Re: MPC5200 PCI Issues

From: Tobias Knutsson <hidden>
Date: 2009-02-06 12:39:55

I have now found drivers/pci/quirks.c which seems to be the place for
these types of things. However, since this code is unlikely to ever
make it to mainline, is it possible to register a fixup hook in a
platform setup file (lite5200.c and friends)?


On Fri, Feb 6, 2009 at 12:46, Tobias Knutsson [off-list ref] w=
rote:

Ok, I have had a breakthrough. I was trying to find out where the
resources were supposed to be allocated and that was when i found this
section of code in pbus_assign_resources_sorted() in
drivers/pci/setup-bus.c:

               /* Don't touch classless devices or host bridges or ioapic=

s.  */

               if (class =3D=3D PCI_CLASS_NOT_DEFINED ||
                   class =3D=3D PCI_CLASS_BRIDGE_HOST)
                       continue;

After that i took a quick look at the PCI-info reported by u-boot.
Sure enough, all boards were reported having class 0. At first i tried
to write a different class type, but as i suspected, the register
seems to be read-only. Therefore I simply commented out the first part
of the if-statement and now the resources seems to get allocated
properly.

However, I do realize that this is not the proper way of solving this
issue. Any ideas on how to fix it without modifying any kernel code?


On Thu, Feb 5, 2009 at 21:03, Tobias Knutsson [off-list ref]=

 wrote:

quoted

Hi,

That is exactly what the bootlog shows. Debug output on in both
mpc52xx_pci.c and in pci-common.c. And the discussed code section enable=

d.

quoted

H=E4lsningar / Regards
Tobias Knutsson

On 5 feb 2009, at 17.44, Grant Likely [off-list ref] wrote:

quoted

On Thu, Feb 5, 2009 at 12:06 AM, Tobias Knutsson
[off-list ref] wrote:

quoted

When enabling the debug-printouts, it came to me that I forgot to
mention something kind of important. Using the standard mpc52xx_pci.c,
even the memory resources are mapped to zero.

After disabling this code block, the memory regions are mapped out
fine (but still no IO):

Reenable this code, enable debug in arch/powerpc/kernel/pci-common.c,
and post the results.

When this code is disabled then you're just getting the leftovers from
U-Boot's scan of the PCI bus.  For MPC5200 support Linux does its own
scan and its own allocation of memory/IO address space and this code
is needed to make sure that happens.

g.

quoted

static void
mpc52xx_pci_fixup_resources(struct pci_dev *dev)
{
     int i;

     pr_debug("mpc52xx_pci_fixup_resources() %.4x:%.4x\n",
              dev->vendor, dev->device);

     /* We don't rely on boot loader for PCI and resets all
        devices */
     for (i =3D 0; i < DEVICE_COUNT_RESOURCE; i++) {
             struct resource *res =3D &dev->resource[i];
             if (res->end > res->start) {    /* Only valid resources *=

quoted

                     res->end -=3D res->start;
                     res->start =3D 0;
                     res->flags |=3D IORESOURCE_UNSET;
             }
     }

What is the purpose of this code? Since all resources are reset here,
where are they inteded to be re-allocated?

Here is the bootlog with the aforementioned code enabled:


reading cuImage.lite5200b

1833699 bytes read
## Booting kernel from Legacy Image at 00201400 ...
 Image Name:   Linux-2.6.28
 Created:      2009-02-05   6:54:20 UTC
 Image Type:   PowerPC Linux Kernel Image (gzip compressed)
 Data Size:    1833635 Bytes =3D  1.7 MB
 Load Address: 00400000
 Entry Point:  004005a0
 Verifying Checksum ... OK
 Uncompressing Kernel Image ... OK
[    0.000000] Using lite5200 machine description
[    0.000000] Linux version 2.6.28 (tknutsson@ztn-ubuntu) (gcc
version 4.3.2 (Sourcery G++ Lite 4.3-50) ) #66 Thu Feb 5 07:51:46 CET
2009
[    0.000000] Adding MPC52xx PCI host bridge /pci@f0000d00
[    0.000000] PCI host bridge /pci@f0000d00 (primary) ranges:
[    0.000000]  MEM 0x0000000050000000..0x000000005fffffff ->
0x0000000050000000
[    0.000000]   IO 0x0000000060000000..0x0000000060ffffff ->
0x0000000000000000
[    0.000000] mpc52xx_pci_setup(hose=3Dc03ec000, pci_regs=3Dfdffad00)
[    0.000000] mem_resource[0] =3D {.start=3D50000000, .end=3D5fffffff=

