On 11/4/21 09:39, Ard Biesheuvel wrote:
> On Thu, 4 Nov 2021 at 09:31, Helge Deller <deller@xxxxxx> wrote:
>>
>> On 11/4/21 09:26, Ard Biesheuvel wrote:
>>> In commit 2214c0e77259 ("parisc: Move thread_info into task struct")
>>> PA-RISC gained support for THREAD_INFO_IN_TASK while changes were
>>> already underway to keep the CPU field in thread_info rather than move
>>> it into task_struct when THREAD_INFO_IN_TASK is enabled. The result is a
>>> broken build for all PA-RISC configs that enable SMP.
>>>
>>> So let's partially revert that commit, and get rid of the ugly hack to
>>> get at the offset of task_struct::cpu without having to include
>>> linux/sched.h, and put the CPU field back where it was before.
>>>
>>> Reported-by: Guenter Roeck <linux@xxxxxxxxxxxx>
>>> Fixes: bcf9033e5449 ("sched: move CPU field back into thread_info if THREAD_INFO_IN_TASK=y")
>>> Signed-off-by: Ard Biesheuvel <ardb@xxxxxxxxxx>
>>
>> Thanks Ard!
>>
>> I actually had the same patch finished right now too :-)
>>
>
> I was wondering about that :-)
>
>> One small nid though. See below...
>>
>>
>>> ---
>>> arch/parisc/include/asm/smp.h | 19 ++-----------------
>>> arch/parisc/include/asm/thread_info.h | 2 ++
>>> arch/parisc/kernel/asm-offsets.c | 4 +---
>>> arch/parisc/kernel/smp.c | 2 +-
>>> 4 files changed, 6 insertions(+), 21 deletions(-)
>>>
>>> diff --git a/arch/parisc/include/asm/smp.h b/arch/parisc/include/asm/smp.h
>>> index 16d41127500e..cba55abf7bac 100644
>>> --- a/arch/parisc/include/asm/smp.h
>>> +++ b/arch/parisc/include/asm/smp.h
>>> @@ -34,23 +34,8 @@ extern void arch_send_call_function_ipi_mask(const struct cpumask *mask);
>>>
>>> #endif /* !ASSEMBLY */
>>>
>>> -/*
>>> - * This is particularly ugly: it appears we can't actually get the definition
>>> - * of task_struct here, but we need access to the CPU this task is running on.
>>> - * Instead of using task_struct we're using TASK_CPU which is extracted from
>>> - * asm-offsets.h by kbuild to get the current processor ID.
>>> - *
>>> - * This also needs to be safeguarded when building asm-offsets.s because at
>>> - * that time TASK_CPU is not defined yet. It could have been guarded by
>>> - * TASK_CPU itself, but we want the build to fail if TASK_CPU is missing
>>> - * when building something else than asm-offsets.s
>>> - */
>>> -#ifdef GENERATING_ASM_OFFSETS
>>> -#define raw_smp_processor_id() (0)
>>> -#else
>>> -#include <asm/asm-offsets.h>
>>> -#define raw_smp_processor_id() (*(unsigned int *)((void *)current + TASK_CPU))
>>> -#endif
>>> +#define raw_smp_processor_id() (current_thread_info()->cpu)
>>> +
>>> #else /* CONFIG_SMP */
>>>
>>> static inline void smp_send_all_nop(void) { return; }
>>> diff --git a/arch/parisc/include/asm/thread_info.h b/arch/parisc/include/asm/thread_info.h
>>> index 75657c2c54e1..d6268834bfa5 100644
>>> --- a/arch/parisc/include/asm/thread_info.h
>>> +++ b/arch/parisc/include/asm/thread_info.h
>>> @@ -8,12 +8,14 @@
>>>
>>> struct thread_info {
>>> unsigned long flags; /* thread_info flags (see TIF_*) */
>>> + __u32 cpu; /* current CPU */
>>> int preempt_count; /* 0=premptable, <0=BUG; will also serve as bh-counter */
>>> };
>>>
>>> #define INIT_THREAD_INFO(tsk) \
>>> { \
>>> .flags = 0, \
>>> + .cpu = 0, \
>>> .preempt_count = INIT_PREEMPT_COUNT, \
>>> }
>>>
>>> diff --git a/arch/parisc/kernel/asm-offsets.c b/arch/parisc/kernel/asm-offsets.c
>>> index e35154035441..629d38e36e22 100644
>>> --- a/arch/parisc/kernel/asm-offsets.c
>>> +++ b/arch/parisc/kernel/asm-offsets.c
>>> @@ -14,8 +14,6 @@
>>> * Copyright (C) 2003 James Bottomley <jejb at parisc-linux.org>
>>> */
>>>
>>> -#define GENERATING_ASM_OFFSETS /* asm/smp.h */
>>> -
>>> #include <linux/types.h>
>>> #include <linux/sched.h>
>>> #include <linux/thread_info.h>
>>> @@ -40,7 +38,7 @@ int main(void)
>>> DEFINE(TASK_TI_FLAGS, offsetof(struct task_struct, thread_info.flags));
>>> DEFINE(TASK_STACK, offsetof(struct task_struct, stack));
>>> #ifdef CONFIG_SMP
>>> - DEFINE(TASK_CPU, offsetof(struct task_struct, cpu));
>>> + DEFINE(TASK_CPU, offsetof(struct task_struct, thread_info.cpu));
>>> #endif
>>
>> The TASK_CPU define isn't needed any longer.
>> So, the whole part inside #ifdef..#endif can go.
>>
>
> OK let's just drop it then.
Ok
>
>>> BLANK();
>>> DEFINE(TASK_REGS, offsetof(struct task_struct, thread.regs));
>>> diff --git a/arch/parisc/kernel/smp.c b/arch/parisc/kernel/smp.c
>>> index 171925285f3e..0cd97fa004c5 100644
>>> --- a/arch/parisc/kernel/smp.c
>>> +++ b/arch/parisc/kernel/smp.c
>>> @@ -339,7 +339,7 @@ int smp_boot_one_cpu(int cpuid, struct task_struct *idle)
>>> const struct cpuinfo_parisc *p = &per_cpu(cpu_data, cpuid);
>>> long timeout;
>>>
>>> - idle->cpu = cpuid;
>>> + task_thread_info(idle)->cpu = cpuid;
>
> I was wondering if this could simply be dropped altogether? I am
> pretty sure it is redundant, as the core code sets this field as well,
> but I don't have access to a SMP PA-RISC machine to test it.
Good point.
I tested and it can be dropped.
>>> /* Let _start know what logical CPU we're booting
>>> ** (offset into init_tasks[],cpu_data[])
>>>
>
>> If it's Ok for you I'll clean up your patch and submit
>> all with the next pull request to Linus later today.
>>
>
> Whatever works for you is fine with me.
Ok. I've cleaned it up, applied it to my for-next tree and will push later today to Linus
with other patches. Here is the current version:
https://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux.git/commit/?h=for-next
Thanks!
Helge
