PATCH: libata PATA patches

PATCH: libata PATA patches

[Alan Cox]

I've put a new patch versus 2.6.14-mm1 on 
http://zeniv.linux.org.uk/~alan/IDE

Alan

Re: PATCH: libata PATA patches

[Adrian Bunk]

On Tue, Nov 08, 2005 at 02:33:06PM +0000, Alan Cox wrote:

> I've put a new patch versus 2.6.14-mm1 on 
> http://zeniv.linux.org.uk/~alan/IDE

I'm wondering about the -m32 additions at the top of your patch.

Shouldn't the following at the top of arch/i386/Makefile already do the 
same?

HAS_BIARCH      := $(call cc-option-yn, -m32)
ifeq ($(HAS_BIARCH),y)
AS              := $(AS) --32
LD              := $(LD) -m elf_i386
CC              := $(CC) -m32
endif



> Alan

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: PATCH: libata PATA patches

[Alan Cox]

On Maw, 2005-11-08 at 15:13 +0100, Adrian Bunk wrote:
> Shouldn't the following at the top of arch/i386/Makefile already do the 
> same?
> 
> HAS_BIARCH      := $(call cc-option-yn, -m32)
> ifeq ($(HAS_BIARCH),y)
> AS              := $(AS) --32
> LD              := $(LD) -m elf_i386
> CC              := $(CC) -m32
> endif
> 

Its just left overs from my old build patches I forgot to clean out.
Harmless however.

Re: PATCH: libata PATA patches

[Sergei Shtylylov]

Hello.

Alan Cox wrote:

> I've put a new patch versus 2.6.14-mm1 on 
> http://zeniv.linux.org.uk/~alan/IDE

    I found somewhat strange that you check for 0xABCDExxx signature in 32-bit 
PCI config. reg. 0x78 while HighPoint's own drivers read BM reg. 0x90 (i.e. 
PCI config. 0x70) for that. PCI reg. 0x7A is DPLL precision adjust reg. and 
0x7B is the input clock select and IRQ reg., so it'd be quite strange if the 
BIOS used them for any kind of signature...

WBR, Sergei

Re: PATCH: libata PATA patches

[Sergei Shtylylov]

Hello, I wrote:

> Alan Cox wrote:
> 
>> I've put a new patch versus 2.6.14-mm1 on 
>> http://zeniv.linux.org.uk/~alan/IDE
> 
> 
>    I found somewhat strange that you check for 0xABCDExxx signature in 
> 32-bit PCI config. reg. 0x78 while HighPoint's own drivers read BM reg. 
> 0x90 (i.e. PCI config. 0x70) for that. PCI reg. 0x7A is DPLL precision 
> adjust reg. and 0x7B is the input clock select and IRQ reg., so it'd be 
> quite strange if the BIOS used them for any kind of signature...

    Pardon me for being too hasty. :-)
    I was talking about drivers/scsi/pata_hpt37x.c

WBR, Sergei

Re: PATCH: libata PATA patches

[Alan Cox]

On Mer, 2005-11-09 at 16:31 +0300, Sergei Shtylylov wrote:
>     I found somewhat strange that you check for 0xABCDExxx signature in 32-bit 
> PCI config. reg. 0x78 while HighPoint's own drivers read BM reg. 0x90 (i.e. 
> PCI config. 0x70) for that. PCI reg. 0x7A is DPLL precision adjust reg. and 
> 0x7B is the input clock select and IRQ reg., so it'd be quite strange if the 
> BIOS used them for any kind of signature...

Yeah thats a bug. What I get for reading the comments in the old driver
and not double checking against the HPT reference code. Thanks

Alan