more airports
Hi. This patch adds/fixes/extends airport data for some exUSSR airports I've been to. OK? Index: share/misc/airport === RCS file: /cvs/src/share/misc/airport,v retrieving revision 1.47 diff -u -r1.47 airport --- share/misc/airport 7 Jan 2015 16:08:50 - 1.47 +++ share/misc/airport 11 Jan 2015 08:18:53 - @@ -11,7 +11,9 @@ AAA:Anaa, Anaa, French Polynesia AAE:Rabah Bitat, Annaba, Algeria AAL:Aalborg, Denmark +AAQ:Vityazevo, Anapa, Russia AAR:Aarhus, Aarhus, Denmark +ABA:Abakan, Russia ABE:Lehigh Valley International, Pennsylvania, USA ABI:Abilene Regional, Abilene, Texas, USA ABJ:Port Bouet, Abidjan, Cote d'ivoire @@ -39,6 +41,7 @@ ADQ:Kodiak, Alaska, USA ADZ:Gustavo Rojas Pinilla International, San Andres Island, Colombia AEP:Jorge Newberry, Buenos Aires, Argentina +AER:Sochi, Russia AES:Vigra, Aalesund, Norway AET:Allakaket, Alaska, USA AEX:Alexandria International, Louisiana, USA @@ -553,7 +556,7 @@ FRM:Fairmont Municipal, Minnesota, USA FRO:Floro, Norway FRS:Flores, Guatemala -FRU:Bishkek, Kyrgyzstan +FRU:Manas, Bishkek, Kyrgyzstan FSC:Figari, Corsica, France FSD:Joe Foss Field, Sioux Falls, South Dakota, USA FSM:Fort Smith Municipal, Arkansas, USA @@ -701,6 +704,7 @@ HRO:Boone County, Harrison, Arkansas, USA HSI:Hastings, Nebraska, USA HSV:Huntsville International, Alabama, USA +HTA:Chita, Russia HTI:Hamilton Island, Queensland, Australia HUF:Terre Haute Hulman Regional, Indiana, USA HUI:Hue, Vietnam @@ -845,7 +849,8 @@ KEP:Nepalganj, Nepal KER:Kerman, Iran KGC:Kingscote, South Australia, Australia -KGD:Kaliningrad, Russia +KGD:Khrabrovo, Kaliningrad, Russia +KGP:Kogalym, Russia KHE:Kherson International, Kherson, Ukraine KHH:Kaohsiung, Taiwan KHI:Karachi, Pakistan @@ -856,7 +861,7 @@ KIT:Kithira, Greece KIV:Kishinev, Moldova KIX:Kansai International, Osaka, Japan -KJA:Krasnojarsk, Russia +KJA:Yemelyanovo, Krasnoyarsk, Russia KKN:Hoeyburtmoen, Kirkenes, Norway KLO:Kalibo, Philippines KLR:Kalmar, Sweden @@ -894,7 +899,7 @@ KTR:Tindal, Katherine, Northern Territory, Australia KTW:Pyrzowice, Katowice, Poland KUA:Padang Geroda, Kuantan, Malaysia -KUF:Samara, Russia +KUF:Kurumoch, Samara, Russia KUL:Subang Kuala Lumpur International, Malaysia KUM:Yakushima, Japan KUN:Kaunas, Lithuania @@ -907,6 +912,7 @@ KWI:Kuwait International KWG:Kryvyi Rih International, Dnipropetrovs'k, Ukraine KWL:Guilin, China +KXK:Khurba, Komsomolsk-on-Amur, Russia KZI:Kozani, Greece KZN:Kazan, Russia KZS:Kastelorizo, Greece @@ -1043,6 +1049,7 @@ MCP:Macapa, Amapa, Brazil MCT:Seeb, Muscat, Oman MCW:Mason City, Iowa, USA +MCX:Uytash, Makhachkala, Dagestan, Russia MCY:Maroochydore, Sunshine