quoted

.flags=3D200}
[    0.000000] .io_resource=3D{.start=3D0,.end=3Dffffff,.flags=3D100}
.io_base_phys=3D0x60000000
[    0.000000] Zone PFN ranges:
[    0.000000]   DMA      0x00000000 -> 0x00004000
[    0.000000]   Normal   0x00004000 -> 0x00004000
[    0.000000] Movable zone start PFN for each node
[    0.000000] early_node_map[1] active PFN ranges
[    0.000000]     0: 0x00000000 -> 0x00004000
[    0.000000] Built 1 zonelists in Zone order, mobility grouping on.
Total pages: 16256
[    0.000000] Kernel command line: console=3DttyPSC0,115200 root=3D/d=

ev/sda2

quoted

rw
[    0.000000] MPC52xx PIC is up and running!
[    0.000000] PID hash table entries: 256 (order: 8, 1024 bytes)
[    0.000000] clocksource: timebase mult[7d00000] shift[22] registere=

quoted

[    0.000263] console [ttyPSC0] enabled
[    0.121690] Dentry cache hash table entries: 8192 (order: 3, 32768
bytes)
[    0.129012] Inode-cache hash table entries: 4096 (order: 2, 16384
bytes)
[    0.147433] Memory: 60844k/65536k available (3704k kernel code,
4624k reserved, 156k data, 117k bss, 180k init)
[    0.157672] SLUB: Genslabs=3D12, HWalign=3D32, Order=3D0-3, MinObje=

cts=3D0,

quoted

CPUs=3D1, Nodes=3D1
[    0.165373] Calibrating delay loop... 63.87 BogoMIPS (lpj=3D127744)
[    0.236265] Mount-cache hash table entries: 512
[    0.245719] net_namespace: 288 bytes
[    0.250237] NET: Registered protocol family 16
[    0.284887] PCI: Probing PCI hardware
[    0.289670] mpc52xx_pci_fixup_resources() 1057:5809
[    0.294799] PCI:0000:00:17.0 Resource 0
0000000050400000-00000000507fffff [21208] fixup...
[    0.303180] PCI:0000:00:17.0
 0000000050400000-00000000507fffff
[    0.309998] PCI:0000:00:17.0 Resource 1
0000000050800000-0000000050ffffff [20200] fixup...
[    0.318262] PCI:0000:00:17.0
 0000000050800000-0000000050ffffff
[    0.325062] PCI:0000:00:17.0 Resource 2
0000000060000000-000000006000000f [20101] fixup...
[    0.333326] PCI:0000:00:17.0
 0000000060000000-000000006000000f
[    0.340124] mpc52xx_pci_fixup_resources() 104c:9065
[    0.345242] PCI:0000:00:18.0 Resource 0
0000000051000000-00000000513fffff [21208] fixup...
[    0.353519] PCI:0000:00:18.0
 0000000051000000-00000000513fffff
[    0.360319] PCI:0000:00:18.0 Resource 1
0000000051800000-0000000051ffffff [20200] fixup...
[    0.368583] PCI:0000:00:18.0
 0000000051800000-0000000051ffffff
[    0.375383] PCI:0000:00:18.0 Resource 2
0000000060000010-000000006000001f [20101] fixup...
[    0.383647] PCI:0000:00:18.0
 0000000060000010-000000006000001f
[    0.390444] mpc52xx_pci_fixup_resources() 104c:9065
[    0.395549] PCI:0000:00:19.0 Resource 0
0000000052000000-00000000523fffff [21208] fixup...
[    0.403826] PCI:0000:00:19.0
 0000000052000000-00000000523fffff
[    0.410626] PCI:0000:00:19.0 Resource 1
0000000052800000-0000000052ffffff [20200] fixup...
[    0.418889] PCI:0000:00:19.0
 0000000052800000-0000000052ffffff
[    0.425690] PCI:0000:00:19.0 Resource 2
0000000060000020-000000006000002f [20101] fixup...
[    0.433954] PCI:0000:00:19.0
 0000000060000020-000000006000002f
[    0.440751] mpc52xx_pci_fixup_resources() 104c:9065
[    0.445854] PCI:0000:00:1a.0 Resource 0
0000000053000000-00000000533fffff [21208] fixup...
[    0.454130] PCI:0000:00:1a.0
 0000000053000000-00000000533fffff
[    0.460930] PCI:0000:00:1a.0 Resource 1
0000000053800000-0000000053ffffff [20200] fixup...
[    0.469194] PCI:0000:00:1a.0
 0000000053800000-0000000053ffffff
[    0.475994] PCI:0000:00:1a.0 Resource 2
0000000060000030-000000006000003f [20101] fixup...
[    0.484258] PCI:0000:00:1a.0
 0000000060000030-000000006000003f
[    0.491056] mpc52xx_pci_fixup_resources() 104c:9065
[    0.495999] PCI: Fixup bus 0 (PHB)
[    0.499383] PCI: device 0000:00:16.0 OF node: <none>
[    0.504371] PCI: device 0000:00:17.0 OF node: <none>
[    0.509363] PCI: device 0000:00:18.0 OF node: <none>
[    0.514354] PCI: device 0000:00:19.0 OF node: <none>
[    0.519345] PCI: device 0000:00:1a.0 OF node: <none>
[    0.524332] Try to map irq for 0000:00:16.0...
[    0.528808] Try to map irq for 0000:00:17.0...
[    0.533334]  -> got one, spec 3 cells (0x00000001 0x00000001...) on
/soc5200@f0000000/interrupt-controller@500
[    0.543389]  -> mapped to linux irq 65
[    0.547117] Try to map irq for 0000:00:18.0...
[    0.551626]  -> got one, spec 3 cells (0x00000001 0x00000001...) on
/soc5200@f0000000/interrupt-controller@500
[    0.561629]  -> mapped to linux irq 65
[    0.565353] Try to map irq for 0000:00:19.0...
[    0.569858]  -> got one, spec 3 cells (0x00000001 0x00000001...) on
/soc5200@f0000000/interrupt-controller@500
[    0.579863]  -> mapped to linux irq 65
[    0.583587] Try to map irq for 0000:00:1a.0...
[    0.588099]  -> got one, spec 3 cells (0x00000001 0x00000001...) on
/soc5200@f0000000/interrupt-controller@500
[    0.598103]  -> mapped to linux irq 65
[    0.605144] PCI: PHB (bus 0) bridge rsrc 0:
0000000000000000-0000000000ffffff [0x100], parent c03a1fdc (PCI IO)
[    0.615266] PCI: PHB (bus 0) bridge rsrc 1:
0000000050000000-000000005fffffff [0x200], parent c03a1fc0 (PCI mem)
[    0.625539] PCI: Assigning unassigned resouces...
[    0.630304] bus: 00 index 0 io port: [0x00-0xffffff]
[    0.635299] bus: 00 index 1 mmio: [0x50000000-0x5fffffff]
[    0.640909] DMA: MPC52xx BestComm driver
[    0.645214] DMA: MPC52xx BestComm engine @f0001200 ok !
[    0.678739] SCSI subsystem initialized
[    0.709498] NET: Registered protocol family 2
[    0.754980] IP route cache hash table entries: 1024 (order: 0, 4096
bytes)
[    0.762429] TCP established hash table entries: 2048 (order: 2, 163=

