Vincent,
In my experience, most performance issues have been in my driver or hardware platform, not in lwIP. I don’t use a real OS so I can’t speak to its added overhead. I do note you changed PBUF_POOL_BUFSIZE. I believe if smaller than the default you get chained pbufs which is a little less efficient than one packet per pbuf. Bill From: [email protected] [mailto:[email protected]] On Behalf Of vincent cui Sent: Wednesday, May 16, 2012 1:49 AM To: Mailing list for lwIP users Subject: Re: [lwip-users] lwip whole performance down Hi : I enable LWIP_STAT to get LWIP status when performance goes down. but it seems no error detected … Anyone help to have a look ? ETHARP xmit: 2 recv: 761 fw: 0 drop: 0 chkerr: 0 lenerr: 0 memerr: 0 rterr: 0 proterr: 0 opterr: 0 err: 0 cachehit: 1750 IP xmit: 1753 recv: 1971 fw: 0 drop: 0 chkerr: 0 lenerr: 0 memerr: 0 rterr: 0 proterr: 0 opterr: 0 err: 0 cachehit: 0 TCP xmit: 1 recv: 1757 fw: 0 drop: 0 chkerr: 0 lenerr: 0 memerr: 0 rterr: 0 proterr: 0 opterr: 0 err: 0 cachehit: 0 MEM HEAP avail: 12288 used: 1676 max: 1760 err: 0 MEM RAW_PCB avail: 4 used: 0 max: 0 err: 0 MEM UDP_PCB avail: 6 used: 1 max: 2 err: 0 MEM TCP_PCB avail: 16 used: 1 max: 2 err: 0 MEM TCP_PCB_LISTEN avail: 6 used: 4 max: 4 err: 0 MEM TCP_SEG avail: 32 used: 0 max: 1 err: 0 MEM NETBUF avail: 2 used: 0 max: 0 err: 0 MEM NETCONN avail: 8 used: 2 max: 2 err: 0 MEM TCPIP_MSG_API avail: 8 used: 0 max: 0 err: 0 MEM TCPIP_MSG_INPKT avail: 8 used: 1 max: 1 err: 0 MEM SYS_TIMEOUT avail: 10 used: 6 max: 6 err: 0 MEM SNMP_ROOTNODE avail: 30 used: 0 max: 0 err: 0 MEM SNMP_NODE avail: 50 used: 0 max: 0 err: 0 MEM SNMP_VARBIND avail: 2 used: 0 max: 0 err: 0 MEM SNMP_VALUE avail: 3 used: 0 max: 0 err: 0 MEM PBUF_REF/ROM avail: 16 used: 0 max: 0 err: 0 MEM PBUF_POOL avail: 10 used: 2 max: 2 err: 0 From: [email protected] [mailto:[email protected]] On Behalf Of vincent cui Sent: 2012年5月15日 21:31 To: Mailing list for lwIP users Subject: [lwip-users] lwip whole performance down Hi: I am working on LWIP1.4.0 and FreeRTOSv7.1.1 to setup web server with SOCKET API..and found that sending speed becomes lower. In order to know the actual sending speed, I write a simple application to test it. And found that the sending speed is up to 8Mb/s. it is high enough for me. I notice that the application only setup connection one times. In web server, it closed connection after sending all data out. I do n’t know if it may cause system Performance down. I extend the sending buffer as I can. They are following: #define TCP_MSS (1500 - 40) /* TCP_MSS = (Ethernet MTU - IP header size - TCP header size) */ /* TCP sender buffer space (bytes). */ #define TCP_SND_BUF (8*TCP_MSS) /* TCP_SND_QUEUELEN: TCP sender buffer space (pbufs). This must be at least as much as (2 * TCP_SND_BUF/TCP_MSS) for things to work. */ #define TCP_SND_QUEUELEN (4* TCP_SND_BUF/TCP_MSS) /* TCP receive window. */ #define TCP_WND (4*TCP_MSS) /* ---------- Pbuf options ---------- */ /* PBUF_POOL_SIZE: the number of buffers in the pbuf pool. */ #define PBUF_POOL_SIZE 10 /* PBUF_POOL_BUFSIZE: the size of each pbuf in the pbuf pool. */ #define PBUF_POOL_BUFSIZE 1024 /* MEMP_NUM_TCP_SEG: the number of simultaneously queued TCP segments. */ #define MEMP_NUM_TCP_SEG 64 /* MEMP_NUM_TCP_PCB: the number of simulatenously active TCP connections. */ #define MEMP_NUM_TCP_PCB 32 /* MEM_SIZE: the size of the heap memory. If the application will send a lot of data that needs to be copied, this should be set high. */ #define MEM_SIZE (12*1024) /* MEMP_NUM_PBUF: the number of memp struct pbufs. If the application sends a lot of data out of ROM (or other static memory), this should be set high. */ #define MEMP_NUM_PBUF 64 Those value is enough big for me….but why do I get better performance ? vincent
_______________________________________________ lwip-users mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/lwip-users
