Am 18.01.2014 um 00:55 hat Max Reitz geschrieben:
> On 17.01.2014 15:15, Kevin Wolf wrote:
> >There is no easy way to check that a request correctly waits for a
> >different request. With a sleep command we can at least approximate it.
> >
> >Signed-off-by: Kevin Wolf <kw...@redhat.com>
> >---
> > qemu-io-cmds.c | 42 ++++++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 42 insertions(+)
> >
> >diff --git a/qemu-io-cmds.c b/qemu-io-cmds.c
> >index 85e4982..978a3a0 100644
> >--- a/qemu-io-cmds.c
> >+++ b/qemu-io-cmds.c
> >@@ -12,6 +12,7 @@
> > #include "block/block_int.h"
> > #include "block/qapi.h"
> > #include "qemu/main-loop.h"
> >+#include "qemu/timer.h"
> > #define CMD_NOFILE_OK 0x01
> >@@ -2038,6 +2039,46 @@ static const cmdinfo_t abort_cmd = {
> > .oneline = "simulate a program crash using abort(3)",
> > };
> >+static void sleep_cb(void *opaque)
> >+{
> >+ bool *expired = opaque;
> >+ *expired = true;
> >+}
> >+
> >+static int sleep_f(BlockDriverState *bs, int argc, char **argv)
> >+{
> >+ char *endptr;
> >+ long ms;
> >+ struct QEMUTimer *timer;
> >+ bool expired = false;
> >+
> >+ ms = strtol(argv[1], &endptr, 0);
> >+ if (ms < 0 || *endptr != '\0') {
> >+ printf("%s is not a valid number\n", argv[1]);
> >+ return 0;
> >+ }
> >+
> >+ timer = timer_new_ns(QEMU_CLOCK_HOST, sleep_cb, &expired);
> >+ timer_mod(timer, qemu_clock_get_ns(QEMU_CLOCK_HOST) + SCALE_MS * ms);
> >+
> >+ while (!expired) {
>
> I don't know whether compilers don't optimize accesses to variables
> whose addresses have been given to other functions, but shouldn't
> expired be marked volatile just to be sure?
That would be a pretty broken compiler, and we're already using the same
pattern all over qemu without volatile, so no. (And if there was a
problem, I guess we'd rather use barriers than volatile variables.)
Kevin
