Re: [PATCH 37/49] x86/alternatives: Move text_poke_array completion from smp_text_poke_batch_finish() and smp_text_poke_batch_flush() to smp_text_poke_batch_process()

[Nikolay Borisov]

On 3.04.25 г. 17:13 ч., Ingo Molnar wrote:
> * Nikolay Borisov <nik.borisov@xxxxxxxx> wrote:
>
> > On 3.04.25 г. 16:38 ч., Ingo Molnar wrote:
> > > * Nikolay Borisov <nik.borisov@xxxxxxxx> wrote:
> > >
> > > > On 28.03.25 г. 15:26 ч., Ingo Molnar wrote:
> > > > > Simplifies the code and improves code generation a bit:
> > > > >
> > > > >       text	   data	    bss	    dec	    hex	filename
> > > > >      14769	   1017	   4112	  19898	   4dba	alternative.o.before
> > > > >      14742	   1017	   4112	  19871	   4d9f	alternative.o.after
> > > > >
> > > > > Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
> > > > > ---
> > > > >     arch/x86/kernel/alternative.c | 11 +++++------
> > > > >     1 file changed, 5 insertions(+), 6 deletions(-)
> > > > >
> > > > > diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
> > > > > index 1df8fac6740d..5293929488f0 100644
> > > > > --- a/arch/x86/kernel/alternative.c
> > > > > +++ b/arch/x86/kernel/alternative.c
> > > > > @@ -2750,6 +2750,9 @@ static void smp_text_poke_batch_process(void)
> > > > >     		if (unlikely(!atomic_dec_and_test(refs)))
> > > > >     			atomic_cond_read_acquire(refs, !VAL);
> > > > >     	}
> > > > > +
> > > > > +	/* They are all completed: */
> > > > > +	text_poke_array.nr_entries = 0;
> > > > >     }
> > > > >     static void __smp_text_poke_batch_add(void *addr, const void *opcode, size_t len, const void *emulate)
> > > > > @@ -2857,20 +2860,16 @@ static bool text_poke_addr_ordered(void *addr)
> > > > >     void smp_text_poke_batch_finish(void)
> > > > >     {
> > > > > -	if (text_poke_array.nr_entries) {
> > > > > +	if (text_poke_array.nr_entries)
> > > > >     		smp_text_poke_batch_process();
> > > > > -		text_poke_array.nr_entries = 0;
> > > > > -	}
> > > > >     }
> > > >
> > > > This function becomes trivial, why not simply move the check into
> > > > smp_text_poke_batch_process and rename it to smp_text_poke_batch_finish ?
> > >
> > > Yeah, that's pretty much what happens in patch #47:
> > >
> > >     x86/alternatives: Remove 'smp_text_poke_batch_flush()'
> >
> > Well, that patch removes poke_batch_flush but still retains
> > poke_batch_finish + poke_batch_process. My suggestion is to eliminate
> > poke_batch_process name and turn it into poke_batch_finish by moving the
> > check from poke_batch_finish into poke_batch_process.
>
> So, in the context of the full tree at:
>
>    git://git.kernel.org/pub/scm/linux/kernel/git/mingo/tip.git WIP.x86/alternatives
>
> Standalone smp_text_poke_batch_process() is still needed, because
> smp_text_poke_batch_add() uses it to drive the whole 'batching'
> machinery.
>
> If smp_text_poke_batch_process() finishes for each call, if I
> understand your suggestion correctly, that reduces the amount of
> batching done, which is a disadvantage.
>
> Note how right now it's possible to do something like this:
>
> 	smp_text_poke_batch_add(1);
> 	smp_text_poke_batch_add(1);
> 	smp_text_poke_batch_add(1);
> 	smp_text_poke_batch_add(1);
> 	smp_text_poke_batch_finish();
>
> This results in a single call to smp_text_poke_batch_process(), with 4
> entries, a single 4-entry flush in essence.

I meant doing this:

diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 5b1a6252a4b9..b6a781b9de26 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -2587,12 +2587,16 @@ noinstr int 
smp_text_poke_int3_trap_handler(struct pt_regs *regs)
  *               replacing opcode
  *     - SMP sync all CPUs
  */
-static void smp_text_poke_batch_process(void)
+void smp_text_poke_batch_finish(void)
 {
        unsigned char int3 = INT3_INSN_OPCODE;
        unsigned int i;
        int do_sync;

+
+       if (!text_poke_array.nr_entries)
+               return;
+
        lockdep_assert_held(&text_mutex);

        /*
@@ -2832,12 +2836,6 @@ static bool text_poke_addr_ordered(void *addr)
        return true;
 }

-void smp_text_poke_batch_finish(void)
-{
-       if (text_poke_array.nr_entries)
-               smp_text_poke_batch_process();
-}
-
 /**
  * smp_text_poke_batch_add() -- update instruction on live kernel on 
SMP, batched
  * @addr:      address to patch
@@ -2854,7 +2852,7 @@ void smp_text_poke_batch_finish(void)
 void __ref smp_text_poke_batch_add(void *addr, const void *opcode, 
size_t len, const void *emulate)
 {
        if (text_poke_array.nr_entries == TEXT_POKE_ARRAY_MAX || 
!text_poke_addr_ordered(addr))
-               smp_text_poke_batch_process();
+               smp_text_poke_batch_finish();
        __smp_text_poke_batch_add(addr, opcode, len, emulate);
 }


AFAICS this doesn't change the semantics. I.e smp_text_poke_batch_add 
will call poke_batch_finish iff the address to be added violates the 
sorted order of text_poke_array. The net effect is we have 1 less 
function name to care about.

> Thanks,
>
> 	Ingo