| Age | Commit message (Collapse) | Author |
|---|
|
2005-11-24 Paul Jakma <paul.jakma@sun.com>
* privs.c: (zcaps2sys/solaris) remove unused variable.
(zprivs_state_caps/solaris) Format string missing a
specifier.
|
|
2005-11-14 Paul Jakma <paul.jakma@sun.com>
* (general) pass struct work-queue to callback functions.
* workqueue.h: (struct work_queue) move the state flag
variables to end.
Add an opaque pointer to spec, for user-data global to the
queue.
Pass reference to work_queue to all callbacks.
* workqueue.c: (work_queue_item_remove) pass ref to workqueue
to user callbacks.
(work_queue_run) ditto.
|
|
2005-11-14 Paul Jakma <paul.jakma@sun.com>
* configure.ac: Tell gcc we like C99.
[bug #231] Check and test for stdint.h.
* lib/zebra.h: [bug #231] include stdint, if its there.
|
|
* (general) Add state to detect queue floods. There's no sense
trying to be sparing of CPU resources, if the queue is
flooding and using ever more memory resources. we should just
get on with clearing the queue.
The sense of delay and hold were wrong way around, fix.
* workqueue.h: (struct work_queue) Add status bitfield. Add
'flood' integer to workqueue spec. Add runs_since_clear
counter to workqueue.
* workqueue.c: (work_queue_new) set defaults for delay, hold
and flood.
(work_queue_add) initial schedule should use delay, not hold.
(show_work_queues) Print flood field, conserve whitespace.
(work_queue_unplug) use delay, not hold.
(work_queue_run) consecutive runs should be seperated by hold
time, not delay.
Keep track of number of consecutive runs, go into 'overdrive'
if queue is being flooded, we can't avoid making heavy use of
resources, better to use CPU than ever more RAM.
|
|
2005-11-12 Paul Jakma <paul.jakma@sun.com>
* routemap.c: (vty_show_route_map_entry) call action is
seperate from exit action, latter should still be printed
regardless of whether a call is specified.
|
|
* memtypes.h: Update auto-built file.
|
|
* if.c: (connected_free) use MTYPE for connected label.
memtypes.c: Add MTYPE_CONNECTED_LABEL
|
|
* zebra.h: BSD BYTE_ORDER define isn't available everywhere,
define if needs be.
* checksum.h: new file. checksum.c exports in_cksum, provide
a header for it.
* checksum.c: (in_cksum) callers shouldn't have to know it uses
a u_short internally, change to void *.
* Makefile.am: Add checksum.h
* command.h: remove bogus trailling slash.
* md5.c: (general) Update it for the twentieth century. ANSI
declarations are widely supported now.. Don't include system
headers, only include zebra.h. Use POSIX types (the
alternative is to define u_int64_t in a portable way - rest
of Quagga needs same cleanup).
Make endian-conditional code be compiler conditional rather
than preprocessor conditional, so that breakage gets noticed
quicker.
* md5.h: POSIX types. Get rid of the odd __P() non-ANSI capable
compiler compatibility hack.
|
|
* command.c: Use MTYPE_HOST, MTYPE_STRVEC. Some other fixups,
including fixing some likely leaks in config_write_file.
* vty.c: memory macro usage fixes.
(vty_read_config) fix leak where relative config file is
specified.
|
|
* (general) Cleanup a some calls to XFREE,strdup, etc. to use
the memory.h macros.
* memtypes.c: Add MTYPE_IF_RMAP_NAME, MTYPE_PQUEUE,
MTYPE_PQUEUE_DATA and MTYPE_HOST.
* memtypes.h: update auto-built file.
* if_rmap.c: Use MTYPE_IF_RMAP_NAME.
* pqueue.c: Use the two MTYPE_PQUEUE mtypes for allocations.
|
|
* sockopt.c: (setsockopt_multicast_ipv4) If IP_ADD_MEMBERSHIP
fails with errno equal to EADDRINUSE, then issue an info
message and try IP_DROP_MEMBERSHIP followed by IP_ADD_MEMBERSHIP.
|
|
* memory.c: (show_memory_vty) Omit zero statistics. Print separators
only if some non-zero stats have been printed in between.