Coast, Queensland, Australia MDC:Samratulang, Manado, Indonesia MDE:La Playas, Medellin, Colombia @@ -1120,6 +1127,7 @@ MRK:Marco Island, Florida, USA MRS:Marseille, France MRU:Sir Seewoosagur Ramgoolam Airport, Mauritius +MRV:Mineralnye Vody, Russia MRY:Monterey, California, USA MSJ:Misawa, Japan MSL:Muscle Shoals, Muscle Shoals / Florence / Sheffield, Alabama, USA @@ -1166,6 +1174,7 @@ NCL:Newcastle International, England, United Kingdom NDJ:N'djamena, N Djamena, Chad NEC:Necochea, Buenos Aires, Argentina +NER:Chulman, Neryungri, Russia NEV:Nevis, Leeward Islands, Saint Kitts And Nevis NGO:Komaki, Nagoya, Japan NGS:Nagasaki, Japan @@ -1429,7 +1438,7 @@ ROR:Airai, Koror, Palau ROS:Fisherton, Rosario, Santa Fe, Argentina ROU:Rousse, Bulgaria -ROV:Rostov, Russia +ROV:Rostov-on-Don, Russia ROW:Industrial Air Center, Roswell, New Mexico, USA RPR:Raipur, India RRG:Rodrigues Island, Mauritius @@ -1575,6 +1584,7 @@ STR:Echterdingen, Stuttgart, Germany STS:Sonoma County, Santa Rosa, California, USA STT:Cyril E King, St Thomas Island, US Virgin Islands, USA +STW:Stavropol, Russia STX:St Croix Island, US Virgin Islands, USA SUB:Juanda, Surabaya, Indonesia SUE:Sturgeon Bay, Wisconsin, USA @@ -1587,7 +1597,7 @@ SVO:Sheremetyevo, Moscow, Russia SVQ:Sevilla, Spain SVU:Savusavu, Savusavu, Fiji -SVX:Ekaterinburg, Russia +SVX:Koltsovo, Ekaterinburg, Russia SVZ:San Antonio, Venezuela SWA:Shantou, China SWF:Stewart, Newburgh/Poughkeepsie, New York, USA @@ -1648,7 +1658,7 @@ TIV:Tivat, Montenegro TIZ:Tari, Papua New Guinea TJA:Tarija, Bolivia -TJM:Tyumen, Russia +TJM:Roshchino, Tyumen, Russia TKK:Truk, Caroline Islands, Micronesia TKQ:Kigoma, Tanzania TKS:Tokushima, Japan @@ -1683,6 +1693,7 @@ TRV:Trivandrum, India TRZ:Civil, Tiruchirapally, India TSA:Sung Shan, Taipei, Taiwan +TSE:Astana, Kazakhstan TSR:Timisoara, Romania TSS:East 34Th Street Heliport, New York, New York, USA TSV:Townsville, Queensland, Australia @@ -1716,6 +1727,7 @@ UAQ:San Juan, Argentina UBJ:Ube, Japan UCA:Utica/Oneida County, New York, USA +UCT:Ukhta, Russia UDJ:Uzhgorod International, Zakarpattia, Ukraine UDR:Udaipur, India UEL:Quelimane, Mozambique @@ -1787,7 +1799,7 @@ VTZ:Vishakhapatnam, India VUP:Valledupar,
Re: massage volume control task
On 3 Jan 2015, at 9:35 am, David Gwynne da...@gwynne.id.au wrote:
On 2 Jan 2015, at 9:55 pm, Mark Kettenis mark.kette...@xs4all.nl wrote:
Date: Fri, 2 Jan 2015 16:15:07 +1000
From: David Gwynne da...@gwynne.id.au
can someone test this?
it allocates storage for the volume change details rather than cast
arguments to a single global task.
adds some safety while there if audio0 is a hotplug device.
ok?
