[PATCH 1/2] arm64: Move handling of erratum 1418040 into C code

[Marc Zyngier]

Hi Stephen,

On 2020-07-28 21:04, Stephen Boyd wrote:
> Quoting Marc Zyngier (2020-07-28 02:21:11)
>> diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
>> index 35de8ba60e3d..44445d471442 100644
>> --- a/arch/arm64/kernel/entry.S
>> +++ b/arch/arm64/kernel/entry.S
>> @@ -169,19 +169,6 @@ alternative_cb_end
>>         stp     x28, x29, [sp, #16 * 14]
>> 
>>         .if     \el == 0
>> -       .if     \regsize == 32
>> -       /*
>> -        * If we're returning from a 32-bit task on a system affected 
>> by
>> -        * 1418040 then re-enable userspace access to the virtual 
>> counter.
>> -        */
> 
> Can this comment go above the function in C?

It will have to be a slightly different comment, as the logic changes.
Something like:

/*
  * ARM erratum 1418040 handling affecting the 32bit view of CNTVCT.
  *
  * Assuming the virtual counter is enabled at the beginning of times:
  * - disable access when switching from a 64bit task to a 32bit task
  * - enable access when switching from a 32bit task to a 64bit task
  */

> 
>> -#ifdef CONFIG_ARM64_ERRATUM_1418040
>> -alternative_if ARM64_WORKAROUND_1418040
>> -       mrs     x0, cntkctl_el1
>> -       orr     x0, x0, #2      // ARCH_TIMER_USR_VCT_ACCESS_EN
>> -       msr     cntkctl_el1, x0
>> -alternative_else_nop_endif
>> -#endif
>> -       .endif
>>         clear_gp_regs
>>         mrs     x21, sp_el0
>>         ldr_this_cpu    tsk, __entry_task, x20
>> @@ -337,14 +324,6 @@ alternative_else_nop_endif
>>         tst     x22, #PSR_MODE32_BIT            // native task?
>>         b.eq    3f
>> 
>> -#ifdef CONFIG_ARM64_ERRATUM_1418040
>> -alternative_if ARM64_WORKAROUND_1418040
>> -       mrs     x0, cntkctl_el1
>> -       bic     x0, x0, #2                      // 
>> ARCH_TIMER_USR_VCT_ACCESS_EN
>> -       msr     cntkctl_el1, x0
>> -alternative_else_nop_endif
>> -#endif
>> -
>>  #ifdef CONFIG_ARM64_ERRATUM_845719
>>  alternative_if ARM64_WORKAROUND_845719
>>  #ifdef CONFIG_PID_IN_CONTEXTIDR
>> diff --git a/arch/arm64/kernel/process.c b/arch/arm64/kernel/process.c
>> index 6089638c7d43..87c33f7c536b 100644
>> --- a/arch/arm64/kernel/process.c
>> +++ b/arch/arm64/kernel/process.c
>> @@ -515,6 +515,34 @@ static void entry_task_switch(struct task_struct 
>> *next)
>>         __this_cpu_write(__entry_task, next);
>>  }
>> 
>> +static void erratum_1418040_thread_switch(struct task_struct *prev,
> 
> Should it be marked __always_inline so that the cpus_have_const_cap()
> check can avoid the branch to this function when there's nothing to do?

I'd expect the compiler to directly inline this (in my experience it
does, as with most of the other functions called from __switch_to()).

It probably doesn't hurt though, so I'll do that.

Thanks,

         M.
-- 
Jazz is not dead. It just smells funny...