Page MenuHomeFreeBSD
Differential D28292

Replace all uses of kern_mmap with kern_mmap_req
ClosedPublic
Actions

Authored by brooks on Jan 23 2021, 12:25 AM.
Tags
None
Referenced Files
Unknown Object (File)
Sun, Mar 9, 9:51 AM
Unknown Object (File)
Wed, Mar 5, 4:35 PM
Unknown Object (File)
Wed, Mar 5, 10:04 AM
Unknown Object (File)
Thu, Feb 27, 2:51 PM
Unknown Object (File)
Feb 22 2025, 5:52 AM
Unknown Object (File)
Feb 22 2025, 1:36 AM
Unknown Object (File)
Feb 21 2025, 10:15 PM
Unknown Object (File)
Feb 21 2025, 9:33 PM
View All 95 Files
Subscribers
None

Details

Reviewers
kevans
kib
imp
jhb
Commits
rG7a1591c1b625: Rename kern_mmap_req to kern_mmap
Summary

The helper saved some code churn initially, but having multiple
interfaces is sub-optimal.

Remove the old kern_mmap.

Rename kern_mmap_req to kern_mmap

There is no need for the suffix now that there is only one interface.

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
Lint Passed
Unit
No Test Coverage
Build Status
Buildable 36481
Build 33370: arc lint + arc unit

Event Timeline

brooks requested review of this revision.Jan 23 2021, 12:25 AM
brooks created this revision.
Harbormaster completed remote builds in B36418: Diff 82747.Jan 23 2021, 12:25 AM
brooks updated this revision to Diff 82748.Jan 23 2021, 12:33 AM
  • Whitespace fix.
Harbormaster completed remote builds in B36419: Diff 82748.Jan 23 2021, 12:33 AM
brooks added a comment.Jan 23 2021, 12:33 AM

I've done this cleanup in CheriBSD on the path to some local cleanups.

I've chosen to use compound literals because they are IMO tidier than single use temporary variables when the initialization values aren't available at the top of the file. They aren't widely used in the tree so now is probably the time to discuss style. The one aspect of the current style I debated the most was if the &(struct foo){ should be on it's own line or on the line above. I felt splitting it out worked well with the other indentation.

kib accepted this revision.Jan 23 2021, 11:43 AM
kib added inline comments.
sys/compat/cloudabi/cloudabi_mem.c
114

I would fill the previous line by merging with this one.

This revision is now accepted and ready to land.Jan 23 2021, 11:43 AM
brooks updated this revision to Diff 82877.Jan 25 2021, 9:24 PM
  • Rebase
  • Use a slightly more compact style
This revision now requires review to proceed.Jan 25 2021, 9:24 PM
Harbormaster completed remote builds in B36481: Diff 82877.Jan 25 2021, 9:24 PM
brooks marked an inline comment as done.Jan 25 2021, 9:49 PM
kib accepted this revision.Jan 25 2021, 9:49 PM
This revision is now accepted and ready to land.Jan 25 2021, 9:49 PM
Closed by commit rG7a1591c1b625: Rename kern_mmap_req to kern_mmap (authored by brooks). · Explain WhyJan 25 2021, 9:59 PM
This revision was automatically updated to reflect the committed changes.
brooks added a commit: rG7a1591c1b625: Rename kern_mmap_req to kern_mmap.

Revision Contents
Changeset List

PathSize
sys/
compat/
cloudabi/
10 lines
freebsd32/
20 lines
linux/
4 lines
sys/
4 lines
vm/
49 lines

