Re: [PATCH 03/13] dm: pci: Skip setting VGA bridge bits if parent device is the host bus
On Tue, 13 Apr 2021 at 16:23, Masami Hiramatsu masami.hiramatsu@linaro.org wrote:
Commit bbbcb5262839 ("dm: pci: Enable VGA address forwarding on bridges") sets the VGA bridge bits by checking pplat->class, but if the parent device is the pci host bus device, it can be skipped. Moreover, it shouldn't access the pplat because the parent has different plat data.
Without this fix, "pci enum" command cause a synchronous abort.
pci_auto_config_devices: start PCI Autoconfig: Bus Memory region: [78000000-7fffffff], Physical Memory [78000000-7fffffffx] PCI Autoconfig: Bus I/O region: [0-ffff], Physical Memory [77f00000-77f0ffffx] pci_auto_config_devices: device pci_6:0.0 PCI Autoconfig: BAR 0, Mem, size=0x1000000, address=0x78000000 bus_lower=0x79000000
PCI Autoconfig: BAR 1, Mem, size=0x8000000, No room in resource, avail start=79000000 / size=8000000, need=8000000 PCI: Failed autoconfig bar 14
PCI Autoconfig: BAR 2, I/O, size=0x4, address=0x1000 bus_lower=0x1004
PCI Autoconfig: BAR 3, Mem, size=0x2000000, address=0x7a000000 bus_lower=0x7c000000
PCI Autoconfig: BAR 4, I/O, size=0x80, address=0x1080 bus_lower=0x1100
PCI Autoconfig: ROM, size=0x80000, address=0x7c000000 bus_lower=0x7c080000
"Synchronous Abort" handler, esr 0x96000006 elr: 00000000e002bd28 lr : 00000000e002bce8 (reloc) elr: 00000000fff6fd28 lr : 00000000fff6fce8 x0 : 0000000000001041 x1 : 000000000000003e x2 : 00000000ffb0f8c8 x3 : 0000000000000001 x4 : 0000000000000080 x5 : 0000000000000000 x6 : 00000000fff718fc x7 : 000000000000000f x8 : 00000000ffb0f238 x9 : 0000000000000008 x10: 0000000000000000 x11: 0000000000000010 x12: 0000000000000006 x13: 000000000001869f x14: 00000000ffb0fcd0 x15: 0000000000000020 x16: 00000000fff71cc4 x17: 0000000000000000 x18: 00000000ffb13d90 x19: 00000000ffb14320 x20: 0000000000000000 x21: 00000000ffb14090 x22: 00000000ffb0f8c8 x23: 0000000000000001 x24: 00000000ffb14c10 x25: 0000000000000000 x26: 0000000000000000 x27: 0000000000000000 x28: 00000000ffb14c70 x29: 00000000ffb0f830
Code: 52800843 52800061 52800e00 97ffcf65 (b9400280) Resetting CPU ...
Signed-off-by: Masami Hiramatsu masami.hiramatsu@linaro.org
drivers/pci/pci-uclass.c | 3 +++ 1 file changed, 3 insertions(+)
Reviewed-by: Simon Glass sjg@chromium.org
Hi,
On Thu, Apr 15, 2021 at 3:39 AM Simon Glass sjg@chromium.org wrote:
On Tue, 13 Apr 2021 at 16:23, Masami Hiramatsu masami.hiramatsu@linaro.org wrote:
Commit bbbcb5262839 ("dm: pci: Enable VGA address forwarding on bridges") sets the VGA bridge bits by checking pplat->class, but if the parent device is the pci host bus device, it can be skipped. Moreover, it shouldn't access the pplat because the parent has different plat data.
Without this fix, "pci enum" command cause a synchronous abort.
pci_auto_config_devices: start PCI Autoconfig: Bus Memory region: [78000000-7fffffff], Physical Memory [78000000-7fffffffx] PCI Autoconfig: Bus I/O region: [0-ffff], Physical Memory [77f00000-77f0ffffx] pci_auto_config_devices: device pci_6:0.0 PCI Autoconfig: BAR 0, Mem, size=0x1000000, address=0x78000000 bus_lower=0x79000000
PCI Autoconfig: BAR 1, Mem, size=0x8000000, No room in resource, avail start=79000000 / size=8000000, need=8000000 PCI: Failed autoconfig bar 14
PCI Autoconfig: BAR 2, I/O, size=0x4, address=0x1000 bus_lower=0x1004
PCI Autoconfig: BAR 3, Mem, size=0x2000000, address=0x7a000000 bus_lower=0x7c000000
PCI Autoconfig: BAR 4, I/O, size=0x80, address=0x1080 bus_lower=0x1100
PCI Autoconfig: ROM, size=0x80000, address=0x7c000000 bus_lower=0x7c080000
"Synchronous Abort" handler, esr 0x96000006 elr: 00000000e002bd28 lr : 00000000e002bce8 (reloc) elr: 00000000fff6fd28 lr : 00000000fff6fce8 x0 : 0000000000001041 x1 : 000000000000003e x2 : 00000000ffb0f8c8 x3 : 0000000000000001 x4 : 0000000000000080 x5 : 0000000000000000 x6 : 00000000fff718fc x7 : 000000000000000f x8 : 00000000ffb0f238 x9 : 0000000000000008 x10: 0000000000000000 x11: 0000000000000010 x12: 0000000000000006 x13: 000000000001869f x14: 00000000ffb0fcd0 x15: 0000000000000020 x16: 00000000fff71cc4 x17: 0000000000000000 x18: 00000000ffb13d90 x19: 00000000ffb14320 x20: 0000000000000000 x21: 00000000ffb14090 x22: 00000000ffb0f8c8 x23: 0000000000000001 x24: 00000000ffb14c10 x25: 0000000000000000 x26: 0000000000000000 x27: 0000000000000000 x28: 00000000ffb14c70 x29: 00000000ffb0f830
Code: 52800843 52800061 52800e00 97ffcf65 (b9400280) Resetting CPU ...
