Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

[Colin Watson]

On Sun, Jan 02, 2011 at 10:41:50PM +0000, Steve McIntyre wrote:
> 1.iso:
> 
>   last bus number printed is b0
> 
> 2.iso:
> 
>   goes all the way through to bus ff and returns to a grub prompt

This is interesting and suggests a measure of coincidence.  What that
patch did was skip remaining functions on a device that doesn't
implement function 0, taking that as an indication that it doesn't
exist.  This was based on:

  http://en.wikipedia.org/wiki/PCI_configuration_space#Bus_enumeration

Vladimir, are you OK with this change to trunk?

2011-01-02  Colin Watson  <cjwatson@ubuntu.com>

	* grub-core/bus/pci.c (grub_pci_iterate): Skip remaining functions
	on devices that do not implement function 0.

=== modified file 'grub-core/bus/pci.c'
--- grub-core/bus/pci.c	2010-06-30 00:30:05 +0000
+++ grub-core/bus/pci.c	2011-01-02 17:31:32 +0000
@@ -90,7 +90,14 @@ grub_pci_iterate (grub_pci_iteratefunc_t
 
 	      /* Check if there is a device present.  */
 	      if (id >> 16 == 0xFFFF)
-		continue;
+		{
+		  if (dev.function == 0)
+		    /* Devices are required to implement function 0, so if
+		       it's missing then there is no device here.  */
+		    break;
+		  else
+		    continue;
+		}
 
 #ifdef GRUB_MACHINE_MIPS_YEELOONG
 	      /* Skip ghosts.  */

Nevertheless, I'm not confident that this will fix the problem on all
machines, so I would like to sort out the bridge handling as well.

> 3.iso: 
> 
>   grub> set debug=pci
>   grub> lspci
>   bus/pci.c:92: bus 0
>   00:00.0 8086:8100 [0600] Host Bridge
>   00:02.0 8086:8108 [0300] VGA Controller
>   00:1b.0 8086:811b [0403] Multimedia device
>   00:1c.0 8086:8110 [0604] PCI-PCI Bridge
>   00:1c.1 8086:8112 [0604] PCI-PCI Bridge
>   00:1d.0 8086:8114 [0c03] USB Controller
>   00:1d.1 8086:8115 [0c03] USB Controller
>   00:1d.2 8086:8116 [0c03] USB Controller
>   00:1d.7 8086:8117 [0c03] USB Controller [PI 20]
>   00:1f.0 8086:8119 [0601] ISA Bridge
>   00:1f.1 8086:811a [0101] IDE Controller [PI 80]
>   grub>

Whoops, I forgot to right-shift the header word.  Can you try 4.iso
instead, at the same location?  I also made it handle PCI-to-CardBus
bridges the same way as PCI-to-PCI bridges since that's what pciutils
does.

(In addition to 'set debug=pci', I'd recommend also doing 'set pager=1'
so that lspci's output will be paged.)

Thanks,

-- 
Colin Watson                                       [cjwatson@debian.org]

Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

[Steve McIntyre]

On Mon, Jan 03, 2011 at 12:13:01AM +0000, Colin Watson wrote:
>
>Whoops, I forgot to right-shift the header word.  Can you try 4.iso
>instead, at the same location?  I also made it handle PCI-to-CardBus
>bridges the same way as PCI-to-PCI bridges since that's what pciutils
>does.
>
>(In addition to 'set debug=pci', I'd recommend also doing 'set pager=1'
>so that lspci's output will be paged.)

4.iso:

  grub> set debug=pci
  grub> set pager=1
  grub> lspci
  bus/pci.c:92: bus 0x0
  00:00.0 8086:8100 [0600] Host Bridge
  00:02.0 8086:8108 [0300] VGA Controller
  00:1b.0 8086:811b [0403] Multimedia device
  bus/pci.c:143: bridge range 0x2-0x2
  00:1c.0 8086:8110 [0604] PCI-PCI Bridge
  bus/pci.c:143: bridge range 0x3-0x3
  00:1c.1 8086:8112 [0604] PCI-PCI Bridge
  00:1d.0 8086:8114 [0c03] USB Controller
  00:1d.1 8086:8115 [0c03] USB Controller
  00:1d.2 8086:8116 [0c03] USB Controller
  00:1d.7 8086:8117 [0c03] USB Controller [PI 20]
  00:1f.0 8086:8119 [0601] ISA Bridge
  00:1f.1 8086:811a [0101] IDE Controller [PI 80]
  bus/pci.c:92: bus 0x2
  02:00.0 10ec:8136 [0200] Ethernet Controller
  bus/pci.c:92: bus 0x3
  03:00.0 168c:001c [0200] Ethernet Controller
  grub>


