Re: [Xen-devel] [PATCH] x86: deal with firmware setting bogus TSC_ADJUST values

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 12.10.18 at 18:29, <wei.liu2@xxxxxxxxxx> wrote:
> On Mon, Oct 01, 2018 at 07:42:12AM -0600, Jan Beulich wrote:
>> The system Intel have handed me for AVX512 emulator work ("Gigabyte
>> Technology Co., Ltd. X299 AORUS Gaming 3 Pro/X299 AORUS Gaming 3
>> Pro-CF, BIOS F3 12/28/2017") would not come up under Xen - it hung in
>> the middle of Dom0 PCI initialization. As it turned out, Xen's time
>> management did not work because of the firmware setting (only) the boot
>> CPU's TSC_ADJUST MSR to a large negative value (on the order of -2^50).
>> 
>> Follow Linux (also shamelessly stealing their comments) in
> 
> Is there a specific commit or a range of commits in Linux that you can
> put here?

I don't think that would be worthwhile, as I only use the
underlying idea, not any of their code.

>> - clearing the register for the boot CPU (we don't have a need for
>>   exceptions here yet, as the only exception in Linux is a class of
>>   systems Xen doesn't work on anyway as far as I'm aware),
>> - forcing non-negative values uniformly,
>> - syncing the registers within sockets.
>> Linux caps at 0x7fffffff as well, but their comment saying "as those
>> wreckage the timer as well" does, to me at least, neither really explain
> 
> I tried to pin down what Linux does by searching the comment here but
> nothing showed up -- searching "wreckage" on Linux master only yielded
> three results, none of which matched the one you wrote here.

Hmm, indeed, that code went away in 4.13; I must have looked at
our 4.12 tree. I'll see if their removal / replacement gives me further
hints as to what else we may want to improve.

Jan



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel