Commit Graph

1007 Commits

Author SHA1 Message Date
chenwei 65074cf4f4 Nginx : The I/O functions with glue code may be called before ngx_max_sockets is initialized, fix this error. 2017-12-14 15:07:30 +08:00
chenwei 2b31b4e34c Nginx : The I/O functions with glue code may be called before ngx_max_sockets is initialized, fix this error. 2017-12-14 15:07:30 +08:00
chenwei fb8ab126f5 Nginx : Follow nginx programming convention - indent. 2017-12-14 14:55:36 +08:00
chenwei df18ef7de7 Nginx : Follow nginx programming convention - indent. 2017-12-14 14:55:36 +08:00
logwang fa74a859d0 Add tool: arp.
The arp utility displays and modifies the Internet-to-Ethernet address translation tables used by the address resolution protocol.
2017-12-13 17:05:48 +08:00
logwang df9cc249f2 Add tool: arp.
The arp utility displays and modifies the Internet-to-Ethernet address translation tables used by the address resolution protocol.
2017-12-13 17:05:48 +08:00
chenwei 03e8baeb66 rename "timers" and "events_posted" 2017-12-09 15:24:58 +08:00
chenwei acc0462acd rename "timers" and "events_posted" 2017-12-09 15:24:58 +08:00
chenwei a7886ec8af rename ngx_event_actions_dy to ngx_ff_host_event_actions 2017-12-09 15:02:13 +08:00
chenwei 590bcd338b rename ngx_event_actions_dy to ngx_ff_host_event_actions 2017-12-09 15:02:13 +08:00
chenwei 6b647d46a9 rename module module ff_channel to ff_host_event 2017-12-09 14:55:48 +08:00
chenwei 142257e3b6 rename module module ff_channel to ff_host_event 2017-12-09 14:55:48 +08:00
chenwei d7bd9e987b Using a more friendly name : belong_to_aeds to belong_to_host 2017-12-09 14:18:11 +08:00
chenwei 5c797d9e1a Using a more friendly name : belong_to_aeds to belong_to_host 2017-12-09 14:18:11 +08:00
chenwei 4b4f027dad Using a more friendly name
1.  remove the assert(ff_fdisused(sockfd))
2017-12-09 14:07:08 +08:00
chenwei 447a82c866 Using a more friendly name
1.  remove the assert(ff_fdisused(sockfd))
2017-12-09 14:07:08 +08:00
chenwei 70bb2888cb Nginx: support kernel network stack, so we can do what fstack can't do,
e.g. unix socket, ipc (with APP on kernel network stack), packet from kernel network stack.
1. Add a new directive kernel_network_stack :
    Syntax: 	kernel_network_stack on | off;
    Default: 	kernel_network_stack off;
   Context: 	http, server
  This directive is available only when NGX_HAVE_FF_STACK is defined.
  Determines whether server should run on kernel network stack or fstack.
2. Use a simpler and  more effective solution to discriminate fstack fd(file descriptor, only socket for now) from kernel fd.
2017-12-08 18:32:08 +08:00
chenwei 3ce9eefdd7 Nginx: support kernel network stack, so we can do what fstack can't do,
e.g. unix socket, ipc (with APP on kernel network stack), packet from kernel network stack.
