Kyösti Mälkki has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/31685 )
Change subject: device/pci_ops: Reuse romstage PCI config for ramstage ......................................................................
Patch Set 5:
(1 comment)
Patch Set 5: Code-Review+1
(1 comment)
I'm not very fond of the FIXME ;) is there a solution in the queue, yet? If not, I fear callers could get used to the fact that the "ops" check for NULL while we should actually encourage them to explicitly check before calling.
For statically declared "ops" functions the call always takes form dev->ops->func(dev), unless you get innovative you will never pass NULL there.
https://review.coreboot.org/#/c/31685/5/src/include/device/pci_type.h File src/include/device/pci_type.h:
https://review.coreboot.org/#/c/31685/5/src/include/device/pci_type.h@32 PS5, Line 32: * the check themselves.
Maybe the following strategy would work: for pcidev_behind() and alike […]
Yep, I just did not get that done yet. Going thru assert() retains behaviour to parent commit and is not a regression per-se.
Also, pcidev_die() could dump return address from stack, to avoid added weight of __func__ strings.