A previous patch to atiixp.c was removed but some code has not been
cleaned. Now we remove these code sine they are no use any longer.
Signed-off-by: Conke Hu <[email protected]>
-----------------
--- linux-2.6.20-rc3-git4/drivers/ide/pci/atiixp.c.orig 2007-01-06
18:45:28.000000000 +0800
+++ linux-2.6.20-rc3-git4/drivers/ide/pci/atiixp.c.1 2007-01-06
19:13:55.000000000 +0800
@@ -318,19 +318,6 @@ static void __devinit init_hwif_atiixp(i
hwif->drives[0].autodma = hwif->autodma;
}
-static void __devinit init_hwif_sb600_legacy(ide_hwif_t *hwif)
-{
-
- hwif->atapi_dma = 1;
- hwif->ultra_mask = 0x7f;
- hwif->mwdma_mask = 0x07;
- hwif->swdma_mask = 0x07;
-
- if (!noautodma)
- hwif->autodma = 1;
- hwif->drives[0].autodma = hwif->autodma;
- hwif->drives[1].autodma = hwif->autodma;
-}
static ide_pci_device_t atiixp_pci_info[] __devinitdata = {
{ /* 0 */
@@ -340,13 +327,7 @@ static ide_pci_device_t atiixp_pci_info[
.autodma = AUTODMA,
.enablebits = {{0x48,0x01,0x00}, {0x48,0x08,0x00}},
.bootable = ON_BOARD,
- },{ /* 1 */
- .name = "ATI SB600 SATA Legacy IDE",
- .init_hwif = init_hwif_sb600_legacy,
- .channels = 2,
- .autodma = AUTODMA,
- .bootable = ON_BOARD,
- }
+ },
};
/**
On 1/6/07, Conke Hu <[email protected]> wrote:
> A previous patch to atiixp.c was removed but some code has not been
This one?
http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=ab17443a3df35abe4b7529e83511a591aa7384f3
Doesn't it break existing setups without giving ANY warning?
theoretical (I don't have hardware in question) scenario:
- user uses atiixp and has modular libata/ahci (or no libata/ahci et all)
- user does kernel upgrade
- boot fails
- ...
If this is true please add something like
printk(KERN_WARNING "PCI: setting SB600 SATA to AHCI mode"
" (please use ahci driver instead of atiixp)\n");
to quirk_sb600_sata() so people will at least know what is wrong...
> cleaned. Now we remove these code sine they are no use any longer.
Acked-by: Bartlomiej Zolnierkiewicz <[email protected]>
[ but the patch is line wrapped and unfortunately doesn't apply ]
PS: please always cc: [email protected] on PATA/SATA patches
Thanks,
Bart
On 1/7/07, Bartlomiej Zolnierkiewicz <[email protected]> wrote:
> On 1/6/07, Conke Hu <[email protected]> wrote:
> > A previous patch to atiixp.c was removed but some code has not been
>
> This one?
>
> http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=ab17443a3df35abe4b7529e83511a591aa7384f3
>
> Doesn't it break existing setups without giving ANY warning?
>
> theoretical (I don't have hardware in question) scenario:
> - user uses atiixp and has modular libata/ahci (or no libata/ahci et all)
> - user does kernel upgrade
> - boot fails
> - ...
>
> If this is true please add something like
>
> printk(KERN_WARNING "PCI: setting SB600 SATA to AHCI mode"
> " (please use ahci driver instead of atiixp)\n");
>
> to quirk_sb600_sata() so people will at least know what is wrong...
>
> > cleaned. Now we remove these code sine they are no use any longer.
>
> Acked-by: Bartlomiej Zolnierkiewicz <[email protected]>
>
> [ but the patch is line wrapped and unfortunately doesn't apply ]
>
> PS: please always cc: [email protected] on PATA/SATA patches
>
> Thanks,
> Bart
>
Hi Bart,
I've tried to access the following link to make sure which it is,
but failed. The internet here is almost broken.
http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=ab17443a3df35abe4b7529e83511a591aa7384f3
I sent out 2 patches for the same SB600 legacy IDE issue. the later
(sb600 pci qurik) is better so we should clean the previous patch
which was applied to atiixp.c. -- that is what this patch does.