The problem with this approach is that things silently fail when
memory allocation fails. Exactly the problem that workq_add_task()
had. Perhaps the best thing is to store the event in the softc,
although then you'd probably need some locking to make sure the task
sees a consistent dir/out pair.
is silently fail a pun here?
you can lose events either way. if you get more than one keypress before the
task runs, you'll lose. if you are in a situation where you cant allocate
memory, you lose.
i think the code generally is bogus so im trying to do as little as possible
to it. the more i look at it the more i want to just remove it.
its been unreliable in all incarnations.
did anyone try this diff?
dlg
Index: audio.c
===
RCS file: /cvs/src/sys/dev/audio.c,v
retrieving revision 1.125
diff -u -p -r1.125 audio.c
--- audio.c 19 Dec 2014 22:44:58 - 1.125
+++ audio.c 2 Jan 2015 06:08:39 -
@@ -465,11 +465,6 @@ audioattach(struct device *parent, struc
}
DPRINTF((audio_attach: inputs ports=0x%x, output ports=0x%x\n,
sc-sc_inports.allports, sc-sc_outports.allports));
-
-#if NWSKBD 0
- task_set(sc-sc_mixer_task, wskbd_set_mixervolume_callback, NULL,
- NULL);
-#endif /* NWSKBD 0 */
}
int
@@ -3432,27 +3427,39 @@ filt_audiowrite(struct knote *kn, long h
}
#if NWSKBD 0
+struct wskbd_vol_change {
+ struct task t;
+ long dir;
+ long out;
+};
+
int
wskbd_set_mixervolume(long dir, long out)
{
struct audio_softc *sc;
+ struct wskbd_vol_change *ch;
if (audio_cd.cd_ndevs == 0 || (sc = audio_cd.cd_devs[0]) == NULL) {
DPRINTF((wskbd_set_mixervolume: audio_cd\n));
return (ENXIO);
}
- task_del(systq, sc-sc_mixer_task);
- task_set(sc-sc_mixer_task, wskbd_set_mixervolume_callback,
- (void *)dir, (void *)out);
- task_add(systq, sc-sc_mixer_task);
+ ch = malloc(sizeof(*ch), M_TEMP, M_NOWAIT);
+ if (ch == NULL)
+ return (ENOMEM);
+
+ task_set(ch-t, wskbd_set_mixervolume_callback, ch, NULL);
+ ch-dir = dir;
+ ch-out = out;
+ task_add(systq, ch-t);
return (0);
}
void
-wskbd_set_mixervolume_callback(void *arg1, void *arg2)
+wskbd_set_mixervolume_callback(void *xch, void *null)
{
+ struct wskbd_vol_change *ch = xch;
struct audio_softc *sc;
struct au_mixer_ports *ports;
mixer_devinfo_t mi;
@@ -3461,19 +3468,19 @@ wskbd_set_mixervolume_callback(void *arg
u_int gain;
int error;
- if (audio_cd.cd_ndevs == 0 || (sc = audio_cd.cd_devs[0]) == NULL) {
- DPRINTF((%s: audio_cd\n, __func__));
- return;
- }
+ dir = ch-dir;
+ out = ch-out;
+ free(ch, M_TEMP, sizeof(*ch));
- dir = (long)arg1;
- out = (long)arg2;
+ sc = (struct audio_softc *)device_lookup(audio_cd, 0);
+ if (sc == NULL)
+ return;
ports = out ? sc-sc_outports : sc-sc_inports;
if (ports-master == -1) {
DPRINTF((%s: master == -1\n, __func__));
- return;
+ goto done;
}
if (dir == 0) {
@@ -3482,7 +3489,7 @@ wskbd_set_mixervolume_callback(void *arg
error = au_get_mute(sc, ports, mute);