quoted

bytes)
[    0.769716] TCP bind hash table entries: 2048 (order: 1, 8192 bytes=

quoted

[    0.776224] TCP: Hash tables configured (established 2048 bind 2048=

quoted

[    0.782604] TCP reno registered
[    0.795054] NET: Registered protocol family 1
[    0.846171] JFFS2 version 2.2. (NAND) (c) 2001-2006 Red Hat, Inc.
[    0.855699] msgmni has been set to 118
[    0.861227] alg: No test for stdrng (krng)
[    0.865651] io scheduler noop registered
[    0.869658] io scheduler anticipatory registered (default)
[    0.875224] io scheduler deadline registered
[    0.879838] io scheduler cfq registered
[    1.442231] Generic RTC Driver v1.07
[    1.447145] Serial: MPC52xx PSC UART driver
[    1.451885] f0002200.serial: ttyPSC2 at MMIO 0xf0002200 (irq =3D 13=

0)

quoted

is a MPC52xx PSC
[    1.461497] f0002400.serial: ttyPSC3 at MMIO 0xf0002400 (irq =3D 13=

1)

quoted

is a MPC52xx PSC
[    1.471257] f0002600.serial: ttyPSC4 at MMIO 0xf0002600 (irq =3D 13=

9)

quoted

is a MPC52xx PSC
[    1.480866] f0002800.serial: ttyPSC5 at MMIO 0xf0002800 (irq =3D 14=

0)

quoted

is a MPC52xx PSC
[    1.490461] f0002c00.serial: ttyPSC0 at MMIO 0xf0002c00 (irq =3D 13=

