diff options
| author | Bjorn Helgaas <bhelgaas@google.com> | 2015-09-22 02:26:45 +0300 |
|---|---|---|
| committer | Bjorn Helgaas <helgaas@kernel.org> | 2015-09-23 22:14:02 +0300 |
| commit | 0b26cd69480d4d99f39042c6621a05af7496b3da (patch) | |
| tree | 65dcf5c8702c803e5d527fb2dccdd0eac2b08e61 /Documentation/SAK.txt | |
| parent | 6ff33f3902c3b1c5d0db6b1e2c70b6d76fba357f (diff) | |
| download | linux-0b26cd69480d4d99f39042c6621a05af7496b3da.tar.xz | |
PCI: Clear IORESOURCE_UNSET when reverting to firmware-assigned address
If pci_assign_resource() fails to assign space for a BAR, we may restore
the BAR to whatever firmware left there at boot-time (this depends on
whether the arch implements pcibios_retrieve_fw_addr()). The messages we
print are not as useful as they could be:
pci 0000:00:01.0: BAR 15: assigned [mem 0xc0000000-0xc01fffff 64bit pref]
pci 0000:01:00.0: BAR 0: no space for [mem size 0x10000000 pref]
pci 0000:01:00.0: BAR 0: trying firmware assignment [mem size 0x10000000 pref]
pci 0000:01:00.0: BAR 0: [mem size 0x10000000 pref] conflicts with PCI Bus 0000:00 [mem 0xc0000000-0xffffffff window]
The last two lines should contain the actual BAR address, not the size.
Clear IORESOURCE_UNSET so we print the address. If requesting the
firmware-assigned resource fails, mark it IORESOURCE_UNSET again.
This is a cosmetic change to clarify the message: previously, if
pci_revert_fw_address() succeeded, pci_assign_resource() cleared
IORESOURCE_UNSET anyway, so this isn't really a functional change.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=85491#c50
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Diffstat (limited to 'Documentation/SAK.txt')
0 files changed, 0 insertions, 0 deletions