Re: [PATCH -tip 3/4]Text Edit Lock - Smp alternatives support

[Masami Hiramatsu <mhiramat@redhat.com>]

Mathieu Desnoyers wrote:
> * Masami Hiramatsu (mhiramat@redhat.com) wrote:
>> Use the mutual exclusion provided by the text edit lock in alternatives code.
>> Since alternative_smp_* will be called from module init code, etc,
>> we'd better protect it from other subsystems.
>>
>> Signed-off-by: Masami Hiramatsu <mhiramat@redhat.com>
>> Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
>> Cc: Ingo Molnar <mingo@elte.hu>
>> ---
>>  arch/x86/kernel/alternative.c |    5 +++++
>>  1 file changed, 5 insertions(+)
>>
>> Index: 2.6.29-rc7/arch/x86/kernel/alternative.c
>> ===================================================================
>> --- 2.6.29-rc7.orig/arch/x86/kernel/alternative.c
>> +++ 2.6.29-rc7/arch/x86/kernel/alternative.c
>> @@ -5,6 +5,7 @@
>>  #include <linux/kprobes.h>
>>  #include <linux/mm.h>
>>  #include <linux/vmalloc.h>
>> +#include <linux/memory.h>
>>  #include <asm/alternative.h>
>>  #include <asm/sections.h>
>>  #include <asm/pgtable.h>
>> @@ -226,6 +227,7 @@ static void alternatives_smp_lock(u8 **s
>>  {
>>  	u8 **ptr;
>>
>> +	mutex_lock(&text_mutex);
>>  	for (ptr = start; ptr < end; ptr++) {
>>  		if (*ptr < text)
>>  			continue;
>> @@ -234,6 +236,7 @@ static void alternatives_smp_lock(u8 **s
>>  		/* turn DS segment override prefix into lock prefix */
>>  		text_poke(*ptr, ((unsigned char []){0xf0}), 1);
>>  	};
>> +	mutex_unlock(&text_mutex);
>>  }
>>
>>  static void alternatives_smp_unlock(u8 **start, u8 **end, u8 *text, u8 *text_end)
>> @@ -243,6 +246,7 @@ static void alternatives_smp_unlock(u8 *
>>  	if (noreplace_smp)
>>  		return;
>>
>> +	mutex_lock(&text_mutex);
>>  	for (ptr = start; ptr < end; ptr++) {
>>  		if (*ptr < text)
>>  			continue;
>> @@ -251,6 +255,7 @@ static void alternatives_smp_unlock(u8 *
>>  		/* turn lock prefix into DS segment override prefix */
>>  		text_poke(*ptr, ((unsigned char []){0x3E}), 1);
>>  	};
>> +	mutex_unlock(&text_mutex);
> 
> And for these cases, the "alternatives_smp_lock" and
> "alternatives_smp_unlock", the system is running as uniprocessor.
> Therefore it's enough to disable interrupts within text_poke to make it
> work correctly.

Hmm, you are right. With my atomic-text_poke() take2 patch,
we don't need this patch.

Thank you,

> 
> Mathieu
> 
>>  }
>>
>>  struct smp_alt_module {
>> -- 
>> Masami Hiramatsu
>>
>> Software Engineer
>> Hitachi Computer Products (America) Inc.
>> Software Solutions Division
>>
>> e-mail: mhiramat@redhat.com
>>
> 

-- 
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America) Inc.
Software Solutions Division

e-mail: mhiramat@redhat.com