This branch is needed to allow preemptible IRQs to be scheduled, > otherwise the kernel hangs during boot on my e300 based platforms. > This branch is needed for both CONFIG_PREEMPT and CONFIG_PREEMPT_RT, > so broaden the check to CONFIG_PREEMPTION, which is selected by both. > > This was previously set to CONFIG_PREEMPTION, but was reverted as part > of removing LAZY_PREEMPT support. I remember I pointed this out while Shrikanth posted the preempt-lazy patches but didn't realize that I need this in RT tree, too. The lazy-preempt patches for powerpc were merged upstream, I am going to pull them into the v6.6-series. I am going to Cc you, hopefully you can confirm that it is working. > Fixes: b33022a48c79 > Signed-off-by: Robert Joslyn <robert_joslyn@selinc.com> > --- > arch/powerpc/kernel/interrupt.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/powerpc/kernel/interrupt.c b/arch/powerpc/kernel/interrupt.c > index af62ec974b97..970039af8e92 100644 > --- a/arch/powerpc/kernel/interrupt.c > +++ b/arch/powerpc/kernel/interrupt.c > @@ -396,7 +396,7 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs) > /* Returning to a kernel context with local irqs enabled. */ > WARN_ON_ONCE(!(regs->msr & MSR_EE)); > again: > - if (IS_ENABLED(CONFIG_PREEMPT)) { > + if (IS_ENABLED(CONFIG_PREEMPTION)) { > /* Return to preemptible kernel context */ > if (unlikely(read_thread_flags() & _TIF_NEED_RESCHED)) { > if (preempt_count() == 0) Sebastian
On 2025-01-07 12:56:08 [+0100], To Robert Joslyn wrote: > On 2024-12-26 11:12:38 [-0800], Robert Joslyn wrote: > > This branch is needed to allow preemptible IRQs to be scheduled, > > otherwise the kernel hangs during boot on my e300 based platforms. > > This branch is needed for both CONFIG_PREEMPT and CONFIG_PREEMPT_RT, > > so broaden the check to CONFIG_PREEMPTION, which is selected by both. > > > > This was previously set to CONFIG_PREEMPTION, but was reverted as part > > of removing LAZY_PREEMPT support. > > I remember I pointed this out while Shrikanth posted the preempt-lazy > patches but didn't realize that I need this in RT tree, too. > > The lazy-preempt patches for powerpc were merged upstream, I am going to > pull them into the v6.6-series. > I am going to Cc you, hopefully you can confirm that it is working. Robert, you are referring to v6.6-RT, correct? Sebastian
On 1/7/25 17:26, Sebastian Andrzej Siewior wrote: > On 2024-12-26 11:12:38 [-0800], Robert Joslyn wrote: >> This branch is needed to allow preemptible IRQs to be scheduled, >> otherwise the kernel hangs during boot on my e300 based platforms. >> This branch is needed for both CONFIG_PREEMPT and CONFIG_PREEMPT_RT, Hi Robert. Are you running with preempt=full or enabled preempt_rt ? rt is yet to be supported on powerpc. are you having it enabled out of tree? Also, I have sent out patch to enable dynamic preemption. It uses preempt_model_preemptible. I think it should work for you as well. If you could give that a try and ensure it works, it would be great. https://lore.kernel.org/all/20250106051919.55020-2-sshegde@linux.ibm.com/ >> so broaden the check to CONFIG_PREEMPTION, which is selected by both. >> >> This was previously set to CONFIG_PREEMPTION, but was reverted as part >> of removing LAZY_PREEMPT support. > > I remember I pointed this out while Shrikanth posted the preempt-lazy > patches but didn't realize that I need this in RT tree, too. > > The lazy-preempt patches for powerpc were merged upstream, I am going to > pull them into the v6.6-series. > I am going to Cc you, hopefully you can confirm that it is working. > >> Fixes: b33022a48c79 >> Signed-off-by: Robert Joslyn <robert_joslyn@selinc.com> >> --- >> arch/powerpc/kernel/interrupt.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/arch/powerpc/kernel/interrupt.c b/arch/powerpc/kernel/interrupt.c >> index af62ec974b97..970039af8e92 100644 >> --- a/arch/powerpc/kernel/interrupt.c >> +++ b/arch/powerpc/kernel/interrupt.c >> @@ -396,7 +396,7 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs) >> /* Returning to a kernel context with local irqs enabled. */ >> WARN_ON_ONCE(!(regs->msr & MSR_EE)); >> again: >> - if (IS_ENABLED(CONFIG_PREEMPT)) { >> + if (IS_ENABLED(CONFIG_PREEMPTION)) { >> /* Return to preemptible kernel context */ >> if (unlikely(read_thread_flags() & _TIF_NEED_RESCHED)) { >> if (preempt_count() == 0) > > Sebastian
