On Fri, 2013-03-29 at 09:14 +0100, Paolo Bonzini wrote:
Il 29/03/2013 03:53, Nicholas A. Bellinger ha scritto:
On Thu, 2013-03-28 at 06:13 -0400, Paolo Bonzini wrote:
I think it's the right thing to do, but maybe not the right place to do this, need to reset after all IO is done, before ring memory is write protected.
Our emails are crossing each other unfortunately, but I want to reinforce this: ring memory is not write protected.
Understood. However, AFAICT the act of write protecting these ranges for ROM generates the offending callbacks to vhost_set_memory().
The part that I'm missing is if ring memory is not being write protected by make_bios_readonly_intel(), why are the vhost_set_memory() calls being invoked..?
Because mappings change for the region that contains the ring. vhost doesn't know yet that the changes do not affect ring memory, vhost_set_memory() is called exactly to ascertain that.
Hi Paolo & Co,
Here's a bit more information on what is going on with the same cpu_physical_memory_map() failure in vhost_verify_ring_mappings()..
So as before, at the point that seabios is marking memory as readonly for ROM in src/shadow.c:make_bios_readonly_intel() with the following call:
Calling pci_config_writeb(0x31): bdf: 0x0000 pam: 0x0000005b
the memory API update hook triggers back into vhost_region_del() code, and following occurs:
Entering vhost_region_del section: 0x7fd30a213b60 offset_within_region: 0xc0000 size: 2146697216 readonly: 0 vhost_region_del: is_rom: 0, rom_device: 0 vhost_region_del: readable: 1 vhost_region_del: ram_addr 0x0, addr: 0x0 size: 2147483648 vhost_region_del: name: pc.ram Entering vhost_set_memory, section: 0x7fd30a213b60 add: 0, dev->started: 1 Entering verify_ring_mappings: start_addr 0x00000000000c0000 size: 2146697216 verify_ring_mappings: ring_phys 0x0 ring_size: 0 verify_ring_mappings: ring_phys 0x0 ring_size: 0 verify_ring_mappings: ring_phys 0xed000 ring_size: 5124 verify_ring_mappings: calling cpu_physical_memory_map ring_phys: 0xed000 l: 5124 address_space_map: addr: 0xed000, plen: 5124 address_space_map: l: 4096, len: 5124 phys_page_find got PHYS_MAP_NODE_NIL >>>>>>>>>>>>>>>>>>>>>>.. address_space_map: section: 0x7fd30fabaed0 memory_region_is_ram: 0 readonly: 0 address_space_map: section: 0x7fd30fabaed0 offset_within_region: 0x0 section size: 18446744073709551615 Unable to map ring buffer for ring 2, l: 4096
So the interesting part is that phys_page_find() is not able to locate the corresponding page for vq->ring_phys: 0xed000 from the vhost_region_del() callback with section->offset_within_region: 0xc0000..
Is there any case where this would not be considered a bug..?
register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 Entering vhost_region_add section: 0x7fd30a213aa0 offset_within_region: 0xc0000 size: 32768 readonly: 1 vhost_region_add: is_rom: 0, rom_device: 0 vhost_region_add: readable: 1 vhost_region_add: ram_addr 0x0000000000000000, addr: 0x 0 size: 2147483648 vhost_region_add: name: pc.ram Entering vhost_set_memory, section: 0x7fd30a213aa0 add: 1, dev->started: 1 Entering verify_ring_mappings: start_addr 0x00000000000c0000 size: 32768 verify_ring_mappings: ring_phys 0x0 ring_size: 0 verify_ring_mappings: ring_phys 0x0 ring_size: 0 verify_ring_mappings: ring_phys 0xed000 ring_size: 5124 verify_ring_mappings: Got !ranges_overlap, skipping register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 Entering vhost_region_add section: 0x7fd30a213aa0 offset_within_region: 0xc8000 size: 2146664448 readonly: 0 vhost_region_add: is_rom: 0, rom_device: 0 vhost_region_add: readable: 1 vhost_region_add: ram_addr 0x0000000000000000, addr: 0x 0 size: 2147483648 vhost_region_add: name: pc.ram Entering vhost_set_memory, section: 0x7fd30a213aa0 add: 1, dev->started: 1 Entering verify_ring_mappings: start_addr 0x00000000000c8000 size: 2146664448 verify_ring_mappings: ring_phys 0x0 ring_size: 0 verify_ring_mappings: ring_phys 0x0 ring_size: 0 verify_ring_mappings: ring_phys 0xed000 ring_size: 5124 verify_ring_mappings: calling cpu_physical_memory_map ring_phys: 0xed000 l: 5124 address_space_map: addr: 0xed000, plen: 5124 address_space_map: l: 4096, len: 5124 address_space_map: section: 0x7fd30fabb020 memory_region_is_ram: 1 readonly: 0 address_space_map: section: 0x7fd30fabb020 offset_within_region: 0xc8000 section size: 2146664448 address_space_map: l: 4096, len: 1028 address_space_map: section: 0x7fd30fabb020 memory_region_is_ram: 1 readonly: 0 address_space_map: section: 0x7fd30fabb020 offset_within_region: 0xc8000 section size: 2146664448 address_space_map: Calling qemu_ram_ptr_length: raddr: 0x ed000 rlen: 5124 address_space_map: After qemu_ram_ptr_length: raddr: 0x ed000 rlen: 5124
So here the vhost_region_add() callback for section->offset_within_region: 0xc8000 for vq->ring_phys: 0xed000 is able to locate *section via phys_page_find() within address_space_map(), and cpu_physical_memory_map() completes as expected..
register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 register_multipage : d: 0x7fd30f7d0ed0 section: 0x7fd30a2139b0 phys_page_find got PHYS_MAP_NODE_NIL >>>>>>>>>>>>>>>>>>>>>>..
So while plodding my way through the memory API, the thing that would be useful to know is if the offending *section that is missing for the first phys_page_find() call is getting removed before the callback makes it's way into vhost_verify_ring_mappings() code, or that some other bug is occuring..?
Any idea on how this could be verified..?
Thanks,
--nab