if (error != 0) {
DPRINTF((%s: au_get_mute: %d\n, __func__, error));
- return;
+ goto done;
}
mute = !mute;
@@ -3490,7 +3497,7 @@ wskbd_set_mixervolume_callback(void *arg
error = au_set_mute(sc, ports, mute);
if (error != 0) {
DPRINTF((%s: au_set_mute: %d\n, __func__, error));
- return;
+ goto done;
}
} else {
/* Raise or lower volume */
@@ -3499,7 +3506,7 @@ wskbd_set_mixervolume_callback(void *arg
error = sc-hw_if-query_devinfo(sc-hw_hdl, mi);
if (error != 0) {
DPRINTF((%s: query_devinfo: %d\n, __func__, error));
- return;
+ goto done;
}
au_get_gain(sc, ports, gain, balance);
@@ -3512,8 +3519,11 @@ wskbd_set_mixervolume_callback(void *arg
error = au_set_gain(sc, ports, gain, balance);
if (error != 0) {
DPRINTF((%s: au_set_gain: %d\n, __func__, error));
- return;
+ goto done;
}
}
+
+done:
+ device_unref(sc-dev);
}
Re: bgpd.conf macros on 5.5 and up
On Sat, Jan 10, 2015 at 02:21:40PM +, Stuart Henderson wrote:
moved from misc@ (http://marc.info/?l=openbsd-miscm=141898047318322w=2)
On 2014-12-19, Tony Sarendal t...@polarcap.org wrote:
From 5.5 and up it looks like bgpd macros are broken.
I suspect this is parse.y r1.268.
Also, the example from bgpd.conf man page fails on 5.4-5.6.
I haven't tested on 5.3 and lower.
On 5.6 snapshot:
tonsar@obc1$ uname -mrsv
OpenBSD 5.6 GENERIC.MP#701 amd64
tonsar@obc1$ cat bgpd.conf-2
good={ 192.168.0.0/16, 172.16.0.0/12, 10.0.0.0/8 }
bad={ 224.0.0.0/4 prefixlen = 4, 240.0.0.0/4 prefixlen = 4 }
ugly={ 127.0.0.1/8, 169.254.0.0/16 }
# global configuration
AS 65001
deny from any prefix { $good $bad $ugly }
tonsar@obc1$ bgpd -f bgpd.conf-2 -nv
good = { 192.168.0.0/16, 172.16.0.0/12, 10.0.0.0/8 }
bad = { 224.0.0.0/4 prefixlen = 4, 240.0.0.0/4 prefixlen = 4 }
ugly = { 127.0.0.1/8, 169.254.0.0/16 }
bgpd.conf-2:6: syntax error
tonsar@obc1$
It looks like nested braces no longer work.
{ { 1.1.1.1/30 } { 2.2.2.2/30 } } - fails
{ 1.1.1.1/30 2.2.2.2/30 } - works
With my poor yacc skills the best I can do for now is suggest a manpage
diff, though other parsers (e.g. pf's) do allow nesting in this situation.
Index: bgpd.conf.5
===
RCS file: /cvs/src/usr.sbin/bgpd/bgpd.conf.5,v
retrieving revision 1.132
diff -u -p -r1.132 bgpd.conf.5
--- bgpd.conf.5 10 Nov 2014 20:48:33 - 1.132
+++ bgpd.conf.5 10 Jan 2015 14:13:58 -
@@ -1210,9 +1210,9 @@ deny from any prefix { 192.168.0.0/16, 1
Multiple lists can also be specified, which is useful for
macro expansion:
.Bd -literal -offset indent
-good={ 192.168.0.0/16, 172.16.0.0/12, 10.0.0.0/8 }
-bad={ 224.0.0.0/4 prefixlen = 4, 240.0.0.0/4 prefixlen = 4 }
-ugly={ 127.0.0.1/8, 169.254.0.0/16 }
+good=192.168.0.0/16, 172.16.0.0/12, 10.0.0.0/8
+bad=224.0.0.0/4 prefixlen = 4, 240.0.0.0/4 prefixlen = 4
+ugly=127.0.0.1/8, 169.254.0.0/16
deny from any prefix { $good $bad $ugly }
.Ed
I think we should fix the parser and not the docu.
