pci-resources feedback

pci-resources feedback

[Andreas Tobler]

Hi all,
I have applied the patches from Michel to my 2399p3 tree. The main
reason for that was some hope in getting my CardBus card recognized
properly. Ok, not that successful but I think on the right way. At least
what dmesg says.

Also I'm confused about the ati memory range two. Which base is the one
I have to take?
A wallstreet I.

Below the dmesg from the 2399p3 (only pci-probing) and the corresponding
lspci -vv

as PCIBIOS_MIN_MEM I took 0x80000000.

---2399-dmesg---
PCI: Probing PCI hardware (semiautomatic)
Scanning bus 00
Found 00:00 [1057/0002] 000600 00
Found 00:68 [106b/0010] 00ff00 00
Found 00:80 [106b/0010] 00ff00 00
Found 00:88 [1002/4c47] 000300 00
Found 00:98 [104c/ac15] 000607 02
Found 00:99 [104c/ac15] 000607 02
Fixups for bus 00
Scanning CardBus bridge 00:13.0			<------ sounds good for CardBus!!
Scanning CardBus bridge 00:13.1
Bus scan for 00 returning with max=09
PCI: Fixing device 00:00.0 (1057:0002)
PCI: Fixing device 00:0d.0 (106b:0010)
PCI: Fixing device 00:10.0 (106b:0010)
PCI: Fixing device 00:11.0 (1002:4c47)
PCI: Setting IRQ 24 on device 00:11.0.
PCI: Correcting IO address 1 on device 00:11.0, now fe000400.
PCI: Enabling device 00:11.0 (0086 -> 0087)
PCI: Fixing device 00:13.0 (104c:ac15)
PCI: Setting IRQ 22 on device 00:13.0.
PCI: Fixing device 00:13.1 (104c:ac15)
PCI: Setting IRQ 23 on device 00:13.1.
PCI: Resource f4000000-f407ffff (f=200, d=0, p=0)
PCI: Resource f3000000-f307ffff (f=200, d=0, p=0)
PCI: Resource 82000000-82ffffff (f=200, d=0, p=0)
PCI: Resource fe000400-fe0004ff (f=101, d=0, p=0)
PCI: Resource 82fff000-82ffffff (f=200, d=0, p=0)
PCI: Cannot allocate resource region 2 of device 00:11.0
PCI: Resource 81803000-81803fff (f=200, d=0, p=0)
PCI: Resource 81802000-81802fff (f=200, d=0, p=0)
  for root[0:ffffffff] min[80000000] size[1000]
  got res[80000000:80000fff] for resource 2
---end---

---lspci2399p3---

00:11.0 VGA compatible controller: ATI Technologies Inc 215LG [Mach64
LG] (rev 80)
	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
Stepping+ SERR- FastB2B-
	Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
	Latency: 8 min, 32 set, cache line size 08
	Interrupt: pin A routed to IRQ 24
	Region 0: Memory at 82000000 (32-bit, non-prefetchable)
	Region 1: I/O ports at fe000400
	Region 2: Memory at 80000000 (32-bit, non-prefetchable)
---end---

For comparison reasons a lspci -vv from the actual 2215p14:

---lspci2215p14---

00:11.0 VGA compatible controller: ATI Technologies Inc 215LG [Mach64
LG] (rev 80)
	Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
Stepping+ SERR- FastB2B-
	Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
	Latency: 8 min, 32 set, cache line size 08
	Interrupt: pin A routed to IRQ 24
	Region 0: Memory at 82000000 (32-bit, non-prefetchable)
	Region 1: I/O ports at 0400 [disabled]
	Region 2: Memory at 82fff000 (32-bit, non-prefetchable)
---end---


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: pci-resources feedback

[Michel Lanners]

Hi Andreas,

On  27 Mar, this message from Andreas Tobler echoed through cyberspace:
> Also I'm confused about the ati memory range two. Which base is the one
> I have to take?

As I understand it (which may be wrong ;-) that's the MMIO region
corresponding to the (VGA?) IO ports used for configuring the device.

Problem is by default it lies within the framebuffer region...

> ---2399-dmesg---
> PCI: Probing PCI hardware (semiautomatic)
> Scanning bus 00
> Found 00:00 [1057/0002] 000600 00
> Found 00:68 [106b/0010] 00ff00 00
> Found 00:80 [106b/0010] 00ff00 00

Hmmm... two times the same dev?? You really have two heathrows?

> Found 00:88 [1002/4c47] 000300 00
> Found 00:98 [104c/ac15] 000607 02
> Found 00:99 [104c/ac15] 000607 02
> Fixups for bus 00
> Scanning CardBus bridge 00:13.0			<------ sounds good for CardBus!!
> Scanning CardBus bridge 00:13.1

Cool ;-)

> Bus scan for 00 returning with max=09
> PCI: Fixing device 00:00.0 (1057:0002)
> PCI: Fixing device 00:0d.0 (106b:0010)
> PCI: Fixing device 00:10.0 (106b:0010)
> PCI: Fixing device 00:11.0 (1002:4c47)
> PCI: Setting IRQ 24 on device 00:11.0.
> PCI: Correcting IO address 1 on device 00:11.0, now fe000400.

Hey, it even works on grackle!