(show_separator) New function to print a separator.
(show_memory_all) Keep track of whether a separator is needed
between the different memory statistics groups.
|
|
* memtypes.h: Add MTYPE_OSPF_VERTEX_PARENT (to match memtypes.c).
|
|
* (general) SPF memory management cleanup and fix for rare
double-free bug.
* ospf_spf.h: (struct vertex_parent) New struct to hold parent
specific data, eg the backlink and the parent vertex pointer,
and point to the appropriate general struct vertex_nexthop.
(struct vertex_nexthop) remove parent vertex pointer, so
this struct can be shared across vertices.
(struct vertex) rename list child to list children. Remove
list of nexthops, replace with list of vertex_parents.
* ospf_spf.c: (update_stat) trivial, remove cast from void *.
(vertex_nexthop_new) remove init of parent - field is gone
from struct vertex_nexthop.
(ospf_canonical_nexthops_free) Remove the canonical
vertex_nexthop memory objects. These are the vertex_nexthops
attached to the first level of router vertices from the root.
(vertex_parent_new) new function, create a vertex_parent.
(vertex_parent_free) ditto, but free it.
(ospf_vertex_new) Update to match changes to struct vertex.
(ospf_vertex_free) Recursively free a struct vertex and its
children. The parent list is used as a reference count.
vertex_nexthops must be free seperately, if required.
(ospf_vertex_dump) update to match struct vertex changes.
Print out backlink of parents too.
(ospf_vertex_add_parent) ditto.
(ospf_lsa_has_link) update comment.
(ospf_nexthop_add_unique) removed, not needed anymore.
(ospf_nexthop_merge) ditto.
(ospf_spf_consider_nexthop) renamed to ospf_spf_add_parent.
Simplified to just create vertex_parent and add it.
(ospf_spf_flush_parents) new function, flush out the parent
list.
(ospf_nexthop_calculation) Take the relevant route_lsa_link
as an argument, which simplifies things and removes the need
for the hack in ospf_nexthop_add_unique - ospf_spf_next
already knew exactly which link the cost calculated was for.
Update to match struct vertex changes too.
(ospf_spf_next) Don't create a vertex for W unnecessarily, if
it's there's a vertex already created for W, use it, and
hence there's no need to free it either.
Update some manipulation/comparisons of distance to match.
Flush the parent list if a lower cost path is found.
(ospf_spf_route_free) unused, removed.
(ospf_spf_dump) match the struct vertex changes, and dump the
ifname if possible.
(ospf_spf_calculate) At end of SPF, free the canonical nexthops
and call ospf_vertex_free on the root vertex to free the
entire tree.
* ospf_interface.c: (ospf_vl_set_params) match struct vertex
changes.
* ospf_route.c: (ospf_intra_route_add) ditto
(ospf_route_copy_nexthops_from_vertex) ditto
* memtypes.c: (memory_list_ospf) Add MTYPE_OSPF_VERTEX_PARENT.
|
|
* zebra.h: Declare new functions zebra_route_string() and
zebra_route_char().
* log.c: (zroute_lookup,zebra_route_string,zebra_route_char) New
functions to map zebra route numbers to strings.
* zebra_vty.c: (route_type_str) Remove obsolete function: use new
library function zebra_route_string() instead. Note that there
are a few differences: for IPv6 routes, we now get "ripng" and
"ospf6" instead of the old behavior ("rip" and "ospf").
(route_type_char) Remove obsolete function: ues new library function
zebra_route_char() instead. Note that there is one difference:
the old function returned 'S' for a ZEBRA_ROUTE_SYSTEM route,
whereas the new one returns 'X'.
(vty_show_ip_route_detail,vty_show_ipv6_route_detail) Replace
route_type_str() with zebra_route_string().
(vty_show_ip_route,vty_show_ipv6_route) Replace route_type_char()
with zebra_route_char().
* bgp_vty.c: (bgp_config_write_redistribute) Use new library function
zebra_route_string instead of a local hard-coded table.
* ospf6_asbr.c: Remove local hard-coded tables zroute_name and
zroute_abname. Change the ZROUTE_NAME macro to use new library
function zebra_route_string(). Remove the ZROUTE_ABNAME macro.