Diff 82877

View Options

sys/compat/cloudabi/cloudabi_mem.c

Show First 20 Lines • Show All 147 Lines • ▼ Show 20 LinesSYSCTL_INT(_net_inet_tcp, OID_AUTO, delayed_ack, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_delack_enabled), 0, &VNET_NAME(tcp_delack_enabled), 0,
"Delay ACK to try and piggyback it onto a data packet"); "Delay ACK to try and piggyback it onto a data packet");
VNET_DEFINE(int, drop_synfin) = 0; VNET_DEFINE(int, drop_synfin) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(drop_synfin), 0, &VNET_NAME(drop_synfin), 0,
"Drop TCP packets with SYN+FIN set"); "Drop TCP packets with SYN+FIN set");
VNET_DEFINE(int, tcp_do_prr_conservative) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr_conservative, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_prr_conservative), 0,
"Do conservative Proportional Rate Reduction");
VNET_DEFINE(int, tcp_do_prr) = 1; VNET_DEFINE(int, tcp_do_prr) = 1;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_prr), 1, &VNET_NAME(tcp_do_prr), 1,
"Enable Proportional Rate Reduction per RFC 6937"); "Enable Proportional Rate Reduction per RFC 6937");
VNET_DEFINE(int, tcp_do_newcwv) = 0; VNET_DEFINE(int, tcp_do_newcwv) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, newcwv, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, newcwv, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_newcwv), 0, &VNET_NAME(tcp_do_newcwv), 0,
▲ Show 20 Lines • Show All 2,401 Lines • ▼ Show 20 Lines if (SEQ_LEQ(th->th_ack, tp->snd_una)) {
else if (!tcp_timer_active(tp, TT_REXMT)) else if (!tcp_timer_active(tp, TT_REXMT))
tp->t_dupacks = 0; tp->t_dupacks = 0;
else if (++tp->t_dupacks > tcprexmtthresh || else if (++tp->t_dupacks > tcprexmtthresh ||
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
cc_ack_received(tp, th, nsegs, cc_ack_received(tp, th, nsegs,
CC_DUPACK); CC_DUPACK);
if (V_tcp_do_prr && if (V_tcp_do_prr &&
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
tcp_do_prr_ack(tp, th, &to); /*
* While dealing with DupAcks,
* always use PRR-CRB
*/
tcp_do_prr_ack(tp, th, &to, 2);
} else if ((tp->t_flags & TF_SACK_PERMIT) && } else if ((tp->t_flags & TF_SACK_PERMIT) &&
(to.to_flags & TOF_SACK) && (to.to_flags & TOF_SACK) &&
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
int awnd; int awnd;
/* /*
* Compute the amount of data in flight first. * Compute the amount of data in flight first.
* We can inject new data into the pipe iff * We can inject new data into the pipe iff
▲ Show 20 Lines • Show All 191 Lines • ▼ Show 20 Linesresume_partialack:
* for the other side's cached packets, retract it. * for the other side's cached packets, retract it.
*/ */
if (IN_FASTRECOVERY(tp->t_flags)) { if (IN_FASTRECOVERY(tp->t_flags)) {
if (SEQ_LT(th->th_ack, tp->snd_recover)) { if (SEQ_LT(th->th_ack, tp->snd_recover)) {
if (tp->t_flags & TF_SACK_PERMIT) if (tp->t_flags & TF_SACK_PERMIT)
if (V_tcp_do_prr && to.to_flags & TOF_SACK) { if (V_tcp_do_prr && to.to_flags & TOF_SACK) {
tcp_timer_activate(tp, TT_REXMT, 0); tcp_timer_activate(tp, TT_REXMT, 0);
tp->t_rtttime = 0; tp->t_rtttime = 0;
tcp_do_prr_ack(tp, th, &to); tcp_do_prr_ack(tp, th, &to, sack_changed);
tp->t_flags |= TF_ACKNOW; tp->t_flags |= TF_ACKNOW;
(void) tcp_output(tp); (void) tcp_output(tp);
} else } else
