diff options
author | dimitri staessens <[email protected]> | 2017-04-01 22:42:21 +0200 |
---|---|---|
committer | dimitri staessens <[email protected]> | 2017-04-01 22:42:21 +0200 |
commit | 4c393500bcf2b35166b98f0def4c767db712cc9e (patch) | |
tree | 3a043ed97b99e15bccce38c4b279d1d2615d2c16 /src/ipcpd/normal | |
parent | ab588700c2d55e427f776c03ae9b7ae5e008f2c1 (diff) | |
download | ouroboros-4c393500bcf2b35166b98f0def4c767db712cc9e.tar.gz ouroboros-4c393500bcf2b35166b98f0def4c767db712cc9e.zip |
ipcpd: Revert timeout in fmgr
Since there are no SDUs on most QoS cubes, the fmgr was always timing
out on most QoS cubes, causing considerable delays in most SDUs. The
timeout was reset to a very small value (so the bug in glibc 2.25 will
cause lockups again).
Diffstat (limited to 'src/ipcpd/normal')
-rw-r--r-- | src/ipcpd/normal/fmgr.c | 7 |
1 files changed, 1 insertions, 6 deletions
diff --git a/src/ipcpd/normal/fmgr.c b/src/ipcpd/normal/fmgr.c index 2c94160a..19653430 100644 --- a/src/ipcpd/normal/fmgr.c +++ b/src/ipcpd/normal/fmgr.c @@ -52,11 +52,7 @@ #include "flow_alloc.pb-c.h" typedef FlowAllocMsg flow_alloc_msg_t; -/* - * NOTE: setting this too low may lead to missed pthread - * cancellations when using glibc 2.25. Bug reported to glibc. - */ -#define FD_UPDATE_TIMEOUT 10000000 /* nanoseconds */ +#define FD_UPDATE_TIMEOUT 10000 /* nanoseconds */ struct { flow_set_t * np1_set[QOS_CUBE_MAX]; @@ -147,7 +143,6 @@ static void * fmgr_np1_sdu_reader(void * o) } pthread_rwlock_unlock(&fmgr.np1_flows_lock); - } } |