(ospf6_asbr_external_route_show): Replace ZROUTE_ABNAME() with
a call to zebra_route_char(), and be sure to fix the format string,
since we now have a char instead of a char *.
* ospf6_zebra.c: Remove local hard-coded tables zebra_route_name and
zebra_route_abname. Note that the zebra_route_name[] table
contained mixed-case strings, whereas the zebra_route_string()
function returns lower-case strings.
(ospf6_zebra_read_ipv6): Change debug message to use new library
function zebra_route_string() instead of zebra_route_name[].
(show_zebra): Use new library function zebra_route_string() instead
of zebra_route_name[].
* ospf_dump.c: Remove local hard-coded table ospf_redistributed_proto.
(ospf_redist_string) New function implemented using new library
function zebra_route_string(). Note that there are a few differences
in the output that will result: the new function returns strings
that are lower-case, whereas the old table was mixed case. Also,
the old table mapped ZEBRA_ROUTE_OSPF6 to "OSPFv3", whereas the
new function returns "ospf6".
* ospfd.h: Remove extern struct message ospf_redistributed_proto[],
and add extern const char *ospf_redist_string(u_int route_type)
instead.
* ospf_asbr.c: (ospf_external_info_add) In two messages, use
ospf_redist_string instead of LOOKUP(ospf_redistributed_proto).
* ospf_vty.c: Remove local hard-coded table distribute_str.
(config_write_ospf_redistribute,config_write_ospf_distribute): Use
new library function zebra_route_string() instead of distribute_str[].
* ospf_zebra.c: (ospf_redistribute_set,ospf_redistribute_unset,
ospf_redistribute_default_set,ospf_redistribute_check)
In debug messages, use ospf_redist_string() instead of
LOOKUP(ospf_redistributed_proto).
* rip_zebra.c: (config_write_rip_redistribute): Remove local hard-coded
table str[]. Replace str[] with calls to new library function
zebra_route_string().
* ripd.c: Remove local hard-coded table route_info[].
(show_ip_rip) Replace uses of str[] with calls to new library
functions zebra_route_char and zebra_route_string.
* ripng_zebra.c: (ripng_redistribute_write) Remove local hard-coded
table str[]. Replace str[i] with new library function
zebra_route_string(i).
* ripngd.c: Remove local hard-coded table route_info[].
(show_ipv6_ripng) Use new library function zebra_route_char() instead
of table route_info[].
|
|
* lib/smux.[ch]: allow to retreive global OID (identified by <0
namelen).
* ospf_ism.c: generate SNMP traps on Interface state change
* ospf_nsm.c: generate SNMP traps on Neighbour state change
* ospf_snmp.[ch]: support for SNMP traps for interface and neighbours.
|
|
* configure.ac: Add the test for Solaris least-privileges. Set
defines for whether capabilities are supported and whether of
the linux or solaris variety.
Add missing-prototypes, missing-declarations, char-subscripts
and cast-qual warnings to default cflags, cause Hasso enjoys warnings,
and we really should clean the remaining ones up. (ie isisd..).
* (*/*main.c) Update the zebra_capabilities_t arrays in the various
daemons to match the changes made in lib/privs.h.
* zebra.h: Solaris capabilities requires priv.h to be included.
* privs.{c,h}: Add support for Solaris Least-Privileges.
privs.h: Reduce some of the abstract capabilities, which do
not have rough equivalents on both systems. Rename the net
related caps to _NET, as they should have been in first
place.
(zprivs_terminate) should take the zebra_privs_t as argument so
that it can update change pointer.
Add an additional privilege state, ZPRIVS_UNKNOWN.
* privs.c: (various capability functions) Add
Solaris privileges variants.
(zprivs_state) Use privs.c specific generic types to
represent various capability/privilege related types, so that
each can be typedef'd as appropriate on each platform.
(zprivs_null_state) static added, to hold the state the null
method should report (should be raised by default, and
LOWERED if zprivs_terminate has been called)
(zprivs_state_null) Report back the zprivs_null_state.
(cap_map) Make it able to map abstract capability to multiple
system capabilities.