tcp_sack_partialack(tp, th); tcp_sack_partialack(tp, th);
else else
tcp_newreno_partial_ack(tp, th); tcp_newreno_partial_ack(tp, th);
} else } else
cc_post_recovery(tp, th); cc_post_recovery(tp, th);
} else if (IN_CONGRECOVERY(tp->t_flags)) { } else if (IN_CONGRECOVERY(tp->t_flags)) {
if (SEQ_LT(th->th_ack, tp->snd_recover)) { if (SEQ_LT(th->th_ack, tp->snd_recover)) {
if (V_tcp_do_prr) { if (V_tcp_do_prr) {
tp->sackhint.delivered_data = BYTES_THIS_ACK(tp, th); tp->sackhint.delivered_data = BYTES_THIS_ACK(tp, th);
tp->snd_fack = th->th_ack; tp->snd_fack = th->th_ack;
tcp_do_prr_ack(tp, th, &to); /*
* During ECN cwnd reduction
* always use PRR-SSRB
*/
tcp_do_prr_ack(tp, th, &to, 1);
(void) tcp_output(tp); (void) tcp_output(tp);
} }
} else } else
cc_post_recovery(tp, th); cc_post_recovery(tp, th);
} }
/* /*
* If we reach this point, ACK is not a duplicate, * If we reach this point, ACK is not a duplicate,
* i.e., it ACKs something we sent. * i.e., it ACKs something we sent.
▲ Show 20 Lines • Show All 1,112 Lines • ▼ Show 20 Lines if (maxmtu && thcmtu)
mss = min(maxmtu, thcmtu) - min_protoh; mss = min(maxmtu, thcmtu) - min_protoh;
else if (maxmtu || thcmtu) else if (maxmtu || thcmtu)
mss = max(maxmtu, thcmtu) - min_protoh; mss = max(maxmtu, thcmtu) - min_protoh;
return (mss); return (mss);
} }
void void
tcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to) tcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to, int sack_changed)
{ {
int snd_cnt = 0, limit = 0, del_data = 0, pipe = 0; int snd_cnt = 0, limit = 0, del_data = 0, pipe = 0;
int maxseg = tcp_maxseg(tp); int maxseg = tcp_maxseg(tp);
INP_WLOCK_ASSERT(tp->t_inpcb); INP_WLOCK_ASSERT(tp->t_inpcb);
/* /*
* Compute the amount of data that this ACK is indicating * Compute the amount of data that this ACK is indicating
Show All 24 Linestcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to, int sack_changed)
if (pipe >= tp->snd_ssthresh) { if (pipe >= tp->snd_ssthresh) {
if (tp->sackhint.recover_fs == 0) if (tp->sackhint.recover_fs == 0)
tp->sackhint.recover_fs = tp->sackhint.recover_fs =
imax(1, tp->snd_nxt - tp->snd_una); imax(1, tp->snd_nxt - tp->snd_una);
snd_cnt = howmany((long)tp->sackhint.prr_delivered * snd_cnt = howmany((long)tp->sackhint.prr_delivered *
tp->snd_ssthresh, tp->sackhint.recover_fs) - tp->snd_ssthresh, tp->sackhint.recover_fs) -
tp->sackhint.prr_out; tp->sackhint.prr_out;
} else { } else {
if (V_tcp_do_prr_conservative || (del_data == 0)) /*
* PRR 6937bis heuristic:
* - A partial ack without SACK block beneath snd_recover
* indicates further loss. This case is not handled here,
* but in tcp_sack_partialack, to send a rescue
* retransmissions.
* - An SACK scoreboard update adding a new hole indicates
* further loss, so be conservative and send at most one
* segment.
*/
if ((sack_changed == 2) || (del_data == 0))
limit = tp->sackhint.prr_delivered - limit = tp->sackhint.prr_delivered -
tp->sackhint.prr_out; tp->sackhint.prr_out;
else else
limit = imax(tp->sackhint.prr_delivered - limit = imax(tp->sackhint.prr_delivered -
tp->sackhint.prr_out, del_data) + tp->sackhint.prr_out, del_data) +
maxseg; maxseg;
snd_cnt = imin((tp->snd_ssthresh - pipe), limit); snd_cnt = imin((tp->snd_ssthresh - pipe), limit);
} }
▲ Show 20 Lines • Show All 94 Lines • Show Last 20 Lines
View Options