--
:wq Claudio
Re: Kernel does not compile with option LOCKDEBUG
On Sat, 10 Jan 2015 19:39:19 -0800
Philip Guenther guent...@gmail.com wrote:
On Mon, 5 Jan 2015, Helg wrote:
The man page for LOCK(9) says that if the kernel option LOCKDEBUG is
enabled, additional facilities are provided to assist in determining
deadlock occurrences.
I created a copy of /sys/arch/amd64/conf/GENERIC and added option
LOCKDEBUG. Executing config and then make clean make results in
warnings like:
implicit declaration of function '__mp_lock'
Adding includes for sys/mplock.h in the offending files resolves the
problem but this just doesn't seem right.
Does anyone have any suggestions?
Yeah, don't use it. It used to be a lot more when there were still traces
of the original simplelocks in the tree, but the code to do more with
LOCKDEBUG was ripped out as those no-ops were felt to be more misleading
than helpful.
Thanks. I figured as much when I started to look deeper into the
implementation of lockinit and lockmgr. These are now just legacy
wrappers for rwlock functions. Removing LOCKDEBUG makes sense to me.
It's dead, Jim, let's bury LOCKDEBUG.
oks?
Philip
Index: sys/arch/i386/i386/trap.c
===
RCS file: /cvs/src/sys/arch/i386/i386/trap.c,v
retrieving revision 1.119
diff -u -p -r1.119 trap.c
--- sys/arch/i386/i386/trap.c 2 Dec 2014 18:13:10 - 1.119
+++ sys/arch/i386/i386/trap.c 11 Jan 2015 03:33:48 -
@@ -381,11 +381,6 @@ trap(struct trapframe *frame)
case T_PAGEFLT: /* allow page faults in kernel mode */
if (p == 0 || p-p_addr == 0)
goto we_re_toast;
-#ifdef LOCKDEBUG
- /* If we page-fault while in scheduler, we're doomed. */
- if (__mp_lock_held(sched_lock))
- goto we_re_toast;
-#endif
pcb = p-p_addr-u_pcb;
#if 0
Index: sys/sys/sched.h
===
RCS file: /cvs/src/sys/sys/sched.h,v
retrieving revision 1.37
diff -u -p -r1.37 sched.h
--- sys/sys/sched.h 17 Oct 2014 01:51:39 - 1.37
+++ sys/sys/sched.h 11 Jan 2015 03:33:48 -
@@ -183,7 +183,7 @@ void remrunqueue(struct proc *);
yield();\
} while (0)
-#if defined(MULTIPROCESSOR) || defined(LOCKDEBUG)
+#if defined(MULTIPROCESSOR)
#include sys/lock.h
/*
@@ -215,7 +215,7 @@ do {
\
splx(s);\
} while (/* CONSTCOND */ 0)
-#else /* ! MULTIPROCESSOR || LOCKDEBUG */
+#else /* ! MULTIPROCESSOR */
#define SCHED_ASSERT_LOCKED() splassert(IPL_SCHED);
#define SCHED_ASSERT_UNLOCKED() /* nothing */
@@ -225,7 +225,7 @@ do {
\
#define SCHED_LOCK(s) s = splsched()
#define SCHED_UNLOCK(s) splx(s)
-#endif /* MULTIPROCESSOR || LOCKDEBUG */
+#endif /* MULTIPROCESSOR */
#endif /* _KERNEL */
#endif /* _SYS_SCHED_H_ */
Index: share/man/man9/lock.9
===
RCS file: /cvs/src/share/man/man9/lock.9,v
retrieving revision 1.22
diff -u -p -r1.22 lock.9
--- share/man/man9/lock.9 9 Jul 2014 14:16:10 - 1.22
+++ share/man/man9/lock.9 11 Jan 2015 03:33:48 -
@@ -57,10 +57,6 @@ single process.