(zcaps2sys) Map to abstract capabilities to multiple system
privileges/capabilities.
(zprivs_init) move capability related init to seperate
function, zprivs_caps_init.
(zprivs_terminate) ditto, moved to zprivs_caps_terminate.
Set the change_state callback to the NULL state, so the
user can continue to run and use the callbacks.
|
|
* lib/filer.c: show protocol name in filter_show()
* lib/plist.c: show protocol name in vty_show_prefix_entry()
* routemap.c: show protocol name in vty_show_route_map_entry()
* lib/vty.c: in vty_command(), show protocol name if command unknown
* zebra/zserv.c: Always provide distance fo route add
* ripd/rip_snmp.c: rip2IfConfReceive() sends values in conformance
with RFC. Also PeerDomain is now set to a STRING type.
* ripd/ripd.h: rip_redistribute_add() API includes metric and distance
* ripd/ripd.c: rip_redistribute_add() API i.e. stores metric and distance
Now allows a RIP-route to overcome a redistributed route coming
from a protocol with worse (higher) administrative distance
Metrics from redistribution are shown in show ip rip
* ripd/rip_zebra.c: adapt to the rip_redistribute_add() API, i.e.
provide distance and metric
* ripd/rip_interface.c: adapt to the rip_redistribute_add() API
* ripd/rip_routemap.c: no RMAP_COMPILE_ERROR on (metric > 16) usage
rather a CMD_WARNING, because set metric ius shared with other
protocols using larger values (such as OSPF)
The match metric action takes first external metric if present
(from redistribution) then RIP metric.
|
|
* lib/md5-gnu.h: removed
* lib/md5.h: replaces md5-gnu.h
* lib/Makefile.am: use correct md5.h
* lib/md5.c: import from WIDE
* ospfd/ospf_packet.c: use new md5 API
* ripd/ripd.c: use new md5 API
|
|
* memtypes.{c,h}: Add MTYPE_AS_SEG_DATA.
|
|
* lib/str.[ch]: Add strndup() from glibc.
|
|
* command.c: (install_element) be more robust. Eg, cmd_init
need not have been called, some applications may use other
library subsystems, which call install_element, without the
application wanting commands and hence not calling cmd_init.
|
|
* command.h: (enum node_type) Add BGP_IPV6M_NODE
* command.c: (node_parent) Handle BGP_IPV6M_NODE node
(config_exit, config_end) ditto
* vty.c: (vty_end_config) Handle BGP_IPV6M_NODE node
|
|
* getopt.h: Don't declare getopt (rather than getopt_long), since
quagga doesn't need it.
* getopt.c (getopt): Don't define getopt.
Fixes build breakage on NetBSD, and seems likely to work on most
platforms since it avoids the entire issue of system getopt
declarations and whether they conform to POSIX.2. Note that this
change doesn't address system getopt_long declarations, but also
doesn't change anything about getopt_long.
|
|
* prefix.c: (prefix_ipv4_new, prefix_ipv6_new): Call prefix_new
to allocate the memory to make sure that all struct prefix pointers
point to objects of the same length (avoids memory overruns
on struct prefix assignments).
(prefix_ipv4_free, prefix_ipv6_free): Simply call prefix_free.
It is interesting to note that these functions are never actually
called anywhere in the code. Instead prefix_free was already
being called directly, despite the previous MTYPE incompatibility.
[backport candidate]
|
|
* prefix.c: (ip_masklen) While loop should test that 'pnt' pointer is
in range before dereferencing it.
[backport candidate]
|
|
* getopt.h: add further tests for full getopt declaration on
various systems.
|
|
* memtypes.h: update autobuilt file to match memtypes.c changes
|
|
* bgpd/(general) refcount struct peer and bgp_info, hence allowing us
add work_queues for bgp_process.
* bgpd/bgp_route.h: (struct bgp_info) Add 'lock' field for refcount.
Add bgp_info_{lock,unlock} helper functions.
Add bgp_info_{add,delete} helpers, to remove need for
users managing locking/freeing of bgp_info and bgp_node's.
