Re: [PATCH 03/11] x86: fix math_emu register frame access

From: Tejun Heo <tj@kernel.org>
To: Brian Gerst <brgerst@gmail.com>
Cc: hpa@zytor.com, jeremy@goop.org, tglx@linutronix.de,
	mingo@elte.hu, linux-kernel@vger.kernel.org, x86@kernel.org,
	rusty@rustcorp.com.au
Subject: Re: [PATCH 03/11] x86: fix math_emu register frame access
Date: Tue, 10 Feb 2009 10:08:39 +0900	[thread overview]
Message-ID: <4990D397.8060705@kernel.org> (raw)
In-Reply-To: <73c1f2160902090913n20283872r2cb5ae774560d047@mail.gmail.com>

Hello, Brian.

Brian Gerst wrote:
>>  {
>>        printk(KERN_EMERG
>>                "math-emulation not enabled and no coprocessor found.\n");
>> @@ -910,16 +910,19 @@ asmlinkage void math_emulate(long arg)
>>  }
>>  #endif /* CONFIG_MATH_EMULATION */
>>
>> -dotraplinkage void __kprobes
>> -do_device_not_available(struct pt_regs *regs, long error)
>> +dotraplinkage void __kprobes do_device_not_available(struct pt_regs regs)
> 
> Why did you change this to pass by value?  This is called from the
> generic trap handler (error_code), which already passes the pt_regs
> pointer in %eax.

Ah.. right, my bad.  I somehow thought dotraplinkage is asmlinkage on
x86_32.  Yeah, the original code was fine in that area, but then again
most other trap functions take struct pt_regs.  Anyways, you're gonna
post patches to convert trap functions to take pointer to pt_regs,
right?  Can you please fix this one up together with those?

Thanks.

-- 
tejun