On 21 January 2012 20:46, Xin Tong wrote:
> If the unlinking does not happen. ( i.e., interrupt checking in every
> TB is used instead of unlinking and force an exit). is the lock still
> needed ?
I don't think so.
-- PMM
If the unlinking does not happen. ( i.e., interrupt checking in every
TB is used instead of unlinking and force an exit). is the lock still
needed ?
Thanks
Xin
On Sat, Jan 21, 2012 at 2:55 PM, Peter Maydell wrote:
> On 21 January 2012 19:39, Xin Tong wrote:
>> There is a TB lock in the cpu-
On 21 January 2012 19:39, Xin Tong wrote:
> There is a TB lock in the cpu-exec.c. It is held before a vcpu is
> trying to find the next tb it is trying to execute. I am wondering
> what this lock is for. Currently, qemu is time multiplexing on a host
> cpu to emulate smp. No other vcpu can be runn
There is a TB lock in the cpu-exec.c. It is held before a vcpu is
trying to find the next tb it is trying to execute. I am wondering
what this lock is for. Currently, qemu is time multiplexing on a host
cpu to emulate smp. No other vcpu can be running while the tb
lookup/translation is performed.