> -----Original Message----- > From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> > Sent: Tuesday, January 7, 2025 3:59 AM > To: Robert Joslyn <robert_joslyn@selinc.com> > Cc: linux-kernel@vger.kernel.org; linux-rt-users@vger.kernel.org; > tglx@linutronix.de; rostedt@goodmis.org; Shrikanth Hegde > <sshegde@linux.ibm.com> > Subject: Re: [PATCH] powerpc: Allow scheduling of IRQ with PREEMPT_RT > > [Caution - External] > > On 2025-01-07 12:56:08 [+0100], To Robert Joslyn wrote: > > On 2024-12-26 11:12:38 [-0800], Robert Joslyn wrote: > > > This branch is needed to allow preemptible IRQs to be scheduled, > > > otherwise the kernel hangs during boot on my e300 based platforms. > > > This branch is needed for both CONFIG_PREEMPT and > CONFIG_PREEMPT_RT, > > > so broaden the check to CONFIG_PREEMPTION, which is selected by both. > > > > > > This was previously set to CONFIG_PREEMPTION, but was reverted as > > > part of removing LAZY_PREEMPT support. > > > > I remember I pointed this out while Shrikanth posted the preempt-lazy > > patches but didn't realize that I need this in RT tree, too. > > > > The lazy-preempt patches for powerpc were merged upstream, I am going > > to pull them into the v6.6-series. > > I am going to Cc you, hopefully you can confirm that it is working. > > Robert, you are referring to v6.6-RT, correct? > > Sebastian Yes, I'm using 6.6.58-rt45 at the moment. We found this regression while trying to update from 5.10-rt. I haven't played with lazy-preempt yet but I can certainly test the patches, presently we run everything with PREEMPT_RT. Thanks, Robert
> -----Original Message----- > From: Shrikanth Hegde <sshegde@linux.ibm.com> > Sent: Tuesday, January 7, 2025 5:10 AM > To: Robert Joslyn <robert_joslyn@selinc.com> > Cc: linux-kernel@vger.kernel.org; linux-rt-users@vger.kernel.org; > tglx@linutronix.de; rostedt@goodmis.org; Sebastian Andrzej Siewior > <bigeasy@linutronix.de> > Subject: Re: [PATCH] powerpc: Allow scheduling of IRQ with PREEMPT_RT > > [Caution - External] > > On 1/7/25 17:26, Sebastian Andrzej Siewior wrote: > > On 2024-12-26 11:12:38 [-0800], Robert Joslyn wrote: > >> This branch is needed to allow preemptible IRQs to be scheduled, > >> otherwise the kernel hangs during boot on my e300 based platforms. > >> This branch is needed for both CONFIG_PREEMPT and > CONFIG_PREEMPT_RT, > > Hi Robert. > > Are you running with preempt=full or enabled preempt_rt ? > rt is yet to be supported on powerpc. are you having it enabled out of tree? I'm running PREEMPT_RT using the out of tree patches on 6.6.58-rt45: https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-stable-rt.git/ I haven't tried using the mainline kernel, since as you said, it doesn't have powerpc support for PREEMPT_RT yet. > > Also, I have sent out patch to enable dynamic preemption. It uses > preempt_model_preemptible. > I think it should work for you as well. If you could give that a try and ensure it > works, it would be great. > https://urldefense.com/v3/__https://lore.kernel.org/all/20250106051919.5 > 5020-2- > sshegde@linux.ibm.com/__;!!O7uE89YCNVw!IDeomstGeuiEZi9FmLHH8C2xo > PPo_ewL- > k2kiQ9CmVID35U4J_v3Dkafjo3RnS9ygPoUZWwyyp91SP2euRkKArk$ Does the dynamic preemption support let me enable the same functionality as PREEMPT_RT with the out of tree patches? I can test it out either way. Thanks, Robert
