Regression caused by "PCI: Shrink decoding-disabled window while sizing BARs"

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hello,

Since next-20141117, PCI support is broken on ARM Marvell EBU
platforms. next-20141114 was working fine. The problem we see is:

pci 0000:03:00.0: reg 0x10: can't handle BAR larger than 4GB (size 0xfffffffffff00000)

Therefore, the Marvell-specific MBus window is not setup, and the first
access to the PCI device registers causes a kernel panic.

You can compare a successful boot:

  http://storage.armcloud.us/kernel-ci/next/next-20141114/arm-mvebu_v7_defconfig/boot-armada-370-mirabox.html

With a failed boot:

  http://storage.armcloud.us/kernel-ci/next/next-20141117/arm-mvebu_v7_defconfig/boot-armada-370-mirabox.html

It turns out that if I take next-20141117 and revert the two following
commits:

  7ea945f0bb49423451dbf222f62c6fabc857aaac PCI: Shrink decoding-disabled window while sizing BARs
  3a02517d5e2aef7cd4524468525a6292c81cfad1 PCI: Support 64-bit bridge windows if we have 64-bit dma_addr_t

Then the PCI support works again.

Note that the Marvell pci-mvebu driver emulates its own PCI bridge, in
order to support the dynamic creation of MBus windows (this is a
Marvell specific concept). Therefore, there is a possibility that the
issue might be that our emulation is not 100% correct, which I'll be
checking today.

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [DMA Engine]     [Linux Coverity]     [Linux USB]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Greybus]

  Powered by Linux