Hi to all.
I recently got MiNt PCI IDE ATA/133 RAID controller based on SiI 0680
chipset. I browsed through the archives and I know that the driver is
known to be broken and simply doesn't work.
That was some time ago, meanwhile there were some fixes/hdparm
almost-solutions, but none of them seems to work. I remember that
Andre Hendrick said he's talking with SiI guys to find a final
solution, but no more info after that...
Therefore my question is: what is the current status of this driver?
Are there any patches available? Or I will have to just buy another
controller? (No problems with HPT or Promise?)
t.
ps. my problems are "standard" (lost interrupts, timeouts, read_intr
errors etc.) and were described on this list by several other people.
I use almost vanilla 2.4.22 (only with Openwall patches)
--
Tomasz B?tor e-mail: [email protected] ICQ: 101194886
------ ---- -- - - - - - - - - - - - -- ---- ------
"What do you mean, they threw it back?"
-- last words in a RPG game
On Maw, 2003-09-02 at 17:55, Tomasz B±tor wrote:
> I recently got MiNt PCI IDE ATA/133 RAID controller based on SiI 0680
> chipset. I browsed through the archives and I know that the driver is
> known to be broken and simply doesn't work.
It just works. You do want 2.4.22 ideally, and you want 2.4.22-ac to use
hotplug.
Alan
>
> Hi to all.
> I recently got MiNt PCI IDE ATA/133 RAID controller based on SiI 0680
> chipset. I browsed through the archives and I know that the driver is
> known to be broken and simply doesn't work.
>
> That was some time ago, meanwhile there were some fixes/hdparm
> almost-solutions, but none of them seems to work. I remember that
> Andre Hendrick said he's talking with SiI guys to find a final
> solution, but no more info after that...
>
> Therefore my question is: what is the current status of this driver?
> Are there any patches available? Or I will have to just buy another
> controller? (No problems with HPT or Promise?)
>
I recently posted a problem with a Promise
UltraTrak 133 TX2 controller (subject: PROBLEM: Powerquest Drive
Image let the kernel panic). With kernel 2.4.20 every time moving
several GB of data I get an OOPS. I already thought trying a
controller with a SI680-chipset... Maybe that's not a good idea.
Mit freundlichen Gruessen/Best regards,
Sebastian Piecha
EMail: [email protected]
Witam,
Dnia Wed, Sep 03, 2003 at 12:49:40PM +0100 Alan Cox napisal(a):
> On Maw, 2003-09-02 at 17:55, Tomasz B??tor wrote:
> > I recently got MiNt PCI IDE ATA/133 RAID controller based on SiI 0680
> > chipset. I browsed through the archives and I know that the driver is
> > known to be broken and simply doesn't work.
>
> It just works. You do want 2.4.22 ideally, and you want 2.4.22-ac to use
> hotplug.
It doesn't for me. I have no idea what could I possibly do wrong, but
I've tried dozens of possibilities without any luck. Compiling in
siimage.c = drive errors, ide2 reset and infinite loop of "lost
interrupt" messages at boot time. Without siimage.c compiled and with
ide2=xxx ide3=xxx parameters in lilo, disks are visible, but of course
there is no DMA.
Any ideas?
t.
ps. yes, I use 2.4.22, without hotplug.
--
Tomasz B?tor e-mail: [email protected] ICQ: 101194886
------ ---- -- - - - - - - - - - - - -- ---- ------
"The most important job is not to be governor, or first lady in my case."
-- George W. Bush
siimage.c is not broken, works, and may be perfect, since workarounds
point to ioapic
handling for which siimage.c is just downstream. ioapic code could be
fine, too, and
boards or chips responsible for ioapic problems, but ioapic is involved
when things
go wrong. Total ACPI lobotomy is unnecessary, acpi is valuable.
Alan Cox wrote:
>On Maw, 2003-09-02 at 17:55, Tomasz B±tor wrote:
>
>
>>I recently got MiNt PCI IDE ATA/133 RAID controller based on SiI 0680
>>chipset. I browsed through the archives and I know that the driver is
>>known to be broken and simply doesn't work.
>>
>>
>
>It just works. You do want 2.4.22 ideally, and you want 2.4.22-ac to use
>hotplug.
>
>Alan
>
SiI680 siig ata/133 card, siimage.c(and no separate cmd640 fix option)
kern 2.6.0-test4,
kernel APIC on but kernel IOAPIC off then hdparm -c1 -d1 -p9 -u0 -X70
the drives
on the sii680
cmos setup setting APIC off gets rtl8139c card to ping by letting linux
turn apic on
That's nforce2 mbo with rtl8139c on **card** and sii680 on card.
Then rtl8139a **onboard ** with all via chips in a Shuttle SK41G toaster
FX41 mbo had
the same no-ping problem fixed by using kernel 2.6.0-test2 or test4
instead of 2.4! That
defines 2.6 as "bleating edge" for sheeple--turning acpi off is more
drastic than necessary,
and the siimage.c driver is not "broken" or perhaps even involved with
the IOAPIC problem,
and using 2.6 can make something work where 2.4 doesn't.
I guess sii680 qualifies as bleatiing edge. I was close to going scsi-bear.
Actually I pulled the raid jumper off sii680 so sii680's two drives and
two more on
onboard amd74xx nforce2 are handled by linux software raid0, reiserfs.
-Bob
On Iau, 2003-09-04 at 02:42, Tomasz B±tor wrote:
> It doesn't for me. I have no idea what could I possibly do wrong, but
> I've tried dozens of possibilities without any luck. Compiling in
> siimage.c = drive errors, ide2 reset and infinite loop of "lost
> interrupt" messages at boot time. Without siimage.c compiled and with
> ide2=xxx ide3=xxx parameters in lilo, disks are visible, but of course
> there is no DMA.
>
> Any ideas?
If you disable both APIC and ACPI support is it any happier ?
Witam,
Dnia Thu, Sep 04, 2003 at 11:06:06AM +0100 Alan Cox napisal(a):
> On Iau, 2003-09-04 at 02:42, Tomasz B??tor wrote:
> > It doesn't for me. I have no idea what could I possibly do wrong, but
> > I've tried dozens of possibilities without any luck. Compiling in
> > siimage.c = drive errors, ide2 reset and infinite loop of "lost
> > interrupt" messages at boot time. Without siimage.c compiled and with
> > ide2=xxx ide3=xxx parameters in lilo, disks are visible, but of course
> > there is no DMA.
> >
> > Any ideas?
>
> If you disable both APIC and ACPI support is it any happier ?
Now it's:
CONFIG_X86_GOOD_APIC=y
# CONFIG_X86_UP_APIC is not set
# CONFIG_X86_UP_IOAPIC is not set
# CONFIG_ACPI is not set
At night I'll try turning APIC on (I've never used this before) and disable
it in cmos like Bob said. We'll see.
t.
--
Tomasz Bator e-mail: [email protected] ICQ: 101194886
------ ---- -- - - - - - - - - - - - -- ---- ------
"Today it's a race between programmers making better idiot proof software,
and the universe making better idiots. So far the universe is winning"