* bgpd/bgp_table.h: (struct bgp_node) Add a flags field, and
BGP_NODE_PROCESS_SCHEDULED to merge redundant processing of
nodes.
* bgpd/bgp_fsm.h: Make the ON/OFF/ADD/REMOVE macros lock and unlock
peer reference as appropriate.
* bgpd/bgp_damp.c: Remove its internal prototypes for
bgp_info_delete/free. Just use bgp_info_delete.
* bgpd/bgpd.h: (struct bgp_master) Add work_queue pointers.
(struct peer) Add reference count 'lock'
(peer_lock,peer_unlock) New helpers to take/release reference
on struct peer.
* bgpd/bgp_advertise.c: (general) Add peer and bgp_info refcounting
and balance how references are taken and released.
(bgp_advertise_free) release bgp_info reference, if appropriate
(bgp_adj_out_free) unlock peer
(bgp_advertise_clean) leave the adv references alone, or else
call bgp_advertise_free cant unlock them.
(bgp_adj_out_set) lock the peer on new adj's, leave the reference
alone otherwise. lock the new bgp_info reference.
(bgp_adj_in_set) lock the peer reference
(bgp_adj_in_remove) and unlock it here
(bgp_sync_delete) make hash_free on peer conditional, just in
case.
* bgpd/bgp_fsm.c: (general) document that the timers depend on
bgp_event to release a peer reference.
(bgp_fsm_change_status) moved up the file, unchanged.
(bgp_stop) Decrement peer lock as many times as cancel_event
canceled - shouldnt be needed but just in case.
stream_fifo_clean of obuf made conditional, just in case.
(bgp_event) always unlock the peer, regardless of return value
of bgp_fsm_change_status.
* bgpd/bgp_packet.c: (general) change several bgp_stop's to BGP_EVENT's.
(bgp_read) Add a mysterious extra peer_unlock for ACCEPT_PEERs
along with a comment on it.
* bgpd/bgp_route.c: (general) Add refcounting of bgp_info, cleanup
some of the resource management around bgp_info. Refcount peer.
Add workqueues for bgp_process and clear_table.
(bgp_info_new) make static
(bgp_info_free) Ditto, and unlock the peer reference.
(bgp_info_lock,bgp_info_unlock) new exported functions
(bgp_info_add) Add a bgp_info to a bgp_node in correct fashion,
taking care of reference counts.
(bgp_info_delete) do the opposite of bgp_info_add.
(bgp_process_rsclient) Converted into a work_queue work function.
(bgp_process_main) ditto.
(bgp_processq_del) process work queue item deconstructor
(bgp_process_queue_init) process work queue init
(bgp_process) call init function if required, set up queue item
and add to queue, rather than calling process functions directly.
(bgp_rib_remove) let bgp_info_delete manage bgp_info refcounts
(bgp_rib_withdraw) ditto
(bgp_update_rsclient) let bgp_info_add manage refcounts
(bgp_update_main) ditto
(bgp_clear_route_node) clear_node_queue work function, does
per-node aspects of what bgp_clear_route_table did previously
(bgp_clear_node_queue_del) clear_node_queue item delete function
(bgp_clear_node_complete) clear_node_queue completion function,
it unplugs the process queues, which have to be blocked while
clear_node_queue is being processed to prevent a race.
(bgp_clear_node_queue_init) init function for clear_node_queue
work queues
(bgp_clear_route_table) Sets up items onto a workqueue now, rather
than clearing each node directly. Plugs both process queues to
avoid potential race.
(bgp_static_withdraw_rsclient) let bgp_info_{add,delete} manage
bgp_info refcounts.
(bgp_static_update_rsclient) ditto
(bgp_static_update_main) ditto
(bgp_static_update_vpnv4) ditto, remove unneeded cast.
(bgp_static_withdraw) see bgp_static_withdraw_rsclient
(bgp_static_withdraw_vpnv4) ditto
(bgp_aggregate_{route,add,delete}) ditto
(bgp_redistribute_{add,delete,withdraw}) ditto
* bgpd/bgp_vty.c: (peer_rsclient_set_vty) lock rsclient list peer
reference
(peer_rsclient_unset_vty) ditto, but unlock same reference
* bgpd/bgpd.c: (peer_free) handle frees of info to be kept for lifetime
of struct peer.
