Am 09.04.2013 16:10, schrieb Peter Meerwald:
> From: Peter Meerwald <p.meerwald at bct-electronic.com>
>
> pa_write() knows two types of operation:
> calling send() and calling write()
>
> there is a flag (a pointer to an int) passed to pa_write()
> which can remember which write type was successful
>
> if the pointer is NULL or the int is 0, send() is tried first,
> with a fallback to write() if send() resulted in ENOTSOCK
>
> pa_fdsem_post() calls pa_write() with a NULL pointer;
> unfortunately (at least with HAVE_SYS_EVENTFD_H #define'd) send()
> always fails here and write() is called -- causing an extra syscall
> quite frequently
>
> strace:
> send(17, "\1\0\0\0\0\0\0\0", 8, MSG_NOSIGNAL) = -1 ENOTSOCK (Socket operation on non-socket)
> write(17, "\1\0\0\0\0\0\0\0", 8) = 8
>
> the patch adds a write_type field to pa_fdsem to the successful
> pa_write() type can be remembered and unnecessary send() calls are
> avoided
>
> Signed-off-by: Peter Meerwald <p.meerwald at bct-electronic.com>
> ---
> src/pulsecore/fdsem.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/src/pulsecore/fdsem.c b/src/pulsecore/fdsem.c
> index 14fcbd6..1d3d44a 100644
> --- a/src/pulsecore/fdsem.c
> +++ b/src/pulsecore/fdsem.c
> @@ -52,7 +52,7 @@ struct pa_fdsem {
> #ifdef HAVE_SYS_EVENTFD_H
> int efd;
> #endif
> -
> + int write_type;
> pa_fdsem_data *data;
> };
>
> @@ -196,7 +196,7 @@ void pa_fdsem_post(pa_fdsem *f) {
> if (f->efd >= 0) {
> uint64_t u = 1;
>
> - if ((r = pa_write(f->efd, &u, sizeof(u), NULL)) != sizeof(u)) {
> + if ((r = pa_write(f->efd, &u, sizeof(u), &f->write_type)) != sizeof(u)) {
> if (r >= 0 || errno != EINTR) {
> pa_log_error("Invalid write to eventfd: %s", r < 0 ? pa_cstrerror(errno) : "EOF");
> pa_assert_not_reached();
> @@ -207,7 +207,7 @@ void pa_fdsem_post(pa_fdsem *f) {
> } else
> #endif
>
> - if ((r = pa_write(f->fds[1], &x, 1, NULL)) != 1) {
> + if ((r = pa_write(f->fds[1], &x, 1, &f->write_type)) != 1) {
> if (r >= 0 || errno != EINTR) {
> pa_log_error("Invalid write to pipe: %s", r < 0 ? pa_cstrerror(errno) : "EOF");
> pa_assert_not_reached();
The type should be initialized to 0. This doesn't happen in the
functions that allocate pa_fdsem objects.
Best regards.
