linux-hyperv.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v2 0/3] hyperv: Fixes for get_vtl(void)
@ 2024-12-26 20:30 Roman Kisel
  2024-12-26 20:30 ` [PATCH v2 1/3] hyperv: Define struct hv_output_get_vp_registers Roman Kisel
                   ` (2 more replies)
  0 siblings, 3 replies; 4+ messages in thread
From: Roman Kisel @ 2024-12-26 20:30 UTC (permalink / raw)
  To: hpa, kys, bp, dave.hansen, decui, eahariha, haiyangz, mingo,
	mhklinux, nunodasneves, tglx, tiala, wei.liu, linux-hyperv,
	linux-kernel, x86
  Cc: apais, benhill, ssengar, sunilmut, vdso

The get_vtl(void) function

* has got one bug when the code started using a wrong pointer type after
  refactoring, and also
* it doesn't adhere to the requirements of the Hypervisor Top-Level Funactional
  Specification[1, 2] as the code overlaps the input and output areas for a hypercall.

The first issue leads to a wrong 100% reproducible computation due to reading
a byte worth of data at a wrong offset. That in turn leads to using a nonsensical
value ("fortunately", could catch it easily!) for the current VTL when initiating
VMBus communications. As a repercussion from that, the system wouldn't boot. The
fix is straightforward: use the correct pointer type.

The second issue doesn't seem to lead to any reproducible breakage just yet. It is
fixed with using the output hypercall pages allocated per-CPU, and that isn't the
only or the most obvious choice so let me elaborate why that fix appears to be the
best one in my opinion out of the options I could conceive of.

The approach chosen for fixing the second issue makes two things shine through:

* `get_vtl(void)` is just getting a vCPU register, no special treatment needs
  to be involved,
* VTLs and dom0 can and should share code as both exist to provide services to
  a guest(s), be that from within the partition or from outside of it.

The projected benefits include replacing the function in question with a future
`hv_get_vp_registers` one shared between dom0 and VTLs to allow for a better test
coverage.

I have validated the fixes by booting the fixed kernel in VTL2 up using OpenVMM and
OpenHCL[3, 4].

[1] https://learn.microsoft.com/en-us/virtualization/hyper-v-on-windows/tlfs/hypercall-interface
[2] https://github.com/MicrosoftDocs/Virtualization-Documentation/tree/main/tlfs
[3] https://openvmm.dev/guide/user_guide/openhcl.html
[4] https://github.com/microsoft/OpenVMM

[v2]
  - Used the suggestions to define an additional structure to improve code readability,
  - Split out the patch with that definition.

[v1]: https://lore.kernel.org/lkml/20241218205421.319969-1-romank@linux.microsoft.com/

Roman Kisel (3):
  hyperv: Define struct hv_output_get_vp_registers
  hyperv: Fix pointer type for the output of the hypercall in
    get_vtl(void)
  hyperv: Do not overlap the input and output hypercall areas in
    get_vtl(void)

 arch/x86/hyperv/hv_init.c   |  6 ++--
 drivers/hv/hv_common.c      |  6 ++--
 include/hyperv/hvgdk_mini.h | 58 +++++++++++++++++++++++++++++++++++--
 3 files changed, 62 insertions(+), 8 deletions(-)


base-commit: 4d4ace979a3066e5c940331571e6c1c3f280d1d3
-- 
2.34.1


^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2024-12-26 20:30 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-12-26 20:30 [PATCH v2 0/3] hyperv: Fixes for get_vtl(void) Roman Kisel
2024-12-26 20:30 ` [PATCH v2 1/3] hyperv: Define struct hv_output_get_vp_registers Roman Kisel
2024-12-26 20:30 ` [PATCH v2 2/3] hyperv: Fix pointer type for the output of the hypercall in get_vtl(void) Roman Kisel
2024-12-26 20:30 ` [PATCH v2 3/3] hyperv: Do not overlap the input and output hypercall areas " Roman Kisel

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).