-- 
Steve McIntyre, Cambridge, UK.                                steve@einval.com
"Because heaters aren't purple!" -- Catherine Pitt

Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

[Colin Watson]

On Mon, Jan 03, 2011 at 11:04:17AM +0000, Steve McIntyre wrote:
> On Mon, Jan 03, 2011 at 12:13:01AM +0000, Colin Watson wrote:
> >Whoops, I forgot to right-shift the header word.  Can you try 4.iso
> >instead, at the same location?  I also made it handle PCI-to-CardBus
> >bridges the same way as PCI-to-PCI bridges since that's what pciutils
> >does.
> >
> >(In addition to 'set debug=pci', I'd recommend also doing 'set pager=1'
> >so that lspci's output will be paged.)
> 
> 4.iso:
> 
>   grub> set debug=pci
>   grub> set pager=1
>   grub> lspci
>   bus/pci.c:92: bus 0x0
>   00:00.0 8086:8100 [0600] Host Bridge
>   00:02.0 8086:8108 [0300] VGA Controller
>   00:1b.0 8086:811b [0403] Multimedia device
>   bus/pci.c:143: bridge range 0x2-0x2
>   00:1c.0 8086:8110 [0604] PCI-PCI Bridge
>   bus/pci.c:143: bridge range 0x3-0x3
>   00:1c.1 8086:8112 [0604] PCI-PCI Bridge
>   00:1d.0 8086:8114 [0c03] USB Controller
>   00:1d.1 8086:8115 [0c03] USB Controller
>   00:1d.2 8086:8116 [0c03] USB Controller
>   00:1d.7 8086:8117 [0c03] USB Controller [PI 20]
>   00:1f.0 8086:8119 [0601] ISA Bridge
>   00:1f.1 8086:811a [0101] IDE Controller [PI 80]
>   bus/pci.c:92: bus 0x2
>   02:00.0 10ec:8136 [0200] Ethernet Controller
>   bus/pci.c:92: bus 0x3
>   03:00.0 168c:001c [0200] Ethernet Controller
>   grub>

This looks right to me.  Excellent.  Vladimir, how does this patch look,
on top of my previous one?  (This is edited slightly relative to what
Steve tested, to make sure that the bus number never overflows
bus_present; this will never happen on x86 but might happen on other
architectures.  I've smoke-tested this change.)

2011-01-03  Colin Watson  <cjwatson@ubuntu.com>

	* grub-core/bus/pci.c (grub_pci_iterate): Only scan bus 0 plus any
	buses linked by PCI-to-PCI or PCI-to-CardBus bridges.
	* include/grub/pci.h: Add definitions for bridges.

=== modified file 'grub-core/bus/pci.c'
--- grub-core/bus/pci.c	2010-06-30 00:30:05 +0000
+++ grub-core/bus/pci.c	2011-01-03 00:05:27 +0000
@@ -20,6 +20,7 @@
 #include <grub/dl.h>
 #include <grub/pci.h>
 #include <grub/mm.h>
+#include <grub/misc.h>
 
 /* FIXME: correctly support 64-bit architectures.  */
 /* #if GRUB_TARGET_SIZEOF_VOID_P == 4 */
@@ -78,9 +79,18 @@ grub_pci_iterate (grub_pci_iteratefunc_t
   grub_pci_address_t addr;
   grub_pci_id_t id;
   grub_uint32_t hdr;
+  grub_uint8_t bus_present[(GRUB_PCI_NUM_BUS + 7) / 8];
+
+  grub_memset (bus_present, 0, sizeof (bus_present));
+  bus_present[0] = 1; /* bus 0 is always enabled */
 
   for (dev.bus = 0; dev.bus < GRUB_PCI_NUM_BUS; dev.bus++)
     {
+      if (!(bus_present[dev.bus / 8] & (1 << (dev.bus % 8))))
+	continue;
+
+      grub_dprintf ("pci", "bus 0x%x\n", dev.bus);
+
       for (dev.device = 0; dev.device < GRUB_PCI_NUM_DEVICES; dev.device++)
 	{
 	  for (dev.function = 0; dev.function < 8; dev.function++)
@@ -112,6 +119,38 @@ grub_pci_iterate (grub_pci_iteratefunc_t
 		continue;
 #endif
 
