On 22.04.15 at 22:30,
wrote: Booting opensuse 13.2 to uname -rm 3.19.4-1.g74c332b-xen x86_64
I notice in the boot logs
dmesg | grep -i " pci " | egrep "fail|space" [ 0.338499] pci 0000:08:00.0: BAR 14: no space for [mem size 0x00400000] [ 0.338500] pci 0000:08:00.0: BAR 14: failed to assign [mem size 0x00400000] [ 0.338501] pci 0000:08:00.0: BAR 14: no space for [mem size 0x00400000] [ 0.338502] pci 0000:08:00.0: BAR 14: failed to assign [mem size 0x00400000] [ 0.338505] pci 0000:09:01.0: BAR 14: no space for [mem size 0x00400000] [ 0.338506] pci 0000:09:01.0: BAR 14: failed to assign [mem size 0x00400000] [ 0.338507] pci 0000:09:01.0: BAR 14: no space for [mem size 0x00400000] [ 0.338508] pci 0000:09:01.0: BAR 14: failed to assign [mem size 0x00400000]
Is this indeed only happening under Xen? If so, the difference to the native case would need to be investigated (i.e. open a bug, attaching the _full_ boot log of the kernel and hypervisor). If not, perhaps there's a BIOS setup option allowing more space for the needed MMIO ranges? Also without the full log it's hard to tell whether these are legitimate failures (taking into consideration the allocation strategy of the kernel), indications of bugs, or indications of that strategy needing improvement.
added 'pci=realloc'
cat /proc/cmdline root=... pci=realloc
recreated initrd & rebooted.
Unfortunately, it made it worse.
dmesg | grep -i " pci " | egrep "fail|space" [ 0.338475] pci 0000:08:00.0: BAR 14: no space for [mem size 0x00400000] [ 0.338476] pci 0000:08:00.0: BAR 14: failed to assign [mem size 0x00400000] [ 0.338477] pci 0000:09:01.0: BAR 14: no space for [mem size 0x00400000] [ 0.338478] pci 0000:09:01.0: BAR 14: failed to assign [mem size 0x00400000] [...] What are these errors, and what's a fix, if not the 'pci=realloc'?
That's a hint at there indeed not being enough MMIO space, or the kernel not (always) falling back to using 64-bit address ranges when short on 32-bit ones. But again - nothing specific can be said without seeing the full log. And also again it should be made clear whether this is a problem with _only_ the Xen kernel. Jan -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org