sys/compat/freebsd32/freebsd32_misc.c

Show First 20 Lines • Show All 147 Lines • ▼ Show 20 LinesSYSCTL_INT(_net_inet_tcp, OID_AUTO, delayed_ack, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_delack_enabled), 0, &VNET_NAME(tcp_delack_enabled), 0,
"Delay ACK to try and piggyback it onto a data packet"); "Delay ACK to try and piggyback it onto a data packet");
VNET_DEFINE(int, drop_synfin) = 0; VNET_DEFINE(int, drop_synfin) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(drop_synfin), 0, &VNET_NAME(drop_synfin), 0,
"Drop TCP packets with SYN+FIN set"); "Drop TCP packets with SYN+FIN set");
VNET_DEFINE(int, tcp_do_prr_conservative) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr_conservative, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_prr_conservative), 0,
"Do conservative Proportional Rate Reduction");
VNET_DEFINE(int, tcp_do_prr) = 1; VNET_DEFINE(int, tcp_do_prr) = 1;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_prr), 1, &VNET_NAME(tcp_do_prr), 1,
"Enable Proportional Rate Reduction per RFC 6937"); "Enable Proportional Rate Reduction per RFC 6937");
VNET_DEFINE(int, tcp_do_newcwv) = 0; VNET_DEFINE(int, tcp_do_newcwv) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, newcwv, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, newcwv, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_newcwv), 0, &VNET_NAME(tcp_do_newcwv), 0,
▲ Show 20 Lines • Show All 2,401 Lines • ▼ Show 20 Lines if (SEQ_LEQ(th->th_ack, tp->snd_una)) {
else if (!tcp_timer_active(tp, TT_REXMT)) else if (!tcp_timer_active(tp, TT_REXMT))
tp->t_dupacks = 0; tp->t_dupacks = 0;
else if (++tp->t_dupacks > tcprexmtthresh || else if (++tp->t_dupacks > tcprexmtthresh ||
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
cc_ack_received(tp, th, nsegs, cc_ack_received(tp, th, nsegs,
CC_DUPACK); CC_DUPACK);
if (V_tcp_do_prr && if (V_tcp_do_prr &&
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
tcp_do_prr_ack(tp, th, &to); /*
* While dealing with DupAcks,
* always use PRR-CRB
*/
tcp_do_prr_ack(tp, th, &to, 2);
} else if ((tp->t_flags & TF_SACK_PERMIT) && } else if ((tp->t_flags & TF_SACK_PERMIT) &&
(to.to_flags & TOF_SACK) && (to.to_flags & TOF_SACK) &&
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
int awnd; int awnd;
/* /*
* Compute the amount of data in flight first. * Compute the amount of data in flight first.
* We can inject new data into the pipe iff * We can inject new data into the pipe iff
▲ Show 20 Lines • Show All 191 Lines • ▼ Show 20 Linesresume_partialack:
* for the other side's cached packets, retract it. * for the other side's cached packets, retract it.
*/ */
if (IN_FASTRECOVERY(tp->t_flags)) { if (IN_FASTRECOVERY(tp->t_flags)) {
if (SEQ_LT(th->th_ack, tp->snd_recover)) { if (SEQ_LT(th->th_ack, tp->snd_recover)) {
if (tp->t_flags & TF_SACK_PERMIT) if (tp->t_flags & TF_SACK_PERMIT)
if (V_tcp_do_prr && to.to_flags & TOF_SACK) { if (V_tcp_do_prr && to.to_flags & TOF_SACK) {
tcp_timer_activate(tp, TT_REXMT, 0); tcp_timer_activate(tp, TT_REXMT, 0);
tp->t_rtttime = 0; tp->t_rtttime = 0;
tcp_do_prr_ack(tp, th, &to); tcp_do_prr_ack(tp, th, &to, sack_changed);
tp->t_flags |= TF_ACKNOW; tp->t_flags |= TF_ACKNOW;
(void) tcp_output(tp); (void) tcp_output(tp);
} else } else
tcp_sack_partialack(tp, th); tcp_sack_partialack(tp, th);