It also allows upgrading a shared-access lock to an
exclusive-access lock, as well as downgrading an exclusive-access lock
to a shared-access lock.
-.Pp
-If the kernel option LOCKDEBUG is enabled, additional facilities
-are provided to record additional lock information.
-These facilities are provided to assist in determining deadlock occurrences.
.Sh FUNCTIONS
The functions which operate on locks are:
.Bl -tag -width Ds
--
Helg xx...@msn.com
Re: 5.6, IPv6: is autoconf set by default?
On Sat, Jan 10, 2015 at 06:09:21PM +, Stuart Henderson wrote: On 2015/01/10 14:54, Florian Obser wrote: I don't think this is entirely correct. eui64 enables IPv6 on an interface by setting a link local address. For lo0 it also sets ::1. I'm unsure what eui64 was supposed to do when IPv6 was on by default and there was no way to disable it. Historically eui64 was used like this: ifconfig vlan4 inet6 2001:db8:: eui64 and it would fill in the bottom 64 bits for you, i.e. 2001:db8::f2de:f1ff:fea3:bc17/64 And this was the use-case I tried to describe (insufficiently) in my ifconfig.8 patch. That is still the normal use for eui64, prior to it now doing double-duty as re-enable ipv6 on this interface and set things up as they used to be. btw. I'm still trying to figure out what setia6eui64() in ifconfig.c is trying to achieve, the code after addaf() looks peculiar. If it's just error checking it looks complicated. Related to the above. And oh, that uses the interface index terminology in a printf...
Re: Kernel does not compile with option LOCKDEBUG
Philip Guenther guenther at gmail.com writes: It's dead, Jim, let's bury LOCKDEBUG. There is an define AZALIA_LOG_MP and accompanying code in sys/dev/pci/azalia.c which looks like a debug left-over. azalia(4) is considered MP-safe for over a year from now.
Re: env fix
On Sun, Jan 11, 2015 at 9:36 AM, Ted Unangst t...@tedunangst.com wrote:
env won't run a command with an = in its name. This is documented as a
bug, but it's easily fixed in a backwards compatible way.
...
+ if (strcmp(*argv, --) == 0)
+ argv++;
No, env foo=bar -- baz=qux really should try to execute --: the
enviroment variables aren't an option string being processed with
getopt().
What we *can* do is enforce a bit more this requirement on environment names:
In the shell command language, a word consisting solely of
underscores, digits, and alphabetics
from the portable character set. The first character of a name is
not a digit.
Slash is not legal in shell variable names, so if we treat an argument
containing a slash before the = as the command instead of a variable
assignment, users have an escape hatch a do something like:
env foo=bar ./my=cool=program
or
env foo=bar `which another=test`
etc.
Something like this on the code side?
PHilip
--- env.c 8 Mar 2014 00:09:20 - 1.15
+++ env.c 11 Jan 2015 19:20:36 -
@@ -63,6 +63,8 @@ main(int argc, char *argv[])
argv += optind;
for (; *argv (p = strchr(*argv, '=')); ++argv) {
+ if (memchr(*argv, '/', p - *argv) != NULL)
+ break;
*p++ = '\0';
if (setenv(*argv, p, 1) == -1) {
/* reuse 126, it matches the problem most */
Re: env fix
On Sun, Jan 11, 2015 at 11:26, Philip Guenther wrote:
On Sun, Jan 11, 2015 at 9:36 AM, Ted Unangst t...@tedunangst.com wrote:
env won't run a command with an = in its name. This is documented as a
bug, but it's easily fixed in a backwards compatible way.
...
+ if (strcmp(*argv, --) == 0)
+ argv++;
No, env foo=bar -- baz=qux really should try to execute --: the
enviroment variables aren't an option string being processed with
getopt().
