Actions
Bug #14195
openpci_cap_locate and pci_xcap_locate need additional bounds
Start date:
Due date:
% Done:
80%
Estimated time:
Difficulty:
Medium
Tags:
Gerrit CR:
Description
When walking PCI capabilities, it is possible for something to screw up in a way that you start reading garbage. When this happens, your only hope is that you find a register that actually looks like a proper capability that ends the list. However, this might not happen and an example of that is present in #14193. In pcieadm
and others, the amount of iterations if often bounded because there are only so many caps that can fit in either basic or extended configuration space based on the size of the capability.
Related issues
Updated by Robert Mustacchi 7 months ago
- Related to Bug #14192: kstat_delete(9F) should decide how it feels about NULL added
Updated by Robert Mustacchi 7 months ago
- Related to deleted (Bug #14192: kstat_delete(9F) should decide how it feels about NULL)
Updated by Robert Mustacchi 7 months ago
- Related to Bug #14193: PCI config space I/O port mechanisms need to check offsets added
Actions