else else
tcp_newreno_partial_ack(tp, th); tcp_newreno_partial_ack(tp, th);
} else } else
cc_post_recovery(tp, th); cc_post_recovery(tp, th);
} else if (IN_CONGRECOVERY(tp->t_flags)) { } else if (IN_CONGRECOVERY(tp->t_flags)) {
if (SEQ_LT(th->th_ack, tp->snd_recover)) { if (SEQ_LT(th->th_ack, tp->snd_recover)) {
if (V_tcp_do_prr) { if (V_tcp_do_prr) {
tp->sackhint.delivered_data = BYTES_THIS_ACK(tp, th); tp->sackhint.delivered_data = BYTES_THIS_ACK(tp, th);
tp->snd_fack = th->th_ack; tp->snd_fack = th->th_ack;
tcp_do_prr_ack(tp, th, &to); /*
* During ECN cwnd reduction
* always use PRR-SSRB
*/
tcp_do_prr_ack(tp, th, &to, 1);
(void) tcp_output(tp); (void) tcp_output(tp);
} }
} else } else
cc_post_recovery(tp, th); cc_post_recovery(tp, th);
} }
/* /*
* If we reach this point, ACK is not a duplicate, * If we reach this point, ACK is not a duplicate,
* i.e., it ACKs something we sent. * i.e., it ACKs something we sent.
▲ Show 20 Lines • Show All 1,112 Lines • ▼ Show 20 Lines if (maxmtu && thcmtu)
mss = min(maxmtu, thcmtu) - min_protoh; mss = min(maxmtu, thcmtu) - min_protoh;
else if (maxmtu || thcmtu) else if (maxmtu || thcmtu)
mss = max(maxmtu, thcmtu) - min_protoh; mss = max(maxmtu, thcmtu) - min_protoh;
return (mss); return (mss);
} }
void void
tcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to) tcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to, int sack_changed)
{ {
int snd_cnt = 0, limit = 0, del_data = 0, pipe = 0; int snd_cnt = 0, limit = 0, del_data = 0, pipe = 0;
int maxseg = tcp_maxseg(tp); int maxseg = tcp_maxseg(tp);
INP_WLOCK_ASSERT(tp->t_inpcb); INP_WLOCK_ASSERT(tp->t_inpcb);
/* /*
* Compute the amount of data that this ACK is indicating * Compute the amount of data that this ACK is indicating
Show All 24 Linestcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to, int sack_changed)
if (pipe >= tp->snd_ssthresh) { if (pipe >= tp->snd_ssthresh) {
if (tp->sackhint.recover_fs == 0) if (tp->sackhint.recover_fs == 0)
tp->sackhint.recover_fs = tp->sackhint.recover_fs =
imax(1, tp->snd_nxt - tp->snd_una); imax(1, tp->snd_nxt - tp->snd_una);
snd_cnt = howmany((long)tp->sackhint.prr_delivered * snd_cnt = howmany((long)tp->sackhint.prr_delivered *
tp->snd_ssthresh, tp->sackhint.recover_fs) - tp->snd_ssthresh, tp->sackhint.recover_fs) -
tp->sackhint.prr_out; tp->sackhint.prr_out;
} else { } else {
if (V_tcp_do_prr_conservative || (del_data == 0)) /*
* PRR 6937bis heuristic:
* - A partial ack without SACK block beneath snd_recover
* indicates further loss. This case is not handled here,
* but in tcp_sack_partialack, to send a rescue
* retransmissions.
* - An SACK scoreboard update adding a new hole indicates
* further loss, so be conservative and send at most one
* segment.
*/
if ((sack_changed == 2) || (del_data == 0))
limit = tp->sackhint.prr_delivered - limit = tp->sackhint.prr_delivered -
tp->sackhint.prr_out; tp->sackhint.prr_out;
else else
limit = imax(tp->sackhint.prr_delivered - limit = imax(tp->sackhint.prr_delivered -
tp->sackhint.prr_out, del_data) + tp->sackhint.prr_out, del_data) +
maxseg; maxseg;
snd_cnt = imin((tp->snd_ssthresh - pipe), limit); snd_cnt = imin((tp->snd_ssthresh - pipe), limit);
} }
▲ Show 20 Lines • Show All 94 Lines • Show Last 20 Lines
View Options