1. Add a new directive kernel_network_stack :
    Syntax: 	kernel_network_stack on | off;
    Default: 	kernel_network_stack off;
   Context: 	http, server
  This directive is available only when NGX_HAVE_FF_STACK is defined.
  Determines whether server should run on kernel network stack or fstack.
2. Use a simpler and  more effective solution to discriminate fstack fd(file descriptor, only socket for now) from kernel fd.
2017-12-08 18:32:08 +08:00
chenwei 7138141b32 Merge pull request #1 from F-Stack/master
update to data
2017-12-07 21:08:20 +08:00
chenwei 1b895bd9c2
Merge pull request #1 from F-Stack/master
update to data
2017-12-07 21:08:20 +08:00
logwang b226a4cf7e ff_epoll: rewrite `ff_epoll_ctl`. 2017-12-07 12:24:49 +08:00
logwang 4ca4a48737 ff_epoll: rewrite `ff_epoll_ctl`. 2017-12-07 12:24:49 +08:00
logwang 9d39026bdb ff_epoll: support edge-triggered mode.
Convert epoll EPOLLET to kqueue EV_CLEAR.
2017-12-05 17:51:02 +08:00
logwang 794317ab37 ff_epoll: support edge-triggered mode.
Convert epoll EPOLLET to kqueue EV_CLEAR.
2017-12-05 17:51:02 +08:00
logwang 2aa28acdb3 Fix #114: An out of bounds of memory in netinet/libalias/alias_sctp.c.
Run with valgrind, and found this:
==2228== Invalid write of size 8
==2228==    at 0x4E05DA: AliasSctpInit (alias_sctp.c:641)
==2228==    by 0x4DE565: LibAliasInit (alias_db.c:2503)
==2228==    by 0x4E9B3B: nat44_config (ip_fw_nat.c:505)
==2228==    by 0x4E9E91: nat44_cfg (ip_fw_nat.c:599)
==2228==    by 0x4F1719: ipfw_ctl3 (ip_fw_sockopt.c:3666)
==2228==    by 0x4B9954: rip_ctloutput (raw_ip.c:659)
==2228==    by 0x447E11: sosetopt (uipc_socket.c:2505)
==2228==    by 0x44BF4D: kern_setsockopt (uipc_syscalls.c:1407)
==2228==    by 0x409F08: ff_setsockopt (ff_syscall_wrapper.c:412)
==2228==    by 0x5277AA: handle_ipfw_msg (ff_dpdk_if.c:1146)
==2228==    by 0x52788C: handle_msg (ff_dpdk_if.c:1196)
==2228==    by 0x5289B8: process_msg_ring (ff_dpdk_if.c:1213)
==2228==  Address 0x60779b0 is 4,800 bytes inside a block of size 4,802
alloc'd
==2228==    at 0x4C2ABBD: malloc (vg_replace_malloc.c:296)
==2228==    by 0x509F15: ff_malloc (ff_host_interface.c:89)
==2228==    by 0x4053BE: malloc (ff_glue.c:1021)
==2228==    by 0x4E054E: AliasSctpInit (alias_sctp.c:632)
==2228==    by 0x4DE565: LibAliasInit (alias_db.c:2503)
==2228==    by 0x4E9B3B: nat44_config (ip_fw_nat.c:505)
==2228==    by 0x4E9E91: nat44_cfg (ip_fw_nat.c:599)
==2228==    by 0x4F1719: ipfw_ctl3 (ip_fw_sockopt.c:3666)
==2228==    by 0x4B9954: rip_ctloutput (raw_ip.c:659)
==2228==    by 0x447E11: sosetopt (uipc_socket.c:2505)
==2228==    by 0x44BF4D: kern_setsockopt (uipc_syscalls.c:1407)
==2228==    by 0x409F08: ff_setsockopt (ff_syscall_wrapper.c:412)
==2228==

