Re: [Qemu-devel] [RFC/PATCH] Fix guest OS panic when 64bit BAR is present

[Michael S. Tsirkin]

On Wed, Jan 25, 2012 at 06:46:03PM +1300, Alexey Korolev wrote:
> Hi, 
> In this post
> http://lists.gnu.org/archive/html/qemu-devel/2011-12/msg03171.html I've
> mentioned about the issues when 64Bit PCI BAR is present and 32bit
> address range is selected for it.
> The issue affects all recent qemu releases and all
> old and recent guest Linux kernel versions.
> 
> We've done some investigations. Let me explain what happens.
> Assume we have 64bit BAR with size 32MB mapped at [0xF0000000 -
> 0xF2000000]
> 
> When Linux guest starts it does PCI bus enumeration.
> The OS enumerates 64BIT bars using the following procedure.
> 1. Write all FF's to lower half of 64bit BAR
> 2. Write address back to lower half of 64bit BAR
> 3. Write all FF's to higher half of 64bit BAR
> 4. Write address back to higher half of 64bit BAR
> 
> Linux code is here: 
> http://lxr.linux.no/#linux+v3.2.1/drivers/pci/probe.c#L149
> 
> What does it mean for qemu?
> 
> At step 1. qemu pci_default_write_config() recevies all FFs for lower
> part of the 64bit BAR. Then it applies the mask and converts the value
> to "All FF's - size + 1" (FE000000 if size is 32MB).
> Then pci_bar_address() checks if BAR address is valid. Since it is a
> 64bit bar it reads 0x00000000FE000000 - this address is valid. So qemu
> updates topology and sends request to update mappings in KVM with new
> range for the 64bit BAR FE000000 - 0xFFFFFFFF. This usually means kernel
> panic on boot, if there is another mapping in the FE000000 - 0xFFFFFFFF
> range, which is quite common.
> 
> 
> The following patch fixes the issue. It affects 64bit PCI BAR's only. 
> The idea of the patch is: we introduce the states for low and high BARs
> whose can have 3 possible values: BAR_VALID, PCIBAR64_PARTIAL_SIZE_QUERY
> - someone has requested size of one half of the 64bit PCI BAR,
> PCIBAR64_PARTIAL_ADDR_PROGRAM - someone has sent a request to update the
> address of one half of the 64bit PCI BAR. The state becomes BAR_VALID
> when both halfs are in the same state. We ignore BAR value until both
> states become BAR_VALID
> 
> Note: Please use the latest Seabios version (commit
> 139d5ac037de828f89c36e39c6dd15610650cede and later), as older versions
> didn't initialize high part of 64bit BAR. 
> 
> The patch is tested on Linux 2.6.18 - 3.1.0 and Windows 2008 Server
> 
> Signed-off-by: Alexey Korolev <address@hidden>

Interesting. However, looking at guest code,
I note that memory and io are disabled
during BAR sizing unless mmio always on is set.
pci_bar_address should return PCI_BAR_UNMAPPED
in this case, and we should never map this BAR
until it's enabled. What's going on?