summaryrefslogtreecommitdiffstats
path: root/net
diff options
context:
space:
mode:
authorEric Dumazet <edumazet@google.com>2019-04-11 05:55:23 -0700
committerDavid S. Miller <davem@davemloft.net>2019-04-11 21:31:03 -0700
commite3058450965972e67cc0e5492c08c4cdadafc134 (patch)
tree896df9274d78b1d868faef05ff1d6089c5b3bf85 /net
parenta5f622984a623df9a84cf43f6b098d8dd76fbe05 (diff)
downloadlinux-e3058450965972e67cc0e5492c08c4cdadafc134.tar.gz
linux-e3058450965972e67cc0e5492c08c4cdadafc134.tar.bz2
linux-e3058450965972e67cc0e5492c08c4cdadafc134.zip
dctcp: more accurate tracking of packets delivery
After commit e21db6f69a95 ("tcp: track total bytes delivered with ECN CE marks") core TCP stack does a very good job tracking ECN signals. The "sender's best estimate of CE information" Yuchung mentioned in his patch is indeed the best we can do. DCTCP can use tp->delivered_ce and tp->delivered to not duplicate the logic, and use the existing best estimate. This solves some problems, since current DCTCP logic does not deal with losses and/or GRO or ack aggregation very well. This also removes a dubious use of inet_csk(sk)->icsk_ack.rcv_mss (this should have been tp->mss_cache), and a 64 bit divide. Finally, we can see that the DCTCP logic, calling dctcp_update_alpha() for every ACK could be done differently, calling it only once per RTT. Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Yuchung Cheng <ycheng@google.com> Cc: Neal Cardwell <ncardwell@google.com> Cc: Soheil Hassas Yeganeh <soheil@google.com> Cc: Florian Westphal <fw@strlen.de> Cc: Daniel Borkmann <daniel@iogearbox.net> Cc: Lawrence Brakmo <brakmo@fb.com> Cc: Abdul Kabbani <akabbani@google.com> Acked-by: Soheil Hassas Yeganeh <soheil@google.com> Acked-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net')
-rw-r--r--net/ipv4/tcp_dctcp.c45
1 files changed, 17 insertions, 28 deletions
diff --git a/net/ipv4/tcp_dctcp.c b/net/ipv4/tcp_dctcp.c
index 359da68d7c06..477cb4aa456c 100644
--- a/net/ipv4/tcp_dctcp.c
+++ b/net/ipv4/tcp_dctcp.c
@@ -49,9 +49,8 @@
#define DCTCP_MAX_ALPHA 1024U
struct dctcp {
- u32 acked_bytes_ecn;
- u32 acked_bytes_total;
- u32 prior_snd_una;
+ u32 old_delivered;
+ u32 old_delivered_ce;
u32 prior_rcv_nxt;
u32 dctcp_alpha;
u32 next_seq;
@@ -73,8 +72,8 @@ static void dctcp_reset(const struct tcp_sock *tp, struct dctcp *ca)
{
ca->next_seq = tp->snd_nxt;
- ca->acked_bytes_ecn = 0;
- ca->acked_bytes_total = 0;
+ ca->old_delivered = tp->delivered;
+ ca->old_delivered_ce = tp->delivered_ce;
}
static void dctcp_init(struct sock *sk)
@@ -86,7 +85,6 @@ static void dctcp_init(struct sock *sk)
sk->sk_state == TCP_CLOSE)) {
struct dctcp *ca = inet_csk_ca(sk);
- ca->prior_snd_una = tp->snd_una;
ca->prior_rcv_nxt = tp->rcv_nxt;
ca->dctcp_alpha = min(dctcp_alpha_on_init, DCTCP_MAX_ALPHA);
@@ -118,37 +116,25 @@ static void dctcp_update_alpha(struct sock *sk, u32 flags)
{
const struct tcp_sock *tp = tcp_sk(sk);
struct dctcp *ca = inet_csk_ca(sk);
- u32 acked_bytes = tp->snd_una - ca->prior_snd_una;
-
- /* If ack did not advance snd_una, count dupack as MSS size.
- * If ack did update window, do not count it at all.
- */
- if (acked_bytes == 0 && !(flags & CA_ACK_WIN_UPDATE))
- acked_bytes = inet_csk(sk)->icsk_ack.rcv_mss;
- if (acked_bytes) {
- ca->acked_bytes_total += acked_bytes;
- ca->prior_snd_una = tp->snd_una;
-
- if (flags & CA_ACK_ECE)
- ca->acked_bytes_ecn += acked_bytes;
- }
/* Expired RTT */
if (!before(tp->snd_una, ca->next_seq)) {
- u64 bytes_ecn = ca->acked_bytes_ecn;
+ u32 delivered_ce = tp->delivered_ce - ca->old_delivered_ce;
u32 alpha = ca->dctcp_alpha;
/* alpha = (1 - g) * alpha + g * F */
alpha -= min_not_zero(alpha, alpha >> dctcp_shift_g);
- if (bytes_ecn) {
+ if (delivered_ce) {
+ u32 delivered = tp->delivered - ca->old_delivered;
+
/* If dctcp_shift_g == 1, a 32bit value would overflow
- * after 8 Mbytes.
+ * after 8 M packets.
*/
- bytes_ecn <<= (10 - dctcp_shift_g);
- do_div(bytes_ecn, max(1U, ca->acked_bytes_total));
+ delivered_ce <<= (10 - dctcp_shift_g);
+ delivered_ce /= max(1U, delivered);
- alpha = min(alpha + (u32)bytes_ecn, DCTCP_MAX_ALPHA);
+ alpha = min(alpha + delivered_ce, DCTCP_MAX_ALPHA);
}
/* dctcp_alpha can be read from dctcp_get_info() without
* synchro, so we ask compiler to not use dctcp_alpha
@@ -200,6 +186,7 @@ static size_t dctcp_get_info(struct sock *sk, u32 ext, int *attr,
union tcp_cc_info *info)
{
const struct dctcp *ca = inet_csk_ca(sk);
+ const struct tcp_sock *tp = tcp_sk(sk);
/* Fill it also in case of VEGASINFO due to req struct limits.
* We can still correctly retrieve it later.
@@ -211,8 +198,10 @@ static size_t dctcp_get_info(struct sock *sk, u32 ext, int *attr,
info->dctcp.dctcp_enabled = 1;
info->dctcp.dctcp_ce_state = (u16) ca->ce_state;
info->dctcp.dctcp_alpha = ca->dctcp_alpha;
- info->dctcp.dctcp_ab_ecn = ca->acked_bytes_ecn;
- info->dctcp.dctcp_ab_tot = ca->acked_bytes_total;
+ info->dctcp.dctcp_ab_ecn = tp->mss_cache *
+ (tp->delivered_ce - ca->old_delivered_ce);
+ info->dctcp.dctcp_ab_tot = tp->mss_cache *
+ (tp->delivered - ca->old_delivered);
}
*attr = INET_DIAG_DCTCPINFO;