> PCI: Enabling device 00:11.0 (0086 -> 0087)
> PCI: Fixing device 00:13.0 (104c:ac15)
> PCI: Setting IRQ 22 on device 00:13.0.
> PCI: Fixing device 00:13.1 (104c:ac15)
> PCI: Setting IRQ 23 on device 00:13.1.
> PCI: Resource f4000000-f407ffff (f=200, d=0, p=0)
> PCI: Resource f3000000-f307ffff (f=200, d=0, p=0)
> PCI: Resource 82000000-82ffffff (f=200, d=0, p=0)
> PCI: Resource fe000400-fe0004ff (f=101, d=0, p=0)
> PCI: Resource 82fff000-82ffffff (f=200, d=0, p=0)
> PCI: Cannot allocate resource region 2 of device 00:11.0

Here the resource code notices the conflict on the ATI.

> PCI: Resource 81803000-81803fff (f=200, d=0, p=0)
> PCI: Resource 81802000-81802fff (f=200, d=0, p=0)
>   for root[0:ffffffff] min[80000000] size[1000]
>   got res[80000000:80000fff] for resource 2

This is the message as the conflicting (second) region of the ATI gets
remapped. It did work... as can be seen ba the lspci outtput.

So far, so good :-)

Michel

-------------------------------------------------------------------------
Michel Lanners                 |  " Read Philosophy.  Study Art.
23, Rue Paul Henkes            |    Ask Questions.  Make Mistakes.
L-1710 Luxembourg              |
email   mlan@cpu.lu            |
http://www.cpu.lu/~mlan        |                     Learn Always. "


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: pci-resources feedback

[Michael Schmitz]

>
> Here the resource code notices the conflict on the ATI.
>
> > PCI: Resource 81803000-81803fff (f=200, d=0, p=0)
> > PCI: Resource 81802000-81802fff (f=200, d=0, p=0)
> >   for root[0:ffffffff] min[80000000] size[1000]
> >   got res[80000000:80000fff] for resource 2
>
> This is the message as the conflicting (second) region of the ATI gets
> remapped. It did work... as can be seen ba the lspci outtput.

And it even used the right function (the one in drivers/pci/ to do it. Now
what kernel version was that with?

Seems the wallstreet was another one of the 'broken mapping for atyfb'
boxes. XFree86 4.0 should work fine now.

	Michael


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: pci-resources feedback

[Michel Lanners]

On  27 Mar, this message from Michael Schmitz echoed through cyberspace:
>>
>> Here the resource code notices the conflict on the ATI.
>>
>> > PCI: Resource 81803000-81803fff (f=200, d=0, p=0)
>> > PCI: Resource 81802000-81802fff (f=200, d=0, p=0)
>> >   for root[0:ffffffff] min[80000000] size[1000]
>> >   got res[80000000:80000fff] for resource 2
>>
>> This is the message as the conflicting (second) region of the ATI gets
>> remapped. It did work... as can be seen ba the lspci outtput.
>
> And it even used the right function (the one in drivers/pci/ to do it. Now
> what kernel version was that with?

If I am not mistaken, 2.3.99 with the following patch applied:

http://www.cpu.lu/~mlan/linux/dev/2.3.99-pre3-pci.diff

> Seems the wallstreet was another one of the 'broken mapping for atyfb'
> boxes. XFree86 4.0 should work fine now.

;-)

Michel

-------------------------------------------------------------------------
Michel Lanners                 |  " Read Philosophy.  Study Art.
23, Rue Paul Henkes            |    Ask Questions.  Make Mistakes.
L-1710 Luxembourg              |
email   mlan@cpu.lu            |
http://www.cpu.lu/~mlan        |                     Learn Always. "


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: pci-resources feedback

[Andreas Tobler]

[-- Attachment #1: Type: text/plain, Size: 831 bytes --]



Michael Schmitz wrote:
>
> >
> > Here the resource code notices the conflict on the ATI.
> >
> > > PCI: Resource 81803000-81803fff (f=200, d=0, p=0)
> > > PCI: Resource 81802000-81802fff (f=200, d=0, p=0)
> > >   for root[0:ffffffff] min[80000000] size[1000]
> > >   got res[80000000:80000fff] for resource 2
> >
> > This is the message as the conflicting (second) region of the ATI gets
> > remapped. It did work... as can be seen ba the lspci outtput.
>
> And it even used the right function (the one in drivers/pci/ to do it. Now
> what kernel version was that with?
>
> Seems the wallstreet was another one of the 'broken mapping for atyfb'
> boxes. XFree86 4.0 should work fine now.

As written 2399pre3 on a wallstreet I (spring 98)

If needed more input/output?

Thx
Andreas

btw, attached the lspci -vv for the whole tree

[-- Attachment #2: Document --]
[-- Type: image/x-xbitmap, Size: 2723 bytes --]

Re: pci-resources feedback

[Timothy A. Seufert]

At 9:09 PM +0200 3/27/00, Michel Lanners wrote:

>>  ---2399-dmesg---
>>  PCI: Probing PCI hardware (semiautomatic)
>>  Scanning bus 00
>>  Found 00:00 [1057/0002] 000600 00
>>  Found 00:68 [106b/0010] 00ff00 00
>>  Found 00:80 [106b/0010] 00ff00 00
>
>Hmmm... two times the same dev?? You really have two heathrows?

The "Wall Street" and "PDQ" series PB G3s really do have two heathrow
ASICs.  One of them is the main Heathrow.  The other is only present
because these two PB models need three IDE channels and three serial
channels, and a single Heathrow provides just two of each.  (Two IDEs
for the two expansion bays, one IDE for the internal HD, one serial
for IrDA, one serial for the internal modem, and one serial for the
general purpose serial port.)

   Tim Seufert

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/