The error line is:
`la->sctpNatTimer.TimerQ = sn_calloc(SN_TIMER_QUEUE_SIZE, sizeof(struct
sctpTimerQ));`

Since SN_TIMER_QUEUE_SIZE is defined as SN_MAX_TIMER+2, and sn_calloc is
defined as sn_malloc(x * n) if _SYS_MALLOC_H_ is defined, the size of
calloced memory will be wrong, because the macro will be expanded to
sizeof(struct sctpTimerQ)*SN_MAX_TIMER+2.

And the memory will be out of bounds here.
```
/* Initialise circular timer Q*/
for (i = 0; i < SN_TIMER_QUEUE_SIZE; i++)
    LIST_INIT(&la->sctpNatTimer.TimerQ[i]);
```
2017-12-05 15:32:10 +08:00
logwang 7f92df961d Fix #114: An out of bounds of memory in netinet/libalias/alias_sctp.c.
Run with valgrind, and found this:
==2228== Invalid write of size 8
==2228==    at 0x4E05DA: AliasSctpInit (alias_sctp.c:641)
==2228==    by 0x4DE565: LibAliasInit (alias_db.c:2503)
==2228==    by 0x4E9B3B: nat44_config (ip_fw_nat.c:505)
==2228==    by 0x4E9E91: nat44_cfg (ip_fw_nat.c:599)
==2228==    by 0x4F1719: ipfw_ctl3 (ip_fw_sockopt.c:3666)
==2228==    by 0x4B9954: rip_ctloutput (raw_ip.c:659)
==2228==    by 0x447E11: sosetopt (uipc_socket.c:2505)
==2228==    by 0x44BF4D: kern_setsockopt (uipc_syscalls.c:1407)
==2228==    by 0x409F08: ff_setsockopt (ff_syscall_wrapper.c:412)
==2228==    by 0x5277AA: handle_ipfw_msg (ff_dpdk_if.c:1146)
==2228==    by 0x52788C: handle_msg (ff_dpdk_if.c:1196)
==2228==    by 0x5289B8: process_msg_ring (ff_dpdk_if.c:1213)
==2228==  Address 0x60779b0 is 4,800 bytes inside a block of size 4,802
alloc'd
==2228==    at 0x4C2ABBD: malloc (vg_replace_malloc.c:296)
==2228==    by 0x509F15: ff_malloc (ff_host_interface.c:89)
==2228==    by 0x4053BE: malloc (ff_glue.c:1021)
==2228==    by 0x4E054E: AliasSctpInit (alias_sctp.c:632)
==2228==    by 0x4DE565: LibAliasInit (alias_db.c:2503)
==2228==    by 0x4E9B3B: nat44_config (ip_fw_nat.c:505)
==2228==    by 0x4E9E91: nat44_cfg (ip_fw_nat.c:599)
==2228==    by 0x4F1719: ipfw_ctl3 (ip_fw_sockopt.c:3666)
==2228==    by 0x4B9954: rip_ctloutput (raw_ip.c:659)
==2228==    by 0x447E11: sosetopt (uipc_socket.c:2505)
==2228==    by 0x44BF4D: kern_setsockopt (uipc_syscalls.c:1407)
==2228==    by 0x409F08: ff_setsockopt (ff_syscall_wrapper.c:412)
==2228==

The error line is:
`la->sctpNatTimer.TimerQ = sn_calloc(SN_TIMER_QUEUE_SIZE, sizeof(struct
sctpTimerQ));`

Since SN_TIMER_QUEUE_SIZE is defined as SN_MAX_TIMER+2, and sn_calloc is
defined as sn_malloc(x * n) if _SYS_MALLOC_H_ is defined, the size of
calloced memory will be wrong, because the macro will be expanded to
sizeof(struct sctpTimerQ)*SN_MAX_TIMER+2.

And the memory will be out of bounds here.
```
/* Initialise circular timer Q*/
for (i = 0; i < SN_TIMER_QUEUE_SIZE; i++)
    LIST_INIT(&la->sctpNatTimer.TimerQ[i]);
```
2017-12-05 15:32:10 +08:00
logwang bc0ac4755a Fix bug: incorrect usage of `rte_pktmbuf_clone` when dispatching arp packets.
Since f-stack uses `rte_pktmbuf_clone` to copy mbuf to other lcores when dispatching arp packets, but it doesn't real copy the packet data. The buf_addr of pktmbuf is pointed to the same address.

The arp response packet is generated with the same mbuf from the request
packet, it just swaps the src and dst address, so the copied mbufs will also be changed.

What we need is a deep copy function, and the arp packets are really small, so deep copy will not harm performance too much.

Fix #53 #111 #112.
2017-12-01 17:25:47 +08:00
logwang 48b700b73c Fix bug: incorrect usage of `rte_pktmbuf_clone` when dispatching arp packets.
Since f-stack uses `rte_pktmbuf_clone` to copy mbuf to other lcores when dispatching arp packets, but it doesn't real copy the packet data. The buf_addr of pktmbuf is pointed to the same address.

