Re: [SOLVED] Re: 2.6.13-rc2 hangs at boot

Re: [SOLVED] Re: 2.6.13-rc2 hangs at boot

[Mikael Pettersson]

On Fri, 8 Jul 2005 12:38:28 +0300 (EEST), Tero Roponen wrote:
>On Fri, 8 Jul 2005, Ivan Kokshaysky wrote:
>
>> On Fri, Jul 08, 2005 at 10:57:56AM +0300, Tero Roponen wrote:
>> > Thanks! Vanilla 2.6.13-rc2 with below patch applied
>> > works perfectly!
>>
>> Thanks for testing. Though, bad news are that it's still unclear
>> why your machine doesn't like large cardbus windows and therefore
>> how to fix that correctly.
>>
>
>> May I ask you to do couple of another tests with following variations
>> of the latest patch (this will help to determine whether it's IO or
>> MEM ranges to blame)?
>> 1.
>> #define CARDBUS_IO_SIZE		(256)
>> #define CARDBUS_MEM_SIZE	(32*1024*1024)
>
>This works correctly.
>
>
>
>>
>> 2.
>> #define CARDBUS_IO_SIZE		(4096)
>> #define CARDBUS_MEM_SIZE	(4*1024*1024)
>
>This hangs at boot.

Same here: 2.6.13-rc2 vanilla and 2.6.13-rc2 + patch #2 above both hang,
but 2.6.13-rc2 + patch #1 boots fine.

/Mikael

Re: [SOLVED] Re: 2.6.13-rc2 hangs at boot

[Ivan Kokshaysky]

On Fri, Jul 08, 2005 at 03:54:00PM +0200, Mikael Pettersson wrote:
> Same here: 2.6.13-rc2 vanilla and 2.6.13-rc2 + patch #2 above both hang,
> but 2.6.13-rc2 + patch #1 boots fine.

I suspect that those larger (4K) cardbus IO windows are clashing with
some legacy IO ports. If so, this could be avoided by setting ISAEN
bridge control bit.
It would be interesting to see whether the patch below works or not...

Ivan.

--- 2.6.13-rc2-git4/drivers/pci/setup-bus.c	Tue Jul 12 16:11:19 2005
+++ linux/drivers/pci/setup-bus.c	Tue Jul 12 16:43:33 2005
@@ -88,6 +88,7 @@ pci_setup_cardbus(struct pci_bus *bus)
 {
 	struct pci_dev *bridge = bus->self;
 	struct pci_bus_region region;
+	u16 ctrl;
 
 	printk("PCI: Bus %d, cardbus bridge: %s\n",
 		bus->number, pci_name(bridge));
@@ -135,6 +136,12 @@ pci_setup_cardbus(struct pci_bus *bus)
 		pci_write_config_dword(bridge, PCI_CB_MEMORY_LIMIT_1,
 					region.end);
 	}
+
+	/* Disable legacy ISA IO port forwarding, as we do
+	   for PCI-to-PCI bridges. */
+	pci_read_config_word(bridge, PCI_CB_BRIDGE_CONTROL, &ctrl);
+	ctrl |= PCI_CB_BRIDGE_CTL_ISA;
+	pci_write_config_word(bridge, PCI_CB_BRIDGE_CONTROL, ctrl);
 }
 
 /* Initialize bridges with base/limit values we have collected.

Re: [SOLVED] Re: 2.6.13-rc2 hangs at boot

[Tero Roponen]

On Tue, 12 Jul 2005, Ivan Kokshaysky wrote:

> On Fri, Jul 08, 2005 at 03:54:00PM +0200, Mikael Pettersson wrote:
> > Same here: 2.6.13-rc2 vanilla and 2.6.13-rc2 + patch #2 above both hang,
> > but 2.6.13-rc2 + patch #1 boots fine.
>
> I suspect that those larger (4K) cardbus IO windows are clashing with
> some legacy IO ports. If so, this could be avoided by setting ISAEN
> bridge control bit.
> It would be interesting to see whether the patch below works or not...
>
> Ivan.

Hi,

I tried 2.6.13-rc2-git5 with your patch.