BTW, I re-create and re-send the patch (see below) based on
2.6.20-rc4, in last patch I fogot to rename atiixp.c.1 to atiixp.c
which may lead to patch fail, nothing else different.
And maybe no need to re-ACK if last one is accepted:)
Thanks!
--------------------
--- linux-2.6.20-rc4/drivers/ide/pci/atiixp.c.1 2007-01-09
15:19:05.000000000 +0800
+++ linux-2.6.20-rc4/drivers/ide/pci/atiixp.c 2007-01-09
15:17:54.000000000 +0800
@@ -320,20 +320,6 @@ static void __devinit init_hwif_atiixp(i
hwif->drives[0].autodma = hwif->autodma;
}
-static void __devinit init_hwif_sb600_legacy(ide_hwif_t *hwif)
-{
-
- hwif->atapi_dma = 1;
- hwif->ultra_mask = 0x7f;
- hwif->mwdma_mask = 0x07;
- hwif->swdma_mask = 0x07;
-
- if (!noautodma)
- hwif->autodma = 1;
- hwif->drives[0].autodma = hwif->autodma;
- hwif->drives[1].autodma = hwif->autodma;
-}
-
static ide_pci_device_t atiixp_pci_info[] __devinitdata = {
{ /* 0 */
.name = "ATIIXP",
@@ -342,13 +328,7 @@ static ide_pci_device_t atiixp_pci_info[
.autodma = AUTODMA,
.enablebits = {{0x48,0x01,0x00}, {0x48,0x08,0x00}},
.bootable = ON_BOARD,
- },{ /* 1 */
- .name = "ATI SB600 SATA Legacy IDE",
- .init_hwif = init_hwif_sb600_legacy,
- .channels = 2,
- .autodma = AUTODMA,
- .bootable = ON_BOARD,
- }
+ },
};
/**
On 1/9/07, Conke Hu <[email protected]> wrote:
> On 1/7/07, Bartlomiej Zolnierkiewicz <[email protected]> wrote:
> > On 1/6/07, Conke Hu <[email protected]> wrote:
> > > A previous patch to atiixp.c was removed but some code has not been
> >
> > This one?
> >
> > http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=ab17443a3df35abe4b7529e83511a591aa7384f3
> >
> > Doesn't it break existing setups without giving ANY warning?
> >
> > theoretical (I don't have hardware in question) scenario:
> > - user uses atiixp and has modular libata/ahci (or no libata/ahci et all)
> > - user does kernel upgrade
> > - boot fails
> > - ...
> >
> > If this is true please add something like
> >
> > printk(KERN_WARNING "PCI: setting SB600 SATA to AHCI mode"
> > " (please use ahci driver instead of atiixp)\n");
> >
> > to quirk_sb600_sata() so people will at least know what is wrong...
> >
> > > cleaned. Now we remove these code sine they are no use any longer.
> >
> > Acked-by: Bartlomiej Zolnierkiewicz <[email protected]>
> >
> > [ but the patch is line wrapped and unfortunately doesn't apply ]
> >
> > PS: please always cc: [email protected] on PATA/SATA patches
> >
> > Thanks,
> > Bart
> >
>
>
> Hi Bart,
> I've tried to access the following link to make sure which it is,
> but failed. The internet here is almost broken.
> http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=ab17443a3df35abe4b7529e83511a591aa7384f3
Yeah, kernel.org is heaving some load related problems.
> I sent out 2 patches for the same SB600 legacy IDE issue. the later
> (sb600 pci qurik) is better so we should clean the previous patch
> which was applied to atiixp.c. -- that is what this patch does.
Yep, I was talking about "PCI: ATI sb600 sata quirk" which was merged
into -rc2.
> BTW, I re-create and re-send the patch (see below) based on
> 2.6.20-rc4, in last patch I fogot to rename atiixp.c.1 to atiixp.c
> which may lead to patch fail, nothing else different.
> And maybe no need to re-ACK if last one is accepted:)
As stated in the other mail - these patches are in -mm kernel already.
Thanks,
Bart