(peer_lock,peer_unlock) peer refcount helpers
(peer_new) add initial refcounts
(peer_create,peer_create_accept) lock peer as appropriate
(peer_delete) unlock as appropriate, move out some free's to
peer_free.
(peer_group_bind,peer_group_unbind) peer refcounting as
appropriate.
(bgp_create) check CALLOC return value.
(bgp_terminate) free workqueues too.
* lib/memtypes.c: Add MTYPE_BGP_PROCESS_QUEUE and
MTYPE_BGP_CLEAR_NODE_QUEUE
|
|
* memtypes.h: update this auto-built file. (maybe we should just
remove it, is GNU awk a terrible dependency to have?)
|
|
* workqueue.h: Add a WQ_QUEUE_BLOCKED item_status return code,
to allow a queue function to indicate the queue is not
ready/blocked - rather than any problem with the item at hand.
Add a notion of being able to 'plug' and 'unplug' a queue.
Add helpers to plug/unplug a queue.
Add a completion callback, to be called when a queue is emptied.
* workqueue.c: (work_queue_new) remove useless list_free.
(work_queue_schedule) new internal helper function to schedule
queue, if appropriate.
(work_queue_add) use work_queue_schedule
(show_work_queues) Print 'P' if queue is plugged.
(work_queue_plug) new API function, plug a queue - ie prevent it
from 'drained' / processed / scheduled.
(work_queue_unplug) unplug a queue, allowing it to be drained
/ scheduled / processed again.
(work_queue_run) Add support for WQ_QUEUE_BLOCKED.
Add comment for RETRY_NOW case.
Make hysteris more aggresive in ramping up granularity, improves
performance significantly.
Add support for calling completion callback when queue is emptied,
possibly useful for knowing when to unplug a queue.
|
|
* routemap.c: (rmap_onmatch_goto) fix crash if 'continue' command
is used, which does not supply an argv[0].
this is a backport candidate /iff/ the trailing ; is removed
from VTY_GET_INTEGER_RANGE
* vty.h: fix the VTY_GET macros, do {..} while(0) so they have
correct function like syntax in usage.
|
|
* memtypes.awk: use character classes, which work correctly in
all LC_COLLATE environments, unlike A-Z, which doesnt work in
eg estonian collate order. Reported by Hasso.
|
|
* memtypes.c: (memory_list_bgp) add MTYPE_BGP_PEER_HOST
|
|
* sockunion.c: (sockunion_getsockname) use MTYPE_SOCKUNION, not TMP
(sockunion_getpeername) ditto
|
|
* thread.c: (thread_cancel_event) the number of pending events
cancelled is potentially useful information, dont throw it away,
pass it back to the caller.
|
|
* getopt.h: It's not just __GNU_LIBRARY__ which defines
getopt, eg __EXTENSIONS__ does too on SunOS. It still seems
awfully fragile though.
* getopt.c: include zebra.h after config.h, before including
getopt.h so that things at least are consistent..
* getopt1.c: ditto
|
|
* configure.ac: Check for OSes which support passing ifindex in
struct ip_mreq.
* lib/sockopt.c: Add support for BSD style ifindex in ip_mreq.
* ospfd/ospf_network.c: Log ifindex on multicast membership leave/join
events.
|
|
* (general) extern and static'ification of functions in code and
header.
Cleanup any definitions with unspecified arguments.
Add casts for callback assignments where the callback is defined,
typically, as passing void *, but the function being assigned has
some other pointer type defined as its argument, as gcc complains
about casts from void * to X* via function arguments.
Fix some old K&R style function argument definitions.
Add noreturn gcc attribute to some functions, as appropriate.
Add unused gcc attribute to some functions (eg ones meant to help
while debugging)
Add guard defines to headers which were missing them.
* command.c: (install_node) add const qualifier, still doesnt shut
up the warning though, because of the double pointer.
(cmp_node) ditto
* keychain.c: (key_str2time) Add GET_LONG_RANGE() macro, derived
fromn vty.h ones to fix some of the (long) < 0 warnings.