What we *can* do is enforce a bit more this requirement on environment names:
In the shell command language, a word consisting solely of
underscores, digits, and alphabetics
from the portable character set. The first character of a name is
not a digit.
Slash is not legal in shell variable names, so if we treat an argument
containing a slash before the = as the command instead of a variable
assignment, users have an escape hatch a do something like:
env foo=bar ./my=cool=program
or
env foo=bar `which another=test`
etc.
Something like this on the code side?
Even more awesome.
PHilip
--- env.c 8 Mar 2014 00:09:20 - 1.15
+++ env.c 11 Jan 2015 19:20:36 -
@@ -63,6 +63,8 @@ main(int argc, char *argv[])
argv += optind;
for (; *argv (p = strchr(*argv, '=')); ++argv) {
+ if (memchr(*argv, '/', p - *argv) != NULL)
+ break;
*p++ = '\0';
if (setenv(*argv, p, 1) == -1) {
/* reuse 126, it matches the problem most */
env fix
env won't run a command with an = in its name. This is documented as a
bug, but it's easily fixed in a backwards compatible way.
Index: env.1
===
RCS file: /cvs/src/usr.bin/env/env.1,v
retrieving revision 1.19
diff -u -p -r1.19 env.1
--- env.1 8 Mar 2014 01:42:17 - 1.19
+++ env.1 11 Jan 2015 17:32:32 -
@@ -119,9 +119,3 @@ specification.
The historic
.Fl
option has been deprecated but is still supported in this implementation.
-.Sh BUGS
-.Nm
-doesn't handle commands with equal
-.Pq Sq =
-signs in their
-names, for obvious reasons.
Index: env.c
===
RCS file: /cvs/src/usr.bin/env/env.c,v
retrieving revision 1.15
diff -u -p -r1.15 env.c
--- env.c 8 Mar 2014 00:09:20 - 1.15
+++ env.c 11 Jan 2015 17:32:07 -
@@ -69,6 +69,8 @@ main(int argc, char *argv[])
err(126, setenv);
}
}
+ if (strcmp(*argv, --) == 0)
+ argv++;
if (*argv) {
/*
Re: env fix
On Sun, Jan 11, 2015 at 13:53, Philip Guenther wrote: On Sun, 11 Jan 2015, Ted Unangst wrote: Even more awesome. How about enforcing the full rule? And now you've gone full awesome.
Re: env fix
On Sun, 11 Jan 2015, Ted Unangst wrote: Even more awesome. How about enforcing the full rule? IIUC the first diff removed '/' from the characters allowed in an environment variable, so that one can run env(1) and a program whose name contains '='... I've never seen such a program name. I think this is not env(1)'s job to tell which letters can be put in a environment variable name; other env(1) implementations don't seem to care at all. btw, in ksh FOO=bar BAZ=quuz ./your=program works around the no '=' in program name rule; I seriously doubt this is a concern though. I don't see how this improves env. Now it has a difference from other systems. How do you use it portably? You can't. The portable behaviour is that it fails like now.
USB abort tasks deadlock
If you're detaching a USB device while another thread is still holding
a reference on it (ifconfig/dhclient for example) you might end up in
a deadlock situation and you'll have to reboot* to use your USB ports.
This deadlock is triggered if the thread detaching your device sleeps
in the *_detach() function waiting for all the previously submitted
USB transfer to finish. Sleeping in a *_detach() function means that
the USB device has been flagged as `dying'. But our custom USB task
code prevent any task to be scheduled as soon as the device is dying.
Sadly, if your device has been detached, there's a lot of chances that
its transfers time out. Since a timed out transfer needs a task to be
properly completed but our stack does not allow tasks to be scheduled
anymore - deadlock.
If you've ever found yourself with a non working USB, you can use ps(1)
to see if one of your threads is waiting on usbsync, the symptom of
this deadlock.
# ps axk -Owchan |grep usbsync
Diff below changes the USB task code to allow abort tasks to be scheduled
and executed even if the device is dying. This should be safe because
aborting the pipe will remove the task if it hasn't been executed.