The arp response packet is generated with the same mbuf from the request
packet, it just swaps the src and dst address, so the copied mbufs will also be changed.

What we need is a deep copy function, and the arp packets are really small, so deep copy will not harm performance too much.

Fix #53 #111 #112.
2017-12-01 17:25:47 +08:00
logwang ffdb15be57 Configuration: optimize the description. 2017-12-01 16:53:57 +08:00
logwang aa3f528959 Configuration: optimize the description. 2017-12-01 16:53:57 +08:00
logwang 112c74aa2e Fix #107: some invalid usages of kqueue in `ff_epoll`.
1.Both EVFILT_READ and EVFILT_WRITE are values but not flags. It needs to check whether it is equal but not to do logic and.

2.If the read direction of the socket has shutdown, then the filter also sets EV_EOF in `flags`, and returns the socket error (if any) in `fflags`.
2017-11-27 18:39:34 +08:00
logwang 8c317b2231 Fix #107: some invalid usages of kqueue in `ff_epoll`.
1.Both EVFILT_READ and EVFILT_WRITE are values but not flags. It needs to check whether it is equal but not to do logic and.

2.If the read direction of the socket has shutdown, then the filter also sets EV_EOF in `flags`, and returns the socket error (if any) in `fflags`.
2017-11-27 18:39:34 +08:00
logwang 3c4a58ceba Roadmap: to support IPV6. 2017-11-27 14:03:46 +08:00
logwang e76966dcf1 Roadmap: to support IPV6. 2017-11-27 14:03:46 +08:00
logwang 564e18ab25 Merge pull request #103 from tigerjibo/master
ipfw: should init dpdk before all malloc
2017-11-21 14:57:35 +08:00
logwang c598cee34c
Merge pull request #103 from tigerjibo/master
ipfw: should init dpdk before all malloc
2017-11-21 14:57:35 +08:00
Ji Bo 5e0cf829d9 ipfw: should init dpdk before all malloc
Signed-off-by: Ji Bo <jibo@xdja.com>
2017-11-21 14:32:00 +08:00
Ji Bo 896dcc2b9a ipfw: should init dpdk before all malloc
Signed-off-by: Ji Bo <jibo@xdja.com>
2017-11-21 14:32:00 +08:00
logwang 4e73363e2c Release-1.11 tag. 2017-11-21 11:42:39 +08:00
logwang 595de2db4b Release-1.11 tag. 2017-11-21 11:42:39 +08:00
logwang 02610d5877 Example: keep run when accept failed. 2017-11-21 11:42:09 +08:00
logwang c7fe235229 Example: keep run when accept failed. 2017-11-21 11:42:09 +08:00
logwang 49e481768a Use general protocol header length in protocol filter. 2017-11-21 11:20:14 +08:00
logwang f17ba62bb6 Use general protocol header length in protocol filter. 2017-11-21 11:20:14 +08:00
logwang 5ac59bc49a Micro_thread: convert encoding of source files and remove chinese comments. 2017-11-20 22:39:00 +08:00
logwang 35a813994b Micro_thread: convert encoding of source files and remove chinese comments. 2017-11-20 22:39:00 +08:00
logwang 8cfa29232a Fix #100: tcp anomaly when TSO is enabled.
1.Fix the wrong ip header length and tcp header length.
2.Recalculate the pseudo header checksum and set it to tcp header.
2017-11-20 17:07:10 +08:00
logwang ef39c58b67 Fix #100: tcp anomaly when TSO is enabled.
1.Fix the wrong ip header length and tcp header length.
2.Recalculate the pseudo header checksum and set it to tcp header.
2017-11-20 17:07:10 +08:00
logwang b1bb619a90 Nginx: fix that worker may crash when reload or recv signals. 2017-11-20 17:06:02 +08:00
logwang ee61971c95 Nginx: fix that worker may crash when reload or recv signals. 2017-11-20 17:06:02 +08:00