sys/compat/linux/linux_mmap.c

Show First 20 Lines • Show All 147 Lines • ▼ Show 20 LinesSYSCTL_INT(_net_inet_tcp, OID_AUTO, delayed_ack, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_delack_enabled), 0, &VNET_NAME(tcp_delack_enabled), 0,
"Delay ACK to try and piggyback it onto a data packet"); "Delay ACK to try and piggyback it onto a data packet");
VNET_DEFINE(int, drop_synfin) = 0; VNET_DEFINE(int, drop_synfin) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(drop_synfin), 0, &VNET_NAME(drop_synfin), 0,
"Drop TCP packets with SYN+FIN set"); "Drop TCP packets with SYN+FIN set");
VNET_DEFINE(int, tcp_do_prr_conservative) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr_conservative, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_prr_conservative), 0,
"Do conservative Proportional Rate Reduction");
VNET_DEFINE(int, tcp_do_prr) = 1; VNET_DEFINE(int, tcp_do_prr) = 1;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_prr), 1, &VNET_NAME(tcp_do_prr), 1,
"Enable Proportional Rate Reduction per RFC 6937"); "Enable Proportional Rate Reduction per RFC 6937");
VNET_DEFINE(int, tcp_do_newcwv) = 0; VNET_DEFINE(int, tcp_do_newcwv) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, newcwv, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, newcwv, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_newcwv), 0, &VNET_NAME(tcp_do_newcwv), 0,
▲ Show 20 Lines • Show All 2,401 Lines • ▼ Show 20 Lines if (SEQ_LEQ(th->th_ack, tp->snd_una)) {
else if (!tcp_timer_active(tp, TT_REXMT)) else if (!tcp_timer_active(tp, TT_REXMT))
tp->t_dupacks = 0; tp->t_dupacks = 0;
else if (++tp->t_dupacks > tcprexmtthresh || else if (++tp->t_dupacks > tcprexmtthresh ||
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
cc_ack_received(tp, th, nsegs, cc_ack_received(tp, th, nsegs,
CC_DUPACK); CC_DUPACK);
if (V_tcp_do_prr && if (V_tcp_do_prr &&
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
tcp_do_prr_ack(tp, th, &to); /*
* While dealing with DupAcks,
* always use PRR-CRB
*/
tcp_do_prr_ack(tp, th, &to, 2);
} else if ((tp->t_flags & TF_SACK_PERMIT) && } else if ((tp->t_flags & TF_SACK_PERMIT) &&
(to.to_flags & TOF_SACK) && (to.to_flags & TOF_SACK) &&
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
int awnd; int awnd;
/* /*
* Compute the amount of data in flight first. * Compute the amount of data in flight first.
* We can inject new data into the pipe iff * We can inject new data into the pipe iff
▲ Show 20 Lines • Show All 191 Lines • ▼ Show 20 Linesresume_partialack:
* for the other side's cached packets, retract it. * for the other side's cached packets, retract it.
*/ */
if (IN_FASTRECOVERY(tp->t_flags)) { if (IN_FASTRECOVERY(tp->t_flags)) {
if (SEQ_LT(th->th_ack, tp->snd_recover)) { if (SEQ_LT(th->th_ack, tp->snd_recover)) {
if (tp->t_flags & TF_SACK_PERMIT) if (tp->t_flags & TF_SACK_PERMIT)
if (V_tcp_do_prr && to.to_flags & TOF_SACK) { if (V_tcp_do_prr && to.to_flags & TOF_SACK) {
tcp_timer_activate(tp, TT_REXMT, 0); tcp_timer_activate(tp, TT_REXMT, 0);
tp->t_rtttime = 0; tp->t_rtttime = 0;
tcp_do_prr_ack(tp, th, &to); tcp_do_prr_ack(tp, th, &to, sack_changed);
tp->t_flags |= TF_ACKNOW; tp->t_flags |= TF_ACKNOW;
(void) tcp_output(tp); (void) tcp_output(tp);
} else } else
tcp_sack_partialack(tp, th); tcp_sack_partialack(tp, th);
else else
tcp_newreno_partial_ack(tp, th); tcp_newreno_partial_ack(tp, th);
} else } else
cc_post_recovery(tp, th); cc_post_recovery(tp, th);
} else if (IN_CONGRECOVERY(tp->t_flags)) { } else if (IN_CONGRECOVERY(tp->t_flags)) {
if (SEQ_LT(th->th_ack, tp->snd_recover)) { if (SEQ_LT(th->th_ack, tp->snd_recover)) {
if (V_tcp_do_prr) { if (V_tcp_do_prr) {
tp->sackhint.delivered_data = BYTES_THIS_ACK(tp, th); tp->sackhint.delivered_data = BYTES_THIS_ACK(tp, th);
tp->snd_fack = th->th_ack; tp->snd_fack = th->th_ack;
tcp_do_prr_ack(tp, th, &to); /*
* During ECN cwnd reduction
* always use PRR-SSRB
*/
tcp_do_prr_ack(tp, th, &to, 1);
(void) tcp_output(tp); (void) tcp_output(tp);
} }
} else } else
cc_post_recovery(tp, th); cc_post_recovery(tp, th);
} }
/* /*
* If we reach this point, ACK is not a duplicate, * If we reach this point, ACK is not a duplicate,
* i.e., it ACKs something we sent. * i.e., it ACKs something we sent.
▲ Show 20 Lines • Show All 1,112 Lines • ▼ Show 20 Lines if (maxmtu && thcmtu)
mss = min(maxmtu, thcmtu) - min_protoh; mss = min(maxmtu, thcmtu) - min_protoh;
else if (maxmtu || thcmtu) else if (maxmtu || thcmtu)
mss = max(maxmtu, thcmtu) - min_protoh; mss = max(maxmtu, thcmtu) - min_protoh;
return (mss); return (mss);
} }
void void
tcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to) tcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to, int sack_changed)
{ {
int snd_cnt = 0, limit = 0, del_data = 0, pipe = 0; int snd_cnt = 0, limit = 0, del_data = 0, pipe = 0;
int maxseg = tcp_maxseg(tp); int maxseg = tcp_maxseg(tp);
INP_WLOCK_ASSERT(tp->t_inpcb); INP_WLOCK_ASSERT(tp->t_inpcb);
/* /*
* Compute the amount of data that this ACK is indicating * Compute the amount of data that this ACK is indicating
Show All 24 Linestcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to, int sack_changed)
if (pipe >= tp->snd_ssthresh) { if (pipe >= tp->snd_ssthresh) {
if (tp->sackhint.recover_fs == 0) if (tp->sackhint.recover_fs == 0)
tp->sackhint.recover_fs = tp->sackhint.recover_fs =
imax(1, tp->snd_nxt - tp->snd_una); imax(1, tp->snd_nxt - tp->snd_una);
snd_cnt = howmany((long)tp->sackhint.prr_delivered * snd_cnt = howmany((long)tp->sackhint.prr_delivered *
tp->snd_ssthresh, tp->sackhint.recover_fs) - tp->snd_ssthresh, tp->sackhint.recover_fs) -
tp->sackhint.prr_out; tp->sackhint.prr_out;
} else { } else {
if (V_tcp_do_prr_conservative || (del_data == 0)) /*
* PRR 6937bis heuristic:
* - A partial ack without SACK block beneath snd_recover
* indicates further loss. This case is not handled here,
* but in tcp_sack_partialack, to send a rescue
* retransmissions.
* - An SACK scoreboard update adding a new hole indicates
* further loss, so be conservative and send at most one
* segment.
*/
if ((sack_changed == 2) || (del_data == 0))
limit = tp->sackhint.prr_delivered - limit = tp->sackhint.prr_delivered -
tp->sackhint.prr_out; tp->sackhint.prr_out;
else else
limit = imax(tp->sackhint.prr_delivered - limit = imax(tp->sackhint.prr_delivered -
tp->sackhint.prr_out, del_data) + tp->sackhint.prr_out, del_data) +
maxseg; maxseg;
snd_cnt = imin((tp->snd_ssthresh - pipe), limit); snd_cnt = imin((tp->snd_ssthresh - pipe), limit);
} }
▲ Show 20 Lines • Show All 94 Lines • Show Last 20 Lines
View Options

