On 04/13/2010 03:41 PM, H. Peter Anvin wrote: > On 04/13/2010 03:29 PM, Yinghai wrote: >>> >>> We have talked about a need to resolve this before. >> >> current code for mmio that is just below 4g, if some PCI BAR use that range, and those range is falling into E820_RESERVED, >> >> those range still can be claimed, but driver can not use pci_request_region() later. >> >> So We still >> 1. rely that BIOS does not reserve the [0xa0000, 0xe0000) >> 2. kernel only reserve the range when we make sure these is legacy device on that range. >> > > This really isn't sufficient. There are systems in the field which > marks a memory range reserved in E820 because it a device pointed there, > and it doesn't want that device moved because it is used by an SMM handler. > > This was reported quite a while ago (like two years.) I can dig up the > thread if it matters. Are you sure? what is BAR range? greater than 1M ? e820_reserve_resources() will make that range to be reserved and BUSY in resource tree. and if driver for that device want to call pci_request_region, it will get failure... YH -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
