Delete obsolete /* ARGSUSED */ lint comments.

ok miod@ millert@

Apply same cleanup to the control code as in ospfd.
Internalize control_state and ctl_conns.

Pledge ripe and rde.
Use unveil to remove file system access for the parent proc.

Also remove control_cleanup as discussed on tech@. It allows to fully remove
fs access (ripd does not support reload). It's not an issue for ripd and
ripctrl if an unused socket file is lying around.

reads OK for claudio@ and mestre@

Nuke some whitespace that keeps poking me in the eye as I try to
steal code.

Replace hand-rolled for(;;) traversal of ctl_conns TAILQ with
TAILQ_FOREACH().

No intentional functional change.

ok reyk@

Allow specifying an alternate socket path.

This allows one to run multiple ripd instances, for example to serve
multiple rdomains. Diff from Nima GHOTBI, ok claudio@ florian@ benno@

EAGAIN handling for imsg_read. OK henning@ benno@

More session_socket_blockmode() removal. While there make the main UDP
socket non-blocking by default.

Close the control fd when it has reported EOF.

ok henning@

handle msgbuf_write() returning EAGAIN

"looks right" deraadt and sthen.

handle ECONNABORTED errors from accept(). In many code blocks they can be
ignored silently and without aborting, much like EINTR and EWOULDBLOCK are.
ok's from various maintainers of these directories...

delete excessive evtimer_pending; ok claudio

Handle file descriptor exhaustion in the accept() case.
ok claudio

Use calloc() instead of malloc() to allocate the connection structure. This
way the memory is zeroed. This is equivalent to the relayd commit by reyk.
OK reyk, jsg

Kill double declaration of control_imsg_relay().

From Christiano F. Haesbaert.

ok claudio@

porcesses -> processes

log_warn() consistency.

`OK' claudio

Another routing daemon that gets IMSG_CTL_LOG_VERBOSE to allow toggeling of
log_debug(). ok michele, henning, sthen

make ripctl/ripd imsg-in-a-lib ready too.

ok pyr@

Remove redundant imsg_event_add calls; ok claudio@

Better EV_READ and EV_WRITE handling in the libevent callbacks. Again mostly
the same ospfd diff applied with some fuzz.

Fix an invalid pointer dereference in control_close(). If control_connbyfd()
fails -- which should never happen -- the function does not return and is
accession the NULL set control pointer later on.
Found by Matthew Haub. OK deraadt@

do not leak confd on malloc failure in control_accept()
found by Igor Zinovik <zinovik@cs.karelia.ru>

Use an ssize_t instead of int like in all other imsg handlers. Hint by lint

Remove redundant code.

do not signal "connection closed" upstream on EINTR/EAGAIN, from bgpd et al

Typo in fatal() message found by dunceor @ gmail dot com

Welcome ripd
started by Michele Marchetto some time ago by using the imsg/three process framework
of ospfd. He implemented most of the daemon with a little help and guidance from
Claudio and I.

Currently the daemon is more or less complete, with the exception of key lifetime
and rollover.

Not yet connected to the builds.
OK claudio@

Apply same cleanup to the control code as in ospfd.
Internalize control_state and ctl_conns.

Pledge ripe and rde.
Use unveil to remove file system access for the parent proc.

Also remove control_cleanup as discussed on tech@. It allows to fully remove
fs access (ripd does not support reload). It's not an issue for ripd and
ripctrl if an unused socket file is lying around.

reads OK for claudio@ and mestre@

Nuke some whitespace that keeps poking me in the eye as I try to
steal code.

Replace hand-rolled for(;;) traversal of ctl_conns TAILQ with
TAILQ_FOREACH().

No intentional functional change.

ok reyk@

Allow specifying an alternate socket path.

This allows one to run multiple ripd instances, for example to serve
multiple rdomains. Diff from Nima GHOTBI, ok claudio@ florian@ benno@

EAGAIN handling for imsg_read. OK henning@ benno@

More session_socket_blockmode() removal. While there make the main UDP
socket non-blocking by default.

Close the control fd when it has reported EOF.

ok henning@

handle msgbuf_write() returning EAGAIN

"looks right" deraadt and sthen.

handle ECONNABORTED errors from accept(). In many code blocks they can be
ignored silently and without aborting, much like EINTR and EWOULDBLOCK are.
ok's from various maintainers of these directories...

delete excessive evtimer_pending; ok claudio

Handle file descriptor exhaustion in the accept() case.
ok claudio

Use calloc() instead of malloc() to allocate the connection structure. This
way the memory is zeroed. This is equivalent to the relayd commit by reyk.
OK reyk, jsg

Kill double declaration of control_imsg_relay().

From Christiano F. Haesbaert.

ok claudio@

porcesses -> processes

log_warn() consistency.

`OK' claudio

Another routing daemon that gets IMSG_CTL_LOG_VERBOSE to allow toggeling of
log_debug(). ok michele, henning, sthen

make ripctl/ripd imsg-in-a-lib ready too.

ok pyr@

Remove redundant imsg_event_add calls; ok claudio@

Better EV_READ and EV_WRITE handling in the libevent callbacks. Again mostly
the same ospfd diff applied with some fuzz.

Fix an invalid pointer dereference in control_close(). If control_connbyfd()
fails -- which should never happen -- the function does not return and is
accession the NULL set control pointer later on.
Found by Matthew Haub. OK deraadt@

do not leak confd on malloc failure in control_accept()
found by Igor Zinovik <zinovik@cs.karelia.ru>

Use an ssize_t instead of int like in all other imsg handlers. Hint by lint

Remove redundant code.

do not signal "connection closed" upstream on EINTR/EAGAIN, from bgpd et al

Typo in fatal() message found by dunceor @ gmail dot com

Welcome ripd
started by Michele Marchetto some time ago by using the imsg/three process framework
of ospfd. He implemented most of the daemon with a little help and guidance from
Claudio and I.

