linux-mips
[Top] [All Lists]

Re: [RFC] Add 4KSx support (try 2)

To: "Kevin D. Kissell" <kevink@mips.com>
Subject: Re: [RFC] Add 4KSx support (try 2)
From: Franck <vagabon.xyz@gmail.com>
Date: Mon, 31 Oct 2005 17:01:42 +0100
Cc: linux-mips@linux-mips.org, Ralf Baechle <ralf@linux-mips.org>, "Maciej W. Rozycki" <macro@linux-mips.org>
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=ksQNZ35ijZmnzseerMjvVYAjGy63QnJrW3BEi/iZN4SewejDynPSd/ltO6ywLg1s3fFZYLHmL9tpTPCfa/twKRgbWMsNyLjfXgZj2tAUyh3Ee+PuBPU+gUJEwFHQ/b1R9Nk/7XAipM1vn9hnWhbg8WAUBK+1ptzpn2qgvr5givU=
In-reply-to: <4365DF22.8060004@mips.com>
Original-recipient: rfc822;linux-mips@linux-mips.org
References: <cda58cb80510310034k60b273dfm@mail.gmail.com> <4365DF22.8060004@mips.com>
Sender: linux-mips-bounce@linux-mips.org
Thanks Kevin for responding !

2005/10/31, Kevin D. Kissell <kevink@mips.com>:
> I'm not set up to actually apply and test the patch,
> but for whatever it's worth, the functionality in the patch
> appears to be superficially correct, and more-or-less
> what I would have done.  That having been said, I think
> you're creating more changes than are really required.
>
> Having seperate target call-outs for the Sc/Sd in the
> arch/mips/kernel/Makefile just to avoid having r4k_fpu.o
> linked it creates cruft for a savings of 400-odd bytes
> of kernel image, and I'd either have not bothered or have
> figured out a more generic way to strip out FP support
> for FP-less cores.
>
> There are places, for example arch/mips/mm/cache.c, but
> also some of the other makefiles, where you're using your
> new config flags to drive things where the standard
> CONFIG_CPU_MIPS32 (which I guess has now fragmented into
> CONFIG_CPU_MIPS32_R1 and CONFIG_CPU_MIPS32_R2, which would
> apply to the Sc and Sd respectively) would do the right thing
> while creating fewer source file mods.
>

That's correct but CONFIG_CPU_MIPS32_Rx seems to be a fallback case.
Don't other cpu use their own flags whereas they could just use
CONFIG_CPU_MIPS32_Rx flag instead ?

> Have you thought about what the ACX state would mean for
> kernel debuggers in general and kgdb in particular?
>

no, I didn't. I took a look at arch/mips/kernel/gdb-low.S and it seems
to be required....

Thanks
--
               Franck

<Prev in Thread] Current Thread [Next in Thread>