On 1/7/25 23:43, Robert Joslyn wrote: >> -----Original Message----- >> From: Shrikanth Hegde <sshegde@linux.ibm.com> >> Sent: Tuesday, January 7, 2025 5:10 AM >> To: Robert Joslyn <robert_joslyn@selinc.com> >> Cc: linux-kernel@vger.kernel.org; linux-rt-users@vger.kernel.org; >> tglx@linutronix.de; rostedt@goodmis.org; Sebastian Andrzej Siewior >> <bigeasy@linutronix.de> >> Subject: Re: [PATCH] powerpc: Allow scheduling of IRQ with PREEMPT_RT >> >> [Caution - External] >> >> On 1/7/25 17:26, Sebastian Andrzej Siewior wrote: >>> On 2024-12-26 11:12:38 [-0800], Robert Joslyn wrote: >>>> This branch is needed to allow preemptible IRQs to be scheduled, >>>> otherwise the kernel hangs during boot on my e300 based platforms. >>>> This branch is needed for both CONFIG_PREEMPT and >> CONFIG_PREEMPT_RT, >> >> Hi Robert. >> >> Are you running with preempt=full or enabled preempt_rt ? >> rt is yet to be supported on powerpc. are you having it enabled out of tree? > > I'm running PREEMPT_RT using the out of tree patches on 6.6.58-rt45: > https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-stable-rt.git/ > > I haven't tried using the mainline kernel, since as you said, it doesn't have powerpc support for PREEMPT_RT yet. > >> >> Also, I have sent out patch to enable dynamic preemption. It uses >> preempt_model_preemptible. >> I think it should work for you as well. If you could give that a try and ensure it >> works, it would be great. >> https://urldefense.com/v3/__https://lore.kernel.org/all/20250106051919.5 >> 5020-2- >> sshegde@linux.ibm.com/__;!!O7uE89YCNVw!IDeomstGeuiEZi9FmLHH8C2xo >> PPo_ewL- >> k2kiQ9CmVID35U4J_v3Dkafjo3RnS9ygPoUZWwyyp91SP2euRkKArk$ > > Does the dynamic preemption support let me enable the same functionality as PREEMPT_RT with the out of tree patches? I can test it out either way. > with preempt_rt, dynamic preemption allows you to switch between full and lazy. without preempt_rt, you can switch between none, voluntary,full and lazy. > Thanks, > Robert
On 2025-01-07 18:09:46 [+0000], Robert Joslyn wrote: > > Yes, I'm using 6.6.58-rt45 at the moment. We found this regression > while trying to update from 5.10-rt. I haven't played with > lazy-preempt yet but I can certainly test the patches, presently we > run everything with PREEMPT_RT. So I released v6.12.8-rt8 and v6.13-rc6-rt3. Looking at the lazy bits I think I just forward your patch to the stable queue for v6.6. > Thanks, > Robert Sebastian
> -----Original Message----- > From: Sebastian Andrzej Siewior <bigeasy@linutronix.de> > Sent: Thursday, January 9, 2025 3:44 AM > To: Robert Joslyn <robert_joslyn@selinc.com> > Cc: linux-kernel@vger.kernel.org; linux-rt-users@vger.kernel.org; > tglx@linutronix.de; rostedt@goodmis.org; Shrikanth Hegde > <sshegde@linux.ibm.com> > Subject: Re: RE: [PATCH] powerpc: Allow scheduling of IRQ with PREEMPT_RT > > [Caution - External] > > On 2025-01-07 18:09:46 [+0000], Robert Joslyn wrote: > > > > Yes, I'm using 6.6.58-rt45 at the moment. We found this regression > > while trying to update from 5.10-rt. I haven't played with > > lazy-preempt yet but I can certainly test the patches, presently we > > run everything with PREEMPT_RT. > > So I released v6.12.8-rt8 and v6.13-rc6-rt3. Looking at the lazy bits I think I > just forward your patch to the stable queue for v6.6. I did some basic testing on 6.12.8-rt8 and that worked fine with both preempt_lazy and preempt_rt. For what it's worth, we've been using the patch for 6.6-rt for about a month now in internal testing and haven't seen any other issues. Thanks! Robert
diff --git a/arch/powerpc/kernel/interrupt.c b/arch/powerpc/kernel/interrupt.c index af62ec974b97..970039af8e92 100644 --- a/arch/powerpc/kernel/interrupt.c +++ b/arch/powerpc/kernel/interrupt.c @@ -396,7 +396,7 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs) /* Returning to a kernel context with local irqs enabled. */ WARN_ON_ONCE(!(regs->msr & MSR_EE)); again: - if (IS_ENABLED(CONFIG_PREEMPT)) { + if (IS_ENABLED(CONFIG_PREEMPTION)) { /* Return to preemptible kernel context */ if (unlikely(read_thread_flags() & _TIF_NEED_RESCHED)) { if (preempt_count() == 0)
This branch is needed to allow preemptible IRQs to be scheduled, otherwise the kernel hangs during boot on my e300 based platforms. This branch is needed for both CONFIG_PREEMPT and CONFIG_PREEMPT_RT, so broaden the check to CONFIG_PREEMPTION, which is selected by both. This was previously set to CONFIG_PREEMPTION, but was reverted as part of removing LAZY_PREEMPT support. Fixes: b33022a48c79 Signed-off-by: Robert Joslyn <robert_joslyn@selinc.com> --- arch/powerpc/kernel/interrupt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)