Re: [PATCH v7 02/16] arm64: Use update{,_tsk}_thread_flag()

[Dave Martin <Dave.Martin@arm.com>]

On Wed, May 09, 2018 at 05:55:51PM +0100, Marc Zyngier wrote:
> On 09/05/18 17:27, Dave Martin wrote:
> > On Wed, May 09, 2018 at 05:17:28PM +0100, Will Deacon wrote:
> >> Hi Dave,
> >>
> >> On Wed, May 09, 2018 at 05:12:51PM +0100, Dave Martin wrote:
> >>> This patch uses the new update_thread_flag() helpers to simplify a
> >>> couple of if () set; else clear; constructs.
> >>>
> >>> No functional change.
> >>>
> >>> Signed-off-by: Dave Martin <Dave.Martin@arm.com>
> >>> Acked-by: Marc Zyngier <marc.zyngier@arm.com>
> >>> Cc: Catalin Marinas <catalin.marinas@arm.com>
> >>> Cc: Will Deacon <will.deacon@arm.com>
> >>> ---
> >>>  arch/arm64/kernel/fpsimd.c | 19 +++++++------------
> >>>  1 file changed, 7 insertions(+), 12 deletions(-)
> >>>
> >>> diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
> >>> index 87a3536..0c4e7e0 100644
> >>> --- a/arch/arm64/kernel/fpsimd.c
> >>> +++ b/arch/arm64/kernel/fpsimd.c
> > 
> > [...]
> > 
> >>> @@ -902,7 +900,7 @@ void fpsimd_thread_switch(struct task_struct *next)
> >>>  	if (current->mm)
> >>>  		task_fpsimd_save();
> >>>  
> >>> -	if (next->mm) {
> >>> +	if (next->mm)
> >>>  		/*
> >>>  		 * If we are switching to a task whose most recent userland
> >>>  		 * FPSIMD state is already in the registers of *this* cpu,
> >>> @@ -910,13 +908,10 @@ void fpsimd_thread_switch(struct task_struct *next)
> >>>  		 * the TIF_FOREIGN_FPSTATE flag so the state will be loaded
> >>>  		 * upon the next return to userland.
> >>>  		 */
> >>> -		if (__this_cpu_read(fpsimd_last_state.st) ==
> >>> -			&next->thread.uw.fpsimd_state
> >>> -		    && next->thread.fpsimd_cpu == smp_processor_id())
> >>> -			clear_tsk_thread_flag(next, TIF_FOREIGN_FPSTATE);
> >>> -		else
> >>> -			set_tsk_thread_flag(next, TIF_FOREIGN_FPSTATE);
> >>> -	}
> >>> +		update_tsk_thread_flag(next, TIF_FOREIGN_FPSTATE,
> >>> +			__this_cpu_read(fpsimd_last_state.st) !=
> >>> +				&next->thread.uw.fpsimd_state ||
> >>> +			next->thread.fpsimd_cpu != smp_processor_id());
> >>
> >> Given the multi-line comment and this multi-line call, I'd be inclined to
> >> leave the curlies in place and then use a local bool for the complex
> >> condition.
> > 
> > Hey, curlies cost money, you know.
> > 
> >> With that:
> >>
> >> Acked-by: Will Deacon <will.deacon@arm.com>
> > 
> > Are you content to see this merged without the change?  It doesn't seem
> > worth a respin of the whole series just for this.  I agree the code
> > would be clearer, but this patch doesn't actually make it worse IMHO.
> > 
> > If I respin for some reason though, I can address this and add your Ack.
> 
> I'm happy to perform the change myself when applying the series if there
> is no additional comment that would lead to a respin.

OK, feel free to apply the following if it looks good for you.

Cheers
---Dave

--8<--

diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index 32898a8..b0d29b7 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c
@@ -906,7 +906,7 @@ void fpsimd_thread_switch(struct task_struct *next)
 	if (current->mm)
 		fpsimd_save();
 
-	if (next->mm)
+	if (next->mm) {
 		/*
 		 * If we are switching to a task whose most recent userland
 		 * FPSIMD state is already in the registers of *this* cpu,
@@ -914,10 +914,13 @@ void fpsimd_thread_switch(struct task_struct *next)
 		 * the TIF_FOREIGN_FPSTATE flag so the state will be loaded
 		 * upon the next return to userland.
 		 */
+		bool wrong_task = __this_cpu_read(fpsimd_last_state.st) !=
+					&next->thread.uw.fpsimd_state;
+		bool wrong_cpu = next->thread.fpsimd_cpu != smp_processor_id();
+
 		update_tsk_thread_flag(next, TIF_FOREIGN_FPSTATE,
-			__this_cpu_read(fpsimd_last_state.st) !=
-				&next->thread.uw.fpsimd_state ||
-			next->thread.fpsimd_cpu != smp_processor_id());
+				       wrong_task || wrong_cpu);
+	}
 }
 
 void fpsimd_flush_thread(void)