+	      /* On bus 0, look for PCI-to-PCI bridges and mark all buses
+	         within their ranges as present.  */
+	      if (dev.bus == 0)
+		{
+		  addr = grub_pci_make_address (dev, GRUB_PCI_REG_CACHELINE);
+		  hdr = grub_pci_read (addr);
+
+		  switch ((hdr >> 16) & 0x7F) {
+		    case GRUB_PCI_HEADER_PCI_BRIDGE:
+		    case GRUB_PCI_HEADER_CARDBUS_BRIDGE:
+		      {
+			grub_uint32_t bus_numbers;
+			grub_uint32_t secondary, subordinate, i;
+
+			addr = grub_pci_make_address
+			  (dev, GRUB_PCI_REG_SEC_LAT_TIMER);
+			bus_numbers = grub_pci_read (addr);
+			secondary = (bus_numbers >> 8) & 0xFF;
+			subordinate = (bus_numbers >> 16) & 0xFF;
+
+			grub_dprintf ("pci", "bridge range 0x%x-0x%x\n",
+				      secondary, subordinate);
+
+			for (i = secondary;
+			     i <= subordinate && i < GRUB_PCI_NUM_BUS; i++)
+			  bus_present[i / 8] |= (1 << (i % 8));
+
+			break;
+		      }
+		  }
+		}
+
 	      if (hook (dev, id))
 		return;
 

=== modified file 'include/grub/pci.h'
--- include/grub/pci.h	2010-08-11 02:18:07 +0000
+++ include/grub/pci.h	2011-01-02 17:32:28 +0000
@@ -68,6 +68,24 @@
 #define  GRUB_PCI_REG_MIN_GNT      0x3e
 #define  GRUB_PCI_REG_MAX_LAT      0x3f
 
+/* Alternative register meanings if header type is 1 (PCI-to-PCI bridge).  */
+#define  GRUB_PCI_REG_SEC_LAT_TIMER     0x18
+#define  GRUB_PCI_REG_SUB_BUS_NUMBER    0x19
+#define  GRUB_PCI_REG_SEC_BUS_NUMBER    0x1a
+#define  GRUB_PCI_REG_PRI_BUS_NUMBER    0x1b
+#define  GRUB_PCI_REG_SEC_STATUS        0x1c
+#define  GRUB_PCI_REG_IO_LIMIT          0x1e
+#define  GRUB_PCI_REG_IO_BASE           0x1f
+#define  GRUB_PCI_REG_MEM_LIMIT         0x20
+#define  GRUB_PCI_REG_MEM_BASE          0x22
+#define  GRUB_PCI_REG_PF_MEM_LIMIT      0x24
+#define  GRUB_PCI_REG_PF_MEM_BASE       0x26
+#define  GRUB_PCI_REG_PF_MEM_BASE_HI    0x28
+#define  GRUB_PCI_REG_PF_MEM_LIMIT_HI   0x2c
+#define  GRUB_PCI_REG_IO_LIMIT_HI       0x30
+#define  GRUB_PCI_REG_IO_BASE_HI        0x32
+#define  GRUB_PCI_REG_BRIDGE_CONTROL    0x3c
+
 #define  GRUB_PCI_COMMAND_IO_ENABLED    0x0001
 #define  GRUB_PCI_COMMAND_MEM_ENABLED   0x0002
 #define  GRUB_PCI_COMMAND_BUS_MASTER    0x0004
@@ -82,6 +100,10 @@
 #define  GRUB_PCI_STATUS_DEVSEL_TIMING_MASK 0x0600
 #define  GRUB_PCI_CLASS_SUBCLASS_VGA  0x0300
 
+#define  GRUB_PCI_HEADER_GENERAL        0x0
+#define  GRUB_PCI_HEADER_PCI_BRIDGE     0x1
+#define  GRUB_PCI_HEADER_CARDBUS_BRIDGE 0x2
+
 #ifndef ASM_FILE
 typedef grub_uint32_t grub_pci_id_t;
 

Thanks,

-- 
Colin Watson                                       [cjwatson@debian.org]

Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

[Colin Watson]

On Mon, Jan 03, 2011 at 12:13:01AM +0000, Colin Watson wrote:
> On Sun, Jan 02, 2011 at 10:41:50PM +0000, Steve McIntyre wrote:
> > 2.iso:
> > 
> >   goes all the way through to bus ff and returns to a grub prompt
> 
> This is interesting and suggests a measure of coincidence.  What that
> patch did was skip remaining functions on a device that doesn't
> implement function 0, taking that as an indication that it doesn't
> exist.  This was based on:
> 
>   http://en.wikipedia.org/wiki/PCI_configuration_space#Bus_enumeration
> 
> Vladimir, are you OK with this change to trunk?
> 
> 2011-01-02  Colin Watson  <cjwatson@ubuntu.com>
> 
> 	* grub-core/bus/pci.c (grub_pci_iterate): Skip remaining functions
> 	on devices that do not implement function 0.

