From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org, Eric Dumazet <edumazet@google.com>,
Soheil Hassas Yeganeh <soheil@google.com>,
Neal Cardwell <ncardwell@google.com>,
Wei Wang <weiwan@google.com>,
"David S. Miller" <davem@davemloft.net>
Subject: [PATCH 4.9 06/44] tcp: tcp_grow_window() needs to respect tcp_space()
Date: Wed, 24 Apr 2019 19:09:44 +0200 [thread overview]
Message-ID: <20190424170844.368846445@linuxfoundation.org> (raw)
In-Reply-To: <20190424170839.924291114@linuxfoundation.org>
From: Eric Dumazet <edumazet@google.com>
[ Upstream commit 50ce163a72d817a99e8974222dcf2886d5deb1ae ]
For some reason, tcp_grow_window() correctly tests if enough room
is present before attempting to increase tp->rcv_ssthresh,
but does not prevent it to grow past tcp_space()
This is causing hard to debug issues, like failing
the (__tcp_select_window(sk) >= tp->rcv_wnd) test
in __tcp_ack_snd_check(), causing ACK delays and possibly
slow flows.
Depending on tcp_rmem[2], MTU, skb->len/skb->truesize ratio,
we can see the problem happening on "netperf -t TCP_RR -- -r 2000,2000"
after about 60 round trips, when the active side no longer sends
immediate acks.
This bug predates git history.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Wei Wang <weiwan@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
net/ipv4/tcp_input.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -389,11 +389,12 @@ static int __tcp_grow_window(const struc
static void tcp_grow_window(struct sock *sk, const struct sk_buff *skb)
{
struct tcp_sock *tp = tcp_sk(sk);
+ int room;
+
+ room = min_t(int, tp->window_clamp, tcp_space(sk)) - tp->rcv_ssthresh;
/* Check #1 */
- if (tp->rcv_ssthresh < tp->window_clamp &&
- (int)tp->rcv_ssthresh < tcp_space(sk) &&
- !tcp_under_memory_pressure(sk)) {
+ if (room > 0 && !tcp_under_memory_pressure(sk)) {
int incr;
/* Check #2. Increase window, if skb with such overhead
@@ -406,8 +407,7 @@ static void tcp_grow_window(struct sock
if (incr) {
incr = max_t(int, incr, 2 * skb->len);
- tp->rcv_ssthresh = min(tp->rcv_ssthresh + incr,
- tp->window_clamp);
+ tp->rcv_ssthresh += min(room, incr);
inet_csk(sk)->icsk_ack.quick |= 1;
}
}
next prev parent reply other threads:[~2019-04-24 17:25 UTC|newest]
Thread overview: 50+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-04-24 17:09 [PATCH 4.9 00/44] 4.9.171-stable review Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 01/44] bonding: fix event handling for stacked bonds Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 02/44] net: atm: Fix potential Spectre v1 vulnerabilities Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 03/44] net: bridge: fix per-port af_packet sockets Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 04/44] net: bridge: multicast: use rcu to access port list from br_multicast_start_querier Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 05/44] net: fou: do not use guehdr after iptunnel_pull_offloads in gue_udp_recv Greg Kroah-Hartman
2019-04-24 17:09 ` Greg Kroah-Hartman [this message]
2019-04-24 17:09 ` [PATCH 4.9 07/44] team: set slave to promisc if team is already in promisc mode Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 08/44] vhost: reject zero size iova range Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 09/44] ipv4: recompile ip options in ipv4_link_failure Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 10/44] ipv4: ensure rcu_read_lock() in ipv4_link_failure() Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 11/44] crypto: crypto4xx - properly set IV after de- and encrypt Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 12/44] mmc: sdhci: Fix data command CRC error handling Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 13/44] modpost: file2alias: go back to simple devtable lookup Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 14/44] modpost: file2alias: check prototype of handler Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 15/44] tpm/tpm_i2c_atmel: Return -E2BIG when the transfer is incomplete Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 16/44] CIFS: keep FileInfo handle live during oplock break Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 17/44] KVM: x86: Dont clear EFER during SMM transitions for 32-bit vCPU Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 18/44] staging: iio: ad7192: Fix ad7193 channel address Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 19/44] iio/gyro/bmg160: Use millidegrees for temperature scale Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 20/44] iio: ad_sigma_delta: select channel when reading register Greg Kroah-Hartman
2019-04-24 17:09 ` [PATCH 4.9 21/44] iio: adc: at91: disable adc channel interrupt in timeout case Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 22/44] io: accel: kxcjk1013: restore the range after resume Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 23/44] staging: comedi: vmk80xx: Fix use of uninitialized semaphore Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 24/44] staging: comedi: vmk80xx: Fix possible double-free of ->usb_rx_buf Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 25/44] staging: comedi: ni_usb6501: Fix use of uninitialized mutex Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 26/44] staging: comedi: ni_usb6501: Fix possible double-free of ->usb_rx_buf Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 27/44] ALSA: core: Fix card races between register and disconnect Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 28/44] Revert "scsi: fcoe: clear FC_RP_STARTED flags when receiving a LOGO" Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 29/44] Revert "svm: Fix AVIC incomplete IPI emulation" Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 30/44] crypto: x86/poly1305 - fix overflow during partial reduction Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 31/44] arm64: futex: Restore oldval initialization to work around buggy compilers Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 32/44] x86/kprobes: Verify stack frame on kretprobe Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 33/44] kprobes: Mark ftrace mcount handler functions nokprobe Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 34/44] kprobes: Fix error check when reusing optimized probes Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 35/44] rt2x00: do not increment sequence number while re-transmitting Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 36/44] mac80211: do not call driver wake_tx_queue op during reconfig Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 37/44] perf/x86/amd: Add event map for AMD Family 17h Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 38/44] Revert "kbuild: use -Oz instead of -Os when using clang" Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 39/44] sched/fair: Limit sched_cfs_period_timer() loop to avoid hard lockup Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 40/44] device_cgroup: fix RCU imbalance in error case Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 41/44] mm/vmstat.c: fix /proc/vmstat format for CONFIG_DEBUG_TLBFLUSH=y CONFIG_SMP=n Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 42/44] ALSA: info: Fix racy addition/deletion of nodes Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 43/44] percpu: stop printing kernel addresses Greg Kroah-Hartman
2019-04-24 17:10 ` [PATCH 4.9 44/44] i2c-hid: properly terminate i2c_hid_dmi_desc_override_table[] array Greg Kroah-Hartman
2019-04-24 21:25 ` [PATCH 4.9 00/44] 4.9.171-stable review kernelci.org bot
2019-04-25 5:24 ` Naresh Kamboju
2019-04-25 11:55 ` Jon Hunter
2019-04-25 16:26 ` shuah
2019-04-25 21:40 ` Guenter Roeck
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190424170844.368846445@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=ncardwell@google.com \
--cc=soheil@google.com \
--cc=stable@vger.kernel.org \
--cc=weiwan@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox