Re: [RFC 03/16] kgr: initial code

[Jiri Slaby <jslaby@suse.cz>]

On 04/30/2014 04:56 PM, Steven Rostedt wrote:
> On Wed, 30 Apr 2014 16:30:36 +0200
> Jiri Slaby <jslaby@suse.cz> wrote:
>> --- a/arch/x86/kernel/entry_64.S
>> +++ b/arch/x86/kernel/entry_64.S
>> @@ -615,6 +615,7 @@ GLOBAL(system_call_after_swapgs)
>>  	movq  %rax,ORIG_RAX-ARGOFFSET(%rsp)
>>  	movq  %rcx,RIP-ARGOFFSET(%rsp)
>>  	CFI_REL_OFFSET rip,RIP-ARGOFFSET
>> +	movw $0, TI_kgr_in_progress+THREAD_INFO(%rsp,RIP-ARGOFFSET)
> 
> Why is this not a entry flag? Because you just added a store into a
> fast path of the kernel for something that will be hardly ever used.

Actually it is converted later in the series, please see 16/16.

>> --- /dev/null
>> +++ b/include/linux/kgr.h
>> @@ -0,0 +1,71 @@
>> +#ifndef LINUX_KGR_H
>> +#define LINUX_KGR_H
>> +
>> +#include <linux/init.h>
>> +#include <linux/ftrace.h>
>> +
>> +#include <asm/kgr.h>
>> +
>> +#ifdef CONFIG_KGR
>> +
>> +#define KGR_TIMEOUT 30
>> +#define KGR_DEBUG 1
>> +
>> +#ifdef KGR_DEBUG
>> +#define kgr_debug(args...)	\
>> +	pr_info(args);
>> +#else
>> +#define kgr_debug(args...) { }
>> +#endif
> 
> Why not just use pr_debug(), as that's not defined unless you add DEBUG
> as a define anyway?

Yeah, OK.

thanks,
-- 
js
suse labs