First I forgot to install modules and it booted fine.
Then I did a 'sudo make modules_install' and rebooted. It hang.

It seems that everything is fine until some module
loads and does something.

-
Tero Roponen

ps.

In 2.6.12 lsmod gives me this:

Module                  Size  Used by
xirc2ps_cs             15728  1
pcmcia                 21416  5 xirc2ps_cs
sunrpc                128132  1
binfmt_misc             9352  1
yenta_socket           20360  3
rsrc_nonstatic         11488  1 yenta_socket
pcmcia_core            44280  4
xirc2ps_cs,pcmcia,yenta_socket,rsrc_nonstatic
snd_cs4236             13700  0
snd_opl3_lib            9152  1 snd_cs4236
snd_hwdep               7104  1 snd_opl3_lib
snd_cs4236_lib         15008  1 snd_cs4236
snd_mpu401_uart         6048  1 snd_cs4236
snd_rawmidi            20832  1 snd_mpu401_uart
snd_cs4231_lib         23808  2 snd_cs4236,snd_cs4236_lib
snd_seq_oss            33088  0
snd_seq_midi_event      6496  1 snd_seq_oss
snd_seq                50416  4 snd_seq_oss,snd_seq_midi_event
snd_seq_device          6892  4
snd_opl3_lib,snd_rawmidi,snd_seq_oss,snd_seq
snd_pcm_oss            50400  0
snd_mixer_oss          17696  1 snd_pcm_oss
snd_pcm                85096  3 snd_cs4236_lib,snd_cs4231_lib,snd_pcm_oss
snd_timer              21924  4
snd_opl3_lib,snd_cs4231_lib,snd_seq,snd_pcm
snd                    48388  14
snd_cs4236,snd_opl3_lib,snd_hwdep,snd_cs4236_lib,snd_mpu401_uart,snd_rawmidi,snd_cs4231_lib,snd_seq_oss,snd_seq,snd_seq_device,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_timer
soundcore               7456  1 snd
snd_page_alloc          7556  2 snd_cs4231_lib,snd_pcm
floppy                 54772  0
ide_cd                 38436  0
cdrom                  38272  1 ide_cd
evdev                   7456  0

Re: [SOLVED] Re: 2.6.13-rc2 hangs at boot

[Ivan Kokshaysky]

On Tue, Jul 12, 2005 at 07:00:37PM +0300, Tero Roponen wrote:
> It seems that everything is fine until some module
> loads and does something.

What does your /proc/ioports say under 2.6.12 with all modules
loaded?

Ivan.

Re: [SOLVED] Re: 2.6.13-rc2 hangs at boot

[Tero Roponen]

On Tue, 12 Jul 2005, Ivan Kokshaysky wrote:

> On Tue, Jul 12, 2005 at 07:00:37PM +0300, Tero Roponen wrote:
> > It seems that everything is fine until some module
> > loads and does something.
>
> What does your /proc/ioports say under 2.6.12 with all modules
> loaded?
>
> Ivan.

cat /proc/ioports under 2.6.12 gives this:

0000-001f : dma1
0020-0021 : pic1
0040-0043 : timer0
0050-0053 : timer1
0060-006f : keyboard
0070-0077 : rtc
0080-008f : dma page reg
00a0-00a1 : pic2
00c0-00df : dma2
00f0-00ff : fpu
01f0-01f7 : ide0
02f8-02ff : serial
0300-030f : pcmcia_socket1
0330-0331 : MPU401 UART
0388-0389 : OPL2/3 (left)
038a-038b : OPL2/3 (right)
03c0-03df : vga+
03f6-03f6 : ide0
04d0-04d1 : pnp 00:0a
0530-0533 : CS4231
0538-053f : CS4232 Control
0cf8-0cff : PCI conf1
15e0-15ef : pnp 00:0a
4000-40ff : PCI CardBus #01
4400-44ff : PCI CardBus #01
4800-48ff : PCI CardBus #05
4c00-4cff : PCI CardBus #05
9000-901f : 0000:00:06.2
ef00-ef3f : 0000:00:06.3
efa0-efbf : 0000:00:06.3
fcf0-fcff : 0000:00:06.1
  fcf0-fcf7 : ide0


-
Tero Roponen

Re: [SOLVED] Re: 2.6.13-rc2 hangs at boot

[Ivan Kokshaysky]

On Fri, Jul 08, 2005 at 12:38:28PM +0300, Tero Roponen wrote:
> > #define CARDBUS_IO_SIZE		(256)
> > #define CARDBUS_MEM_SIZE	(32*1024*1024)
> 
> This works correctly.

