On Tue, Jun 23, 2015 at 12:49 PM, Fabien Chouteau <chout...@adacore.com>
wrote:
>
> Maybe we can, but the "while" will only loop when threads are running on
> different CPU, so the other thread is already running and calling sleep
> will not help I think.
>
What you think about this:
diff --git a/cpus.c b/cpus.c
index 83d5eb5..2e71221 100644
--- a/cpus.c
+++ b/cpus.c
@@ -1097,8 +1097,10 @@ static void qemu_cpu_kick_thread(CPUState *cpu)
* suspended until we can get the context.
*/
tcgContext.ContextFlags = CONTEXT_CONTROL;
- while (GetThreadContext(cpu->hThread, &tcgContext) == 0) {
- continue;
+ if (GetThreadContext(cpu->hThread, &tcgContext) == 0) {
+ fprintf(stderr, "qemu:%s: GetLastError:%lu\n", __func__,
+ GetLastError());
+ exit(1);
}
cpu_signal(0);
Without THREAD_GET_CONTEXT flag this code instantly fails with error
message.
So if GetThreadContext will fail sometime in future we will get useful
stderr.txt.
Also GetThreadContext is synchronized function(at least this is wrotten by
authoritative MS man - see the link from my previous post). No need to ping
GetThreadContext until it works.
> Yan, if you didn't already, I recommend you extensively test this
> modification. By extensively, I mean running QEMU several thousands of
> time on an SMP host (with many CPUs like 8 or 16 if possible).
>
Temporarily now I have only my home 4-cores CPU.
But this mistake I found in March. Fix it and run Qemu many times. Mostly
with OVMF BIOS.