2)

quoted

is a MPC52xx PSC
[    1.521689] brd: module loaded
[    1.535886] loop: module loaded
[    1.544328] mpc52xx MII bus: probed
[    1.550220] net eth0: Using PHY at MDIO address 0
[    1.556124] Driver 'sd' needs updating - please use bus_type method=

quoted

[    1.563224] ata: MPC52xx IDE/ATA libata driver
[    1.568729] scsi0 : mpc52xx_ata
[    1.573032] ata1: PATA max UDMA/33 ata_regs 0xf0003a00 irq 135
[    1.751224] ata1.00: ATA-0: TRANSCEND, 20070831, max UDMA/66
[    1.756948] ata1.00: 3915072 sectors, multi 0: LBA
[    1.775156] ata1.00: configured for UDMA/33
[    1.803154] ata1.00: configured for UDMA/33
[    1.807423] ata1: EH complete
[    1.811300] scsi 0:0:0:0: Direct-Access     ATA      TRANSCEND
2007 PQ: 0 ANSI: 5
[    1.821337] sd 0:0:0:0: [sda] 3915072 512-byte hardware sectors:
(2.00 GB/1.86 GiB)
[    1.829333] sd 0:0:0:0: [sda] Write Protect is off
[    1.834879] sd 0:0:0:0: [sda] Write cache: disabled, read cache:
enabled, doesn't support DPO or FUA
[    1.844851] sd 0:0:0:0: [sda] 3915072 512-byte hardware sectors:
(2.00 GB/1.86 GiB)
[    1.852841] sd 0:0:0:0: [sda] Write Protect is off
[    1.858259] sd 0:0:0:0: [sda] Write cache: disabled, read cache:
enabled, doesn't support DPO or FUA
[    1.867450]  sda: sda1 sda2
[    1.881490] sd 0:0:0:0: [sda] Attached SCSI disk
[    1.890711] fc000000.flash: Found 1 x16 devices at 0x0 in 8-bit ban=

quoted

[    1.897222]  Amd/Fujitsu Extended Query Table at 0x0040
[    1.902529] fc000000.flash: CFI does not contain boot bank
location. Assuming top.
[    1.910147] number of CFI chips: 1
[    1.913603] cfi_cmdset_0002: Disabling erase-suspend-program due to
code brokenness.
[    1.922009] RedBoot partition parsing not available
[    1.926998] Creating 4 MTD partitions on "fc000000.flash":
[    1.932547] 0x00000000-0x00500000 : "u-boot"
[    1.940601] 0x00500000-0x01000000 : "kernel"
[    1.948071] 0x01000000-0x02000000 : "rootfs"
[    1.955601] 0x02000000-0x03000000 : "application"
[    1.964240] i2c /dev entries driver
[    1.975775] Advanced Linux Sound Architecture Driver Version
1.0.18rc3.
[    1.983814] ASoC version 0.13.2
[    1.987668] tlv320aic23 0-001a: probing tlv320aic23 i2c device
[    2.040451] registering ASoC codec driver:
/soc5200@f0000000/i2c@3d00/codec@1a
[    2.049539] ALSA device list:
[    2.052591]   No soundcards found.
[    2.057430] TCP cubic registered
[    2.060765] NET: Registered protocol family 17
[    2.066886] RPC: Registered udp transport module.
[    2.071683] RPC: Registered tcp transport module.
[    2.087086] EXT2-fs warning: mounting unchecked fs, running e2fsck
is recommended
[    2.094828] VFS: Mounted root (ext2 filesystem).
[    2.099690] Freeing unused kernel memory: 180k init

On Wed, Feb 4, 2009 at 18:54, Grant Likely [off-list ref]
wrote:

quoted

On Wed, Feb 4, 2009 at 9:53 AM, Tobias Knutsson
[off-list ref] wrote:

quoted

            clock-frequency =3D <0x1e84800>;  //<0>; // From boot lo=

ader

quoted

            interrupts =3D <2 8 0 2 9 0 2 10 0>;
            interrupt-parent =3D <&mpc5200_pic>;
            bus-range =3D <0 0>;
            ranges =3D <0x02000000 0 0x50000000 0x50000000 0 0x10000=

quoted

                      0x01000000 0 0x00000000 0x60000000 0 0x0100000=

0>;

quoted

    };

The cards' memory regions are mapped out fine to different addresses
in the 0x50000000-0x60000000 range. However, the IO regions are all
mapped
to zero.

Do you have any idea on what could be wrong or in which direction I
should be looking?

Hmmm.  Your ranges property looks okay to me.  The MPC5200 setup code
may be doing the wrong thing.  Turn on debug at the top of
arch/powerpc/platforms/mpc52xx/mpc52xx_pci.c and post your kernel
output.

g.

--
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.



--

H=E4lsningar/Regards
Tobias Knutsson



--
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.



--

H=E4lsningar/Regards
Tobias Knutsson



--=20

H=E4lsningar/Regards
Tobias Knutsson

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help