* thread.c: (various) use thread_empty
(cpu_record_hash_key) should cast to uintptr_t, a stdint.h type
* vty.h: Add VTY_GET_IPV4_ADDRESS and VTY_GET_IPV4_PREFIX so they
removed from ospfd/ospf_vty.h
* zebra.h: Move definition of ZEBRA_PORT to here, to remove
dependence of lib on zebra/zserv.h
|
|
|
|
* stream.h: Add comment about the special zero-ing ability of
stream_put.
(stream_recvmsg, stream_write) should return ssize_t and size_t
respectively. Should both be extern linkage.
(stream_recvfrom) Stream aware wrapper around recvfrom, in style
of stream_read_try.
* stream.c: (stream_recvfrom) new function, wrapper around recvfrom.
(stream_recvmsg, stream_write) ssize_t and size_t return values
|
|
Add wall-clock timing statistics to 'show thread cpu' output.
* thread.h: Define struct rusage_t to contain wall-clock time
and cpu time. Change GETRUSAGE macro to collect both pieces
of data. Make appropriate changes to struct cpu_thread_history
to track CPU time and real time. Change proto for
thread_consumed_time to return real and cpu time elapsed.
And declare a new global variable 'struct timeval recent_time'.
* thread.c (struct timeval recent_time): New global timestamp variable.
(timeval_adjust): If timeout is negative, set to 0 (not 10
microseconds). And remove upper bound of 1,000,000 seconds, since
this does not seem to make any sense (and it breaks
funcname_thread_add_timer_timeval).
(timeval_cmp): Should return long, not int.
(vty_out_cpu_thread_history): Show CPU time and real time.
(cpu_record_hash_print): Calculate totals for CPU and real time.
(cpu_record_print): Change 'show thread cpu' title to show CPU and
real time.
(thread_timer_remain_second): Put current time in global recent_time.
(funcname_thread_add_timer_timeval): Fix assert. Replace 2-case
switch assignment with a ternary expression. Use global recent_time
variable. Fix use of timeval_adjust (previously, the value was not
actually being adjusted).
(thread_cancel): Add missing "break" statement in case
THREAD_BACKGROUND.
(thread_timer_wait): Use global recent_time value instead of calling
gettimeofday. And there's no need to check for negative timeouts,
since timeval_subtract already sets these to zero.
(thread_timer_process): Timers are sorted, so bail out once we
encounter a timer that has not yet popped. And remove some
extraneous asserts.
(thread_fetch): Do not process foreground timers before calling
select. Instead, add them to the ready list just after the select.
Also, no need to maintain a count of the number of ready threads,
since we don't care how many there are, just whether there's
one at the head of the ready list (which is easily checked).
Stick current time in global variable recent_time to reduce
the number of calls to gettimeofday. Tighten logic for
calculating the select timeout.
(thread_consumed_time): Now returns real time and puts the elapsed
cpu time in an additional argument.
(thread_should_yield): Use real (wall-clock) time to decide whether
to yield.
(thread_call): Maintain CPU and real time statistics.
* vty.c (vty_command): For slow commands, show real and cpu time.
|
|
* workqueue.c (show_work_queues): Remove unused gettimeofday call.
|
|
* memory.h: memtypes is built source, default includes points to
top_builddir, so we should refer to lib/memtypes.h
|
|
* workqueue.h: (struct work_queue_item) change retry_count to ran,
its a count of number item has been run.
* workqueue.c: (show_work_queues) Fix formating of slightly
bugfix: fix SIGFPE if wq->runs is 0.
(work_queue_run) retry logic was slightly wrong.
cycles.best is 0 initialy, granularity is 1, so update best
if cycles >= granularity, not just >.
|
|
* buffer.c (buffer_write): Comment out call to buffer_flush_available.
This should speed up buffering at the expense of a possible increase
in latency in flushing the data if inside a long-running thread.
|
|
|
|
* thread.c: Kill unused TIMER_NO_SORT bits
|
|
* memory.c: Make the string field much wider
* memtypes.c: Correct the prefix list str/entry strings
|
|
* Makefile.am: Refer to source files via srcdir variable, fix
out-of-tree build breakage.
|
