external/srs
1
0
Fork 0
mirror of https://github.com/ossrs/srs.git synced 2025-03-09 15:49:59 +00:00
Code Issues Releases Wiki Activity

for bug #241, support mr(merged-read) config and reload. 2.0.52.

Browse source
This commit is contained in:
winlin 2014-12-04 18:21:04 +08:00
parent 57f844b636
commit 5589b13d2e
18 changed files with 293 additions and 48 deletions
Download patch file Download diff file Expand all files Collapse all files

2
trunk/src/core/srs_core.hpp
View file

@ -31,7 +31,7 @@ CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
// current release version
#define VERSION_MAJOR 2
#define VERSION_MINOR 0
#define VERSION_REVISION 51
#define VERSION_REVISION 52
// server info.
#define RTMP_SIG_SRS_KEY "SRS"
#define RTMP_SIG_SRS_ROLE "origin/edge server"

33
trunk/src/core/srs_core_performance.hpp
View file

@ -42,22 +42,23 @@ CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
* @example, for the default settings, this algorithm will use:
* that is, when got nread bytes smaller than 4KB, sleep(780ms).
*/
#if 1
// to enable merged read.
#define SRS_PERF_MERGED_READ
// the max sleep time in ms
#define SRS_MR_MAX_SLEEP_MS 780
// the max small bytes to group
#define SRS_MR_SMALL_BYTES 4096
// the underlayer api will set to SRS_MR_SOCKET_BUFFER bytes.
// 4KB=4096, 8KB=8192, 16KB=16384, 32KB=32768, 64KB=65536,
// 128KB=131072, 256KB=262144, 512KB=524288
// the buffer should set to SRS_MR_MAX_SLEEP_MS*kbps/8,
// for example, your system delivery stream in 1000kbps,
// sleep 800ms for small bytes, the buffer should set to:
// 800*1000/8=100000B(about 128KB).
#define SRS_MR_SOCKET_BUFFER 65536
#endif
/**
* https://github.com/winlinvip/simple-rtmp-server/issues/241#issuecomment-65554690
* The merged read algorithm is ok and can be simplified for:
* 1. Suppose the client network is ok. All algorithm go wrong when netowrk is not ok.
* 2. Suppose the client send each packet one by one. Although send some together, it's same.
* 3. SRS MR algorithm will read all data then sleep.
* So, the MR algorithm is:
* while true:
* read all data from socket.
* sleep a while
* For example, sleep 120ms. Then there is, and always 120ms data in buffer.
* That is, the latency is 120ms(the sleep time).
*/
// to enable merged read.
#undef SRS_PERF_MERGED_READ
// the max sleep time in ms
#define SRS_MR_MAX_SLEEP_MS 800
/**
* the send cache time in ms.