| Message ID | 1608271876-120934-1-git-send-email-liyonglong@chinatelecom.cn |
|---|---|
| State | New |
| Headers | show |
| Series | tcp: remove obsolete paramter sysctl_tcp_low_latency | expand |
在 2020/12/19 8:46, Jakub Kicinski 写道: > On Fri, 18 Dec 2020 14:11:16 +0800 lyl wrote: >> Remove tcp_low_latency, since it is not functional After commit >> e7942d0633c4 (tcp: remove prequeue support) >> >> Signed-off-by: lyl <liyonglong@chinatelecom.cn> > > I don't think we can remove sysctls, even if they no longer control > the behavior of the kernel. The existence of the file itself is uAPI. > Got it. But a question: why tcp_tw_recycle can be removed totally? it is also part of uAPI
On Thu, 7 Jan 2021 11:08:16 +0800 lll wrote: > 在 2020/12/19 8:46, Jakub Kicinski 写道: > > On Fri, 18 Dec 2020 14:11:16 +0800 lyl wrote: > >> Remove tcp_low_latency, since it is not functional After commit > >> e7942d0633c4 (tcp: remove prequeue support) > >> > >> Signed-off-by: lyl <liyonglong@chinatelecom.cn> > > > > I don't think we can remove sysctls, even if they no longer control > > the behavior of the kernel. The existence of the file itself is uAPI. > > Got it. But a question: why tcp_tw_recycle can be removed totally? > it is also part of uAPI Good question, perhaps with tcp_tw_recycle we wanted to make sure users who depended on it notice removal, since the feature was broken by design? tcp_low_latency is an optimization, not functionality which users may depend on. But I may be wrong so CCing authors.
Jakub Kicinski <kuba@kernel.org> wrote: > > Got it. But a question: why tcp_tw_recycle can be removed totally? > > it is also part of uAPI > > Good question, perhaps with tcp_tw_recycle we wanted to make sure users > who depended on it notice removal, since the feature was broken by > design? > > tcp_low_latency is an optimization, not functionality which users may > depend on. > > But I may be wrong so CCing authors. I guess it was just a case of 'noone noticed'. I'm not sure if anyone would notice dropping lowlatency sysctl, was just a case of 'overly careful'. Personally I'd rather have them gone so 'sysctl tcp.bla' shows if the feature exists/does something.
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c index 3e5f4f2..d03e4c0 100644 --- a/net/ipv4/sysctl_net_ipv4.c +++ b/net/ipv4/sysctl_net_ipv4.c @@ -51,9 +51,6 @@ static u32 u32_max_div_HZ = UINT_MAX / HZ; static int one_day_secs = 24 * 3600; -/* obsolete */ -static int sysctl_tcp_low_latency __read_mostly; - /* Update system visible IP port range */ static void set_local_port_range(struct net *net, int range[2]) { @@ -501,13 +498,6 @@ static int proc_fib_multipath_hash_policy(struct ctl_table *table, int write, .mode = 0644, .proc_handler = proc_doulongvec_minmax, }, - { - .procname = "tcp_low_latency", - .data = &sysctl_tcp_low_latency, - .maxlen = sizeof(int), - .mode = 0644, - .proc_handler = proc_dointvec - }, #ifdef CONFIG_NETLABEL { .procname = "cipso_cache_enable",
Remove tcp_low_latency, since it is not functional After commit e7942d0633c4 (tcp: remove prequeue support) Signed-off-by: lyl <liyonglong@chinatelecom.cn> --- net/ipv4/sysctl_net_ipv4.c | 10 ---------- 1 file changed, 10 deletions(-)