Signed-off-by: Masami Hiramatsu masami.hiramatsu@linaro.org
drivers/pci/pci-uclass.c | 3 +++ 1 file changed, 3 insertions(+)
Reviewed-by: Simon Glass sjg@chromium.org
I can't find this patch in my inbox, nor in the patchwork.
Could you please resend?
Regards, Bin
Hi Bin,
On Thu, 15 Apr 2021 06:30:27 +0800 Bin Meng bmeng.cn@gmail.com wrote:
Hi,
On Thu, Apr 15, 2021 at 3:39 AM Simon Glass sjg@chromium.org wrote:
On Tue, 13 Apr 2021 at 16:23, Masami Hiramatsu masami.hiramatsu@linaro.org wrote:
Commit bbbcb5262839 ("dm: pci: Enable VGA address forwarding on bridges") sets the VGA bridge bits by checking pplat->class, but if the parent device is the pci host bus device, it can be skipped. Moreover, it shouldn't access the pplat because the parent has different plat data.
Without this fix, "pci enum" command cause a synchronous abort.
pci_auto_config_devices: start PCI Autoconfig: Bus Memory region: [78000000-7fffffff], Physical Memory [78000000-7fffffffx] PCI Autoconfig: Bus I/O region: [0-ffff], Physical Memory [77f00000-77f0ffffx] pci_auto_config_devices: device pci_6:0.0 PCI Autoconfig: BAR 0, Mem, size=0x1000000, address=0x78000000 bus_lower=0x79000000
PCI Autoconfig: BAR 1, Mem, size=0x8000000, No room in resource, avail start=79000000 / size=8000000, need=8000000 PCI: Failed autoconfig bar 14
PCI Autoconfig: BAR 2, I/O, size=0x4, address=0x1000 bus_lower=0x1004
PCI Autoconfig: BAR 3, Mem, size=0x2000000, address=0x7a000000 bus_lower=0x7c000000
PCI Autoconfig: BAR 4, I/O, size=0x80, address=0x1080 bus_lower=0x1100
PCI Autoconfig: ROM, size=0x80000, address=0x7c000000 bus_lower=0x7c080000
"Synchronous Abort" handler, esr 0x96000006 elr: 00000000e002bd28 lr : 00000000e002bce8 (reloc) elr: 00000000fff6fd28 lr : 00000000fff6fce8 x0 : 0000000000001041 x1 : 000000000000003e x2 : 00000000ffb0f8c8 x3 : 0000000000000001 x4 : 0000000000000080 x5 : 0000000000000000 x6 : 00000000fff718fc x7 : 000000000000000f x8 : 00000000ffb0f238 x9 : 0000000000000008 x10: 0000000000000000 x11: 0000000000000010 x12: 0000000000000006 x13: 000000000001869f x14: 00000000ffb0fcd0 x15: 0000000000000020 x16: 00000000fff71cc4 x17: 0000000000000000 x18: 00000000ffb13d90 x19: 00000000ffb14320 x20: 0000000000000000 x21: 00000000ffb14090 x22: 00000000ffb0f8c8 x23: 0000000000000001 x24: 00000000ffb14c10 x25: 0000000000000000 x26: 0000000000000000 x27: 0000000000000000 x28: 00000000ffb14c70 x29: 00000000ffb0f830
Code: 52800843 52800061 52800e00 97ffcf65 (b9400280) Resetting CPU ...
Signed-off-by: Masami Hiramatsu masami.hiramatsu@linaro.org
drivers/pci/pci-uclass.c | 3 +++ 1 file changed, 3 insertions(+)
Reviewed-by: Simon Glass sjg@chromium.org
I can't find this patch in my inbox, nor in the patchwork.
Hmm, it is strange... I set up gitconfig but it seems not working well. Let me send patch via my MUA next time.
Could you please resend?
OK, I attached it to this mail. (maybe it is safer in this case)
Thank you,
participants (3)
-
Bin Meng -
Masami Hiramatsu -
Simon Glass