I've applied this patch to trunk following an ack from Vladimir on IRC.
I'll prepare an updated package for unstable shortly.

> Nevertheless, I'm not confident that this will fix the problem on all
> machines, so I would like to sort out the bridge handling as well.

This may be more complicated than I thought.  Seth Goldberg pointed out
that my approach fails to deal with peer host bridges correctly (i.e.
cases where there are multiple trees, not just a single one rooted at
bus 0).  Linux deals with this by asking the PCI BIOS for the last bus
number, but at this point things get complicated as you have to do
things in different ways for different firmware.

I am inclined to try the first piece alone and see how this works out,
and if we can fix the affected systems by just probing function 0 on
every device on every bus then let it stand at that, even if it feels
less elegant.  Inventing new piles of infrastructure to handle a case
I'm unsure about in a subsystem I don't know well isn't my idea of a
good time.

-- 
Colin Watson                                       [cjwatson@debian.org]

Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

[Seth Goldberg]

On Jan 3, 2011, at 4:06 PM, Colin Watson <cjwatson@debian.org> wrote:

> On Mon, Jan 03, 2011 at 12:13:01AM +0000, Colin Watson wrote:
>> On Sun, Jan 02, 2011 at 10:41:50PM +0000, Steve McIntyre wrote:
>>> 2.iso:
>>> 
>>>  goes all the way through to bus ff and returns to a grub prompt
>> 
>> This is interesting and suggests a measure of coincidence.  What that
>> patch did was skip remaining functions on a device that doesn't
>> implement function 0, taking that as an indication that it doesn't
>> exist.  This was based on:
>> 
>>  http://en.wikipedia.org/wiki/PCI_configuration_space#Bus_enumeration
>> 
>> Vladimir, are you OK with this change to trunk?
>> 
>> 2011-01-02  Colin Watson  <cjwatson@ubuntu.com>
>> 
>>    * grub-core/bus/pci.c (grub_pci_iterate): Skip remaining functions
>>    on devices that do not implement function 0.
> 
> I've applied this patch to trunk following an ack from Vladimir on IRC.
> I'll prepare an updated package for unstable shortly.
> 
>> Nevertheless, I'm not confident that this will fix the problem on all
>> machines, so I would like to sort out the bridge handling as well.
> 
> This may be more complicated than I thought.  Seth Goldberg pointed out
> that my approach fails to deal with peer host bridges correctly (i.e.
> cases where there are multiple trees, not just a single one rooted at
> bus 0).  Linux deals with this by asking the PCI BIOS for the last bus
> number, but at this point things get complicated as you have to do
> things in different ways for different firmware.
> 

  The proper way to do this on modern systems is to traverse the system's [DSDT/SSDT] ACPI tables looking for Device objects with the host bridge HID/CID and evaluate the BBN object (which can be a method), if it exists (which it must if there are multiple host bridges).  Since grub2 does not have a full ACPI interpreter (pulling in Intel's acpica would work ;), though the license may force it to be a grub-extra), going that route with anything less would never cover all systems' BBNs, so PCI BIOS would be simplest.  Things get a bit more complicated when a system has multiple PCI segments (i.e.: using the MCFG table, MMIO addresses that may be >4G, etc.), but that can be tackled later.

  --S



> I am inclined to try the first piece alone and see how this works out,
> and if we can fix the affected systems by just probing function 0 on
> every device on every bus then let it stand at that, even if it feels
> less elegant.  Inventing new piles of infrastructure to handle a case
> I'm unsure about in a subsystem I don't know well isn't my idea of a
> good time.
> 
> -- 
> Colin Watson                                       [cjwatson@debian.org]
> 
> _______________________________________________
> Grub-devel mailing list
> Grub-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/grub-devel

Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

[Colin Watson]

On Tue, Jan 04, 2011 at 12:06:54AM +0000, Colin Watson wrote:
> On Mon, Jan 03, 2011 at 12:13:01AM +0000, Colin Watson wrote:
> > 2011-01-02  Colin Watson  <cjwatson@ubuntu.com>
> > 
> > 	* grub-core/bus/pci.c (grub_pci_iterate): Skip remaining functions
> > 	on devices that do not implement function 0.
> 
> I've applied this patch to trunk following an ack from Vladimir on IRC.
> I'll prepare an updated package for unstable shortly.

