add hypervisor support section in README #709
Conversation
| @@ -41,6 +41,20 @@ documentation online at http://www.libvmi.com. | |||
| .. image:: https://scan.coverity.com/projects/14159/badge.svg | |||
| :target: https://scan.coverity.com/projects/libvmi-libvmi | |||
|
|
|||
| Hypervisor support | |||
| ------------------ | |||
There was a problem hiding this comment.
This is a good idea but we should probably have some more in-depth description of what the columns actually mean. Some of the events listed here are also architecture specific and only available on arm or x86.
There was a problem hiding this comment.
would be also pretty cool if you could add some details about the features that are supported by AMD processors
There was a problem hiding this comment.
Could you be more specific ?
There was a problem hiding this comment.
notes/events.txt says: "Currently only the Xen Hypervisor provides these features, and some of these are specifically only available on Intel CPU". Would be nice to know if this is still up to date or if some features still only work on Intel.
There was a problem hiding this comment.
I updated the table to reflect that CPUID and MSR were x86 specific events.
I never worked on AMD architecture, so I'm not sure about the descriptor, control register and debug register.
cc @bentau
There was a problem hiding this comment.
update on this ?
@bentau , @tklengyel
There was a problem hiding this comment.
i will try to contribute for AMD specific behavior soon
There was a problem hiding this comment.
Hi, @Wenzel @bentau I am currently working on AMD issues, and I see 4 problems
- single stepping not supported. AMD lacks the monitor-trap-flag, and xen doesnt implement any other way. i have a patch for xen, but currently I cannot test it.
- anything related to VMI_EVENT_MEMORY (e.g. syscall tracing example) results in a VMI_ERROR: xc_hvm_set_mem_access failed with code: -1. Despite my best efforts I cant seem to track down the underlying issue.
- register monitoring works, but not with MSR_ALL - here I get hangs on either one or even both VMs, the monitor and the one being monitored! Again i am not sure why.
- in the breakpoint example, re-injecting the bp doesnt work, but this is probably related to singlestepping not being available.
The rest actually seems to work fine, from working with the libvmi examples.
Wenzel
force-pushed
the
hypervisor_status
branch
from
b2a6483
to
313053b
Compare
tklengyel
force-pushed
the
master
branch
25 times, most recently
from
cd6a256
to
fb2e006
Compare
Add a section in
README.rstto quickly describe the hypervisors we are currently supporting and the features available.