sys/sys/syscallsubr.h

Show First 20 Lines • Show All 147 Lines • ▼ Show 20 LinesSYSCTL_INT(_net_inet_tcp, OID_AUTO, delayed_ack, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_delack_enabled), 0, &VNET_NAME(tcp_delack_enabled), 0,
"Delay ACK to try and piggyback it onto a data packet"); "Delay ACK to try and piggyback it onto a data packet");
VNET_DEFINE(int, drop_synfin) = 0; VNET_DEFINE(int, drop_synfin) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(drop_synfin), 0, &VNET_NAME(drop_synfin), 0,
"Drop TCP packets with SYN+FIN set"); "Drop TCP packets with SYN+FIN set");
VNET_DEFINE(int, tcp_do_prr_conservative) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr_conservative, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_prr_conservative), 0,
"Do conservative Proportional Rate Reduction");
VNET_DEFINE(int, tcp_do_prr) = 1; VNET_DEFINE(int, tcp_do_prr) = 1;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_prr, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_prr), 1, &VNET_NAME(tcp_do_prr), 1,
"Enable Proportional Rate Reduction per RFC 6937"); "Enable Proportional Rate Reduction per RFC 6937");
VNET_DEFINE(int, tcp_do_newcwv) = 0; VNET_DEFINE(int, tcp_do_newcwv) = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, newcwv, CTLFLAG_VNET | CTLFLAG_RW, SYSCTL_INT(_net_inet_tcp, OID_AUTO, newcwv, CTLFLAG_VNET | CTLFLAG_RW,
&VNET_NAME(tcp_do_newcwv), 0, &VNET_NAME(tcp_do_newcwv), 0,
▲ Show 20 Lines • Show All 2,401 Lines • ▼ Show 20 Lines if (SEQ_LEQ(th->th_ack, tp->snd_una)) {
else if (!tcp_timer_active(tp, TT_REXMT)) else if (!tcp_timer_active(tp, TT_REXMT))
tp->t_dupacks = 0; tp->t_dupacks = 0;
else if (++tp->t_dupacks > tcprexmtthresh || else if (++tp->t_dupacks > tcprexmtthresh ||
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
cc_ack_received(tp, th, nsegs, cc_ack_received(tp, th, nsegs,
CC_DUPACK); CC_DUPACK);
if (V_tcp_do_prr && if (V_tcp_do_prr &&
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
tcp_do_prr_ack(tp, th, &to); /*
* While dealing with DupAcks,
* always use PRR-CRB
*/
tcp_do_prr_ack(tp, th, &to, 2);
} else if ((tp->t_flags & TF_SACK_PERMIT) && } else if ((tp->t_flags & TF_SACK_PERMIT) &&
(to.to_flags & TOF_SACK) && (to.to_flags & TOF_SACK) &&
IN_FASTRECOVERY(tp->t_flags)) { IN_FASTRECOVERY(tp->t_flags)) {
int awnd; int awnd;
/* /*
* Compute the amount of data in flight first. * Compute the amount of data in flight first.
* We can inject new data into the pipe iff * We can inject new data into the pipe iff
▲ Show 20 Lines • Show All 191 Lines • ▼ Show 20 Linesresume_partialack:
* for the other side's cached packets, retract it. * for the other side's cached packets, retract it.
*/ */
if (IN_FASTRECOVERY(tp->t_flags)) { if (IN_FASTRECOVERY(tp->t_flags)) {
if (SEQ_LT(th->th_ack, tp->snd_recover)) { if (SEQ_LT(th->th_ack, tp->snd_recover)) {
if (tp->t_flags & TF_SACK_PERMIT) if (tp->t_flags & TF_SACK_PERMIT)
if (V_tcp_do_prr && to.to_flags & TOF_SACK) { if (V_tcp_do_prr && to.to_flags & TOF_SACK) {
tcp_timer_activate(tp, TT_REXMT, 0); tcp_timer_activate(tp, TT_REXMT, 0);
tp->t_rtttime = 0; tp->t_rtttime = 0;
tcp_do_prr_ack(tp, th, &to); tcp_do_prr_ack(tp, th, &to, sack_changed);
tp->t_flags |= TF_ACKNOW; tp->t_flags |= TF_ACKNOW;
(void) tcp_output(tp); (void) tcp_output(tp);
} else } else