Uploading now.  I'd appreciate confirmation from affected folks that
this is enough to make things boot.  If it is, we can perhaps avoid
worrying about the rest; otherwise, we'll have to get more creative.

Thanks,

-- 
Colin Watson                                       [cjwatson@debian.org]

Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

[Steve McIntyre]

On Tue, Jan 04, 2011 at 12:46:54PM +0000, Colin Watson wrote:
>On Tue, Jan 04, 2011 at 12:06:54AM +0000, Colin Watson wrote:
>> On Mon, Jan 03, 2011 at 12:13:01AM +0000, Colin Watson wrote:
>> > 2011-01-02  Colin Watson  <cjwatson@ubuntu.com>
>> > 
>> > 	* grub-core/bus/pci.c (grub_pci_iterate): Skip remaining functions
>> > 	on devices that do not implement function 0.
>> 
>> I've applied this patch to trunk following an ack from Vladimir on IRC.
>> I'll prepare an updated package for unstable shortly.
>
>Uploading now.  I'd appreciate confirmation from affected folks that
>this is enough to make things boot.  If it is, we can perhaps avoid
>worrying about the rest; otherwise, we'll have to get more creative.

Just tested on Jo's laptop now, and the new version works exactly as
hoped. Thanks for the quick fix! Jo says "many thankyous" too. :-)

-- 
Steve McIntyre, Cambridge, UK.                                steve@einval.com
"Since phone messaging became popular, the young generation has lost the
 ability to read or write anything that is longer than one hundred and sixty
 characters."  -- Ignatios Souvatzis

Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

[Matt Zagrabelny]

On Tue, Jan 4, 2011 at 6:46 AM, Colin Watson <cjwatson@debian.org> wrote:
> On Tue, Jan 04, 2011 at 12:06:54AM +0000, Colin Watson wrote:
>> On Mon, Jan 03, 2011 at 12:13:01AM +0000, Colin Watson wrote:
>> > 2011-01-02  Colin Watson  <cjwatson@ubuntu.com>
>> >
>> >     * grub-core/bus/pci.c (grub_pci_iterate): Skip remaining functions
>> >     on devices that do not implement function 0.
>>
>> I've applied this patch to trunk following an ack from Vladimir on IRC.
>> I'll prepare an updated package for unstable shortly.
>
> Uploading now.  I'd appreciate confirmation from affected folks that
> this is enough to make things boot.  If it is, we can perhaps avoid
> worrying about the rest; otherwise, we'll have to get more creative.

"Graphical" booting is back. Thanks Colin.

-matt

Re: Bug#594967: Bug #594967: [poulsbo] grub-pc Hangs After "Welcome to GRUB!"

[P. J. McDermott]

On 01/04/2011 07:46 AM, Colin Watson wrote:
> On Tue, Jan 04, 2011 at 12:06:54AM +0000, Colin Watson wrote:
>   
>> On Mon, Jan 03, 2011 at 12:13:01AM +0000, Colin Watson wrote:
>>     
>>> 2011-01-02  Colin Watson  <cjwatson@ubuntu.com>
>>>
>>> 	* grub-core/bus/pci.c (grub_pci_iterate): Skip remaining functions
>>> 	on devices that do not implement function 0.
>>>       
>> I've applied this patch to trunk following an ack from Vladimir on IRC.
>> I'll prepare an updated package for unstable shortly.
>>     
> Uploading now.  I'd appreciate confirmation from affected folks that
> this is enough to make things boot.  If it is, we can perhaps avoid
> worrying about the rest; otherwise, we'll have to get more creative.
>
> Thanks,
>   

Sorry I haven't responded to this more quickly; I was busy with some
personal business last week. Thanks for the patch, and it does make GRUB
2's graphical menu load on my AO751h. :)

I noticed an odd lingering aesthetics issue though; most of the
background image is missing. The only parts of Squeeze's
`spacefun-grub.png' that are visible are the bottom of Earth and the two
stars on the top-right corner (basically, anything along the edges).
I've consistently reproduced this on two installations (one is a fresh
installation on a flash drive), even after purging and reinstalling
grub-pc and grub-common. The odd thing though is that it only happens
with desktop-base 6.0.5. The `spacefun-grub.png' from version 6.0.2 is
displayed in full.

P. J.