Currently the daemon is more or less complete, with the exception of key lifetime
and rollover.

Not yet connected to the builds.
OK claudio@

Pledge ripe and rde.
Use unveil to remove file system access for the parent proc.

Also remove control_cleanup as discussed on tech@. It allows to fully remove
fs access (ripd does not support reload). It's not an issue for ripd and
ripctrl if an unused socket file is lying around.

reads OK for claudio@ and mestre@

Nuke some whitespace that keeps poking me in the eye as I try to
steal code.

Replace hand-rolled for(;;) traversal of ctl_conns TAILQ with
TAILQ_FOREACH().

No intentional functional change.

ok reyk@

Allow specifying an alternate socket path.

This allows one to run multiple ripd instances, for example to serve
multiple rdomains. Diff from Nima GHOTBI, ok claudio@ florian@ benno@

EAGAIN handling for imsg_read. OK henning@ benno@

More session_socket_blockmode() removal. While there make the main UDP
socket non-blocking by default.

Close the control fd when it has reported EOF.

ok henning@

handle msgbuf_write() returning EAGAIN

"looks right" deraadt and sthen.

handle ECONNABORTED errors from accept(). In many code blocks they can be
ignored silently and without aborting, much like EINTR and EWOULDBLOCK are.
ok's from various maintainers of these directories...

delete excessive evtimer_pending; ok claudio

Handle file descriptor exhaustion in the accept() case.
ok claudio

Use calloc() instead of malloc() to allocate the connection structure. This
way the memory is zeroed. This is equivalent to the relayd commit by reyk.
OK reyk, jsg

Kill double declaration of control_imsg_relay().

From Christiano F. Haesbaert.

ok claudio@

porcesses -> processes

log_warn() consistency.

`OK' claudio

Another routing daemon that gets IMSG_CTL_LOG_VERBOSE to allow toggeling of
log_debug(). ok michele, henning, sthen

make ripctl/ripd imsg-in-a-lib ready too.

ok pyr@

Remove redundant imsg_event_add calls; ok claudio@

Better EV_READ and EV_WRITE handling in the libevent callbacks. Again mostly
the same ospfd diff applied with some fuzz.

Fix an invalid pointer dereference in control_close(). If control_connbyfd()
fails -- which should never happen -- the function does not return and is
accession the NULL set control pointer later on.
Found by Matthew Haub. OK deraadt@

do not leak confd on malloc failure in control_accept()
found by Igor Zinovik <zinovik@cs.karelia.ru>

Use an ssize_t instead of int like in all other imsg handlers. Hint by lint

Remove redundant code.

do not signal "connection closed" upstream on EINTR/EAGAIN, from bgpd et al

Typo in fatal() message found by dunceor @ gmail dot com

Welcome ripd
started by Michele Marchetto some time ago by using the imsg/three process framework
of ospfd. He implemented most of the daemon with a little help and guidance from
Claudio and I.

Currently the daemon is more or less complete, with the exception of key lifetime
and rollover.

Not yet connected to the builds.
OK claudio@

Nuke some whitespace that keeps poking me in the eye as I try to
steal code.

Replace hand-rolled for(;;) traversal of ctl_conns TAILQ with
TAILQ_FOREACH().

No intentional functional change.

ok reyk@

Allow specifying an alternate socket path.

This allows one to run multiple ripd instances, for example to serve
multiple rdomains. Diff from Nima GHOTBI, ok claudio@ florian@ benno@

EAGAIN handling for imsg_read. OK henning@ benno@

More session_socket_blockmode() removal. While there make the main UDP
socket non-blocking by default.

Close the control fd when it has reported EOF.

ok henning@

handle msgbuf_write() returning EAGAIN

"looks right" deraadt and sthen.

handle ECONNABORTED errors from accept(). In many code blocks they can be
ignored silently and without aborting, much like EINTR and EWOULDBLOCK are.
ok's from various maintainers of these directories...

delete excessive evtimer_pending; ok claudio

Handle file descriptor exhaustion in the accept() case.
ok claudio

Use calloc() instead of malloc() to allocate the connection structure. This
way the memory is zeroed. This is equivalent to the relayd commit by reyk.
OK reyk, jsg

Kill double declaration of control_imsg_relay().

From Christiano F. Haesbaert.

ok claudio@

porcesses -> processes

log_warn() consistency.

`OK' claudio

Another routing daemon that gets IMSG_CTL_LOG_VERBOSE to allow toggeling of
log_debug(). ok michele, henning, sthen

make ripctl/ripd imsg-in-a-lib ready too.

ok pyr@

Remove redundant imsg_event_add calls; ok claudio@

Better EV_READ and EV_WRITE handling in the libevent callbacks. Again mostly
the same ospfd diff applied with some fuzz.

Fix an invalid pointer dereference in control_close(). If control_connbyfd()
fails -- which should never happen -- the function does not return and is
accession the NULL set control pointer later on.
Found by Matthew Haub. OK deraadt@

do not leak confd on malloc failure in control_accept()
found by Igor Zinovik <zinovik@cs.karelia.ru>

Use an ssize_t instead of int like in all other imsg handlers. Hint by lint

Remove redundant code.

do not signal "connection closed" upstream on EINTR/EAGAIN, from bgpd et al

Typo in fatal() message found by dunceor @ gmail dot com

Welcome ripd
started by Michele Marchetto some time ago by using the imsg/three process framework
of ospfd. He implemented most of the daemon with a little help and guidance from
Claudio and I.

Currently the daemon is more or less complete, with the exception of key lifetime
and rollover.

Not yet connected to the builds.
OK claudio@