Thanks a lot. I'll go read the cardbus spec.

Ivan.

Re: [SOLVED] Re: 2.6.13-rc2 hangs at boot

[Tero Roponen]

On Fri, 8 Jul 2005, Ivan Kokshaysky wrote:

> On Fri, Jul 08, 2005 at 10:57:56AM +0300, Tero Roponen wrote:
> > Thanks! Vanilla 2.6.13-rc2 with below patch applied
> > works perfectly!
>
> Thanks for testing. Though, bad news are that it's still unclear
> why your machine doesn't like large cardbus windows and therefore
> how to fix that correctly.
>

> May I ask you to do couple of another tests with following variations
> of the latest patch (this will help to determine whether it's IO or
> MEM ranges to blame)?
> 1.
> #define CARDBUS_IO_SIZE		(256)
> #define CARDBUS_MEM_SIZE	(32*1024*1024)

This works correctly.



>
> 2.
> #define CARDBUS_IO_SIZE		(4096)
> #define CARDBUS_MEM_SIZE	(4*1024*1024)

This hangs at boot.



>
> Ivan.
>

-
Tero Roponen

Re: [SOLVED] Re: 2.6.13-rc2 hangs at boot

[Ivan Kokshaysky]

On Fri, Jul 08, 2005 at 10:57:56AM +0300, Tero Roponen wrote:
> Thanks! Vanilla 2.6.13-rc2 with below patch applied
> works perfectly!

Thanks for testing. Though, bad news are that it's still unclear
why your machine doesn't like large cardbus windows and therefore
how to fix that correctly.

May I ask you to do couple of another tests with following variations
of the latest patch (this will help to determine whether it's IO or
MEM ranges to blame)?
1. 
#define CARDBUS_IO_SIZE		(256)
#define CARDBUS_MEM_SIZE	(32*1024*1024)

2.
#define CARDBUS_IO_SIZE		(4096)
#define CARDBUS_MEM_SIZE	(4*1024*1024)

Ivan.

[SOLVED] Re: 2.6.13-rc2 hangs at boot

[Tero Roponen]

On Fri, 8 Jul 2005, Ivan Kokshaysky wrote:

Thanks! Vanilla 2.6.13-rc2 with below patch applied
works perfectly!

-
Tero Roponen


> On Thu, Jul 07, 2005 at 05:13:49PM +0300, Tero Roponen wrote:
> > I applied your original patch (the no-op one) and the
> > end=0 patch. With those applied I could boot into login
> > prompt.
>
> Puzzling. None of these patches should affect your setup.
> And you still have DMA timeouts...
>
> > Attached are lspci -vv and dmesg outputs from
> > 2.6.12 and 2.6.13-rc2 kernels.
>
> The only difference is that under 2.6.13-rc2 the cardbus ranges
> are a lot bigger. With the patch here your PCI setup should be
> identical to 2.6.12. I don't think this fixes DMA problem,
> but just to be sure...
>
> Ivan.
>
> --- 2.6.13-rc2/drivers/pci/setup-bus.c	Thu Jul  7 01:32:43 2005
> +++ linux/drivers/pci/setup-bus.c	Fri Jul  8 10:25:20 2005
> @@ -40,8 +40,8 @@
>   * FIXME: IO should be max 256 bytes.  However, since we may
>   * have a P2P bridge below a cardbus bridge, we need 4K.
>   */
> -#define CARDBUS_IO_SIZE		(4096)
> -#define CARDBUS_MEM_SIZE	(32*1024*1024)
> +#define CARDBUS_IO_SIZE		(256)
> +#define CARDBUS_MEM_SIZE	(4*1024*1024)
>
>  static void __devinit
>  pbus_assign_resources_sorted(struct pci_bus *bus)
>