tcp_sack_partialack(tp, th); tcp_sack_partialack(tp, th);
else else
tcp_newreno_partial_ack(tp, th); tcp_newreno_partial_ack(tp, th);
} else } else
cc_post_recovery(tp, th); cc_post_recovery(tp, th);
} else if (IN_CONGRECOVERY(tp->t_flags)) { } else if (IN_CONGRECOVERY(tp->t_flags)) {
if (SEQ_LT(th->th_ack, tp->snd_recover)) { if (SEQ_LT(th->th_ack, tp->snd_recover)) {
if (V_tcp_do_prr) { if (V_tcp_do_prr) {
tp->sackhint.delivered_data = BYTES_THIS_ACK(tp, th); tp->sackhint.delivered_data = BYTES_THIS_ACK(tp, th);
tp->snd_fack = th->th_ack; tp->snd_fack = th->th_ack;
tcp_do_prr_ack(tp, th, &to); /*
* During ECN cwnd reduction
* always use PRR-SSRB
*/
tcp_do_prr_ack(tp, th, &to, 1);
(void) tcp_output(tp); (void) tcp_output(tp);
} }
} else } else
cc_post_recovery(tp, th); cc_post_recovery(tp, th);
} }
/* /*
* If we reach this point, ACK is not a duplicate, * If we reach this point, ACK is not a duplicate,
* i.e., it ACKs something we sent. * i.e., it ACKs something we sent.
▲ Show 20 Lines • Show All 1,112 Lines • ▼ Show 20 Lines if (maxmtu && thcmtu)
mss = min(maxmtu, thcmtu) - min_protoh; mss = min(maxmtu, thcmtu) - min_protoh;
else if (maxmtu || thcmtu) else if (maxmtu || thcmtu)
mss = max(maxmtu, thcmtu) - min_protoh; mss = max(maxmtu, thcmtu) - min_protoh;
return (mss); return (mss);
} }
void void
tcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to) tcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to, int sack_changed)
{ {
int snd_cnt = 0, limit = 0, del_data = 0, pipe = 0; int snd_cnt = 0, limit = 0, del_data = 0, pipe = 0;
int maxseg = tcp_maxseg(tp); int maxseg = tcp_maxseg(tp);
INP_WLOCK_ASSERT(tp->t_inpcb); INP_WLOCK_ASSERT(tp->t_inpcb);
/* /*
* Compute the amount of data that this ACK is indicating * Compute the amount of data that this ACK is indicating
Show All 24 Linestcp_do_prr_ack(struct tcpcb *tp, struct tcphdr *th, struct tcpopt *to, int sack_changed)
if (pipe >= tp->snd_ssthresh) { if (pipe >= tp->snd_ssthresh) {
if (tp->sackhint.recover_fs == 0) if (tp->sackhint.recover_fs == 0)
tp->sackhint.recover_fs = tp->sackhint.recover_fs =
imax(1, tp->snd_nxt - tp->snd_una); imax(1, tp->snd_nxt - tp->snd_una);
snd_cnt = howmany((long)tp->sackhint.prr_delivered * snd_cnt = howmany((long)tp->sackhint.prr_delivered *
tp->snd_ssthresh, tp->sackhint.recover_fs) - tp->snd_ssthresh, tp->sackhint.recover_fs) -
tp->sackhint.prr_out; tp->sackhint.prr_out;
} else { } else {
if (V_tcp_do_prr_conservative || (del_data == 0)) /*
* PRR 6937bis heuristic:
* - A partial ack without SACK block beneath snd_recover
* indicates further loss. This case is not handled here,
* but in tcp_sack_partialack, to send a rescue
* retransmissions.
* - An SACK scoreboard update adding a new hole indicates
* further loss, so be conservative and send at most one
* segment.
*/
if ((sack_changed == 2) || (del_data == 0))
limit = tp->sackhint.prr_delivered - limit = tp->sackhint.prr_delivered -
tp->sackhint.prr_out; tp->sackhint.prr_out;
else else
limit = imax(tp->sackhint.prr_delivered - limit = imax(tp->sackhint.prr_delivered -
tp->sackhint.prr_out, del_data) + tp->sackhint.prr_out, del_data) +
maxseg; maxseg;
snd_cnt = imin((tp->snd_ssthresh - pipe), limit); snd_cnt = imin((tp->snd_ssthresh - pipe), limit);
} }
▲ Show 20 Lines • Show All 94 Lines • Show Last 20 Lines
View Options

sys/vm/vm_mmap.c

Loading...