git: dd72b2f3de4f - stable/12 - Track kern.ipc.somaxconn -> kern.ipc.soacceptqueue rename
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Fri, 14 Oct 2022 00:43:08 UTC
The branch stable/12 has been updated by emaste: URL: https://cgit.FreeBSD.org/src/commit/?id=dd72b2f3de4f41ac612bcbf9ac1d0f294b1ed277 commit dd72b2f3de4f41ac612bcbf9ac1d0f294b1ed277 Author: Jose Luis Duran <jlduran@gmail.com> AuthorDate: 2022-10-11 16:39:26 +0000 Commit: Ed Maste <emaste@FreeBSD.org> CommitDate: 2022-10-14 00:42:51 +0000 Track kern.ipc.somaxconn -> kern.ipc.soacceptqueue rename Fixes: 2bdf61ca29d0 ("Hide the unfortunate named sysctl...") (cherry picked from commit 9871ae6aa91f82a6932be7fcfa1023b077be8e2e) (cherry picked from commit f4f6bc636787e110cbea90ca5822cc5209138125) (cherry picked from commit 67d863df0833cdb1ca054a7c4e7d551b4ad28d95) --- release/tools/gce.conf | 2 +- share/man/man7/tuning.7 | 4 ++-- tools/regression/sockets/listen_backlog/listen_backlog.c | 8 ++++---- 3 files changed, 7 insertions(+), 7 deletions(-) diff --git a/release/tools/gce.conf b/release/tools/gce.conf index f156a5bb93eb..30561acb054c 100644 --- a/release/tools/gce.conf +++ b/release/tools/gce.conf @@ -85,7 +85,7 @@ net.inet.icmp.drop_redirect=1 net.inet.ip.redirect=0 net.inet.tcp.blackhole=2 net.inet.udp.blackhole=1 -kern.ipc.somaxconn=1024 +kern.ipc.soacceptqueue=1024 debug.trace_on_panic=1 debug.debugger_on_panic=0 EOF diff --git a/share/man/man7/tuning.7 b/share/man/man7/tuning.7 index 63e09ecce098..6cfb0dd15bc4 100644 --- a/share/man/man7/tuning.7 +++ b/share/man/man7/tuning.7 @@ -24,7 +24,7 @@ .\" .\" $FreeBSD$ .\" -.Dd April 16, 2020 +.Dd October 11, 2022 .Dt TUNING 7 .Os .Sh NAME @@ -482,7 +482,7 @@ By default is set at the maximum allowable port number. .Pp The -.Va kern.ipc.somaxconn +.Va kern.ipc.soacceptqueue sysctl limits the size of the listen queue for accepting new TCP connections. The default value of 128 is typically too low for robust handling of new connections in a heavily loaded web server environment. diff --git a/tools/regression/sockets/listen_backlog/listen_backlog.c b/tools/regression/sockets/listen_backlog/listen_backlog.c index 2276393ee6f1..9d156c36aa60 100644 --- a/tools/regression/sockets/listen_backlog/listen_backlog.c +++ b/tools/regression/sockets/listen_backlog/listen_backlog.c @@ -44,7 +44,7 @@ * * Future things to test: * - * - That if we change the value of kern.ipc.somaxconn, the limits really + * - That if we change the value of kern.ipc.soacceptqueue, the limits really * do change. * * - That limits are, approximately, enforced and implemented. @@ -55,7 +55,7 @@ */ /* - * We retrieve kern.ipc.somaxconn before running the tests in order to use a + * We retrieve kern.ipc.soacceptqueue before running the tests in order to use a * run-time set value of SOMAXCONN, rather than compile-time set. We assume * that no other process will be simultaneously frobbing it, and these tests * may fail if that assumption is not held. @@ -371,9 +371,9 @@ main(void) size_t len; len = sizeof(somaxconn); - if (sysctlbyname("kern.ipc.somaxconn", &somaxconn, &len, NULL, 0) + if (sysctlbyname("kern.ipc.soacceptqueue", &somaxconn, &len, NULL, 0) < 0) - err(-1, "sysctlbyname(kern.ipc.somaxconn)"); + err(-1, "sysctlbyname(kern.ipc.soacceptqueue)"); test_defaults(); test_listen_update();