Comments, ok?
* If you can break into DDB you don't necessarily need to reboot when
such deadlock occurs. First get the xfer address (wait channel) for the
stuck thread:
# ps axk -Onwchan |grep usbsync
Then break into DDB and finish the transfer manually:
ddb call usb_transfer_comple(0xnwchan_addr)
Index: usb.c
===
RCS file: /cvs/src/sys/dev/usb/usb.c,v
retrieving revision 1.103
diff -u -p -r1.103 usb.c
--- usb.c 18 Dec 2014 10:44:17 - 1.103
+++ usb.c 11 Jan 2015 22:41:14 -
@@ -298,8 +298,13 @@ usb_add_task(struct usbd_device *dev, st
{
int s;
- /* Don't add task if the device's root hub is dying. */
- if (usbd_is_dying(dev))
+ /*
+* If the thread detaching ``dev'' is sleeping, waiting
+* for all submitted transfers to finish, we must be able
+* to enqueue abort tasks. Otherwise timeouts can't give
+* back submitted transfers to the stack.
+*/
+ if (usbd_is_dying(dev) (task-type != USB_TASK_TYPE_ABORT))
return;
DPRINTFN(2,(%s: task=%p state=%d type=%d\n, __func__, task,
@@ -455,12 +460,9 @@ usb_abort_task_thread(void *arg)
*/
task-state |= USB_TASK_STATE_RUN;
task-state = ~USB_TASK_STATE_ONQ;
- /* Don't actually execute the task if dying. */
- if (!usbd_is_dying(task-dev)) {
- splx(s);
- task-fun(task-arg);
- s = splusb();
- }
+ splx(s);
+ task-fun(task-arg);
+ s = splusb();
task-state = ~USB_TASK_STATE_RUN;
if (task-state == USB_TASK_STATE_NONE)
wakeup(task);
Re: env fix
On Sun, 11 Jan 2015, Ted Unangst wrote:
Even more awesome.
How about enforcing the full rule?
Index: env.c
===
RCS file: /cvs/src/usr.bin/env/env.c,v
retrieving revision 1.15
diff -u -p -r1.15 env.c
--- env.c 8 Mar 2014 00:09:20 - 1.15
+++ env.c 11 Jan 2015 21:52:32 -
@@ -29,6 +29,7 @@
* SUCH DAMAGE.
*/
+#include ctype.h
#include err.h
#include errno.h
#include locale.h
@@ -39,12 +40,22 @@
__dead void usage(void);
+/*
+ * POSIX XBD definition of 'name':
+ * In the shell command language, a word consisting solely of
+ * underscores, digits, and alphabetics from the portable character
+ * set. The first character of a name is not a digit.
+ */
+static const char varchar[] =
+ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789_;
+
int
main(int argc, char *argv[])
{
extern char **environ;
extern int optind;
char **ep, *p;
+ size_t n;
int ch;
setlocale(LC_ALL, );
@@ -62,9 +73,14 @@ main(int argc, char *argv[])
argc -= optind;
argv += optind;
- for (; *argv (p = strchr(*argv, '=')); ++argv) {
- *p++ = '\0';
- if (setenv(*argv, p, 1) == -1) {
+ for (; p = *argv; ++argv) {
+ /* is the argument an assignment to a legal name? */
+ if (isdigit((unsigned char)*p) ||
+ (n = strspn(p, varchar)) == 0 ||
+ p[n] != '=')
+ break;
+ p[n] = '\0';
+ if (setenv(p, p + n + 1, 1) == -1) {
/* reuse 126, it matches the problem most */
err(126, setenv);
}
Index: env.1
===
RCS file: /cvs/src/usr.bin/env/env.1,v
retrieving revision 1.19
diff -u -p -r1.19 env.1
--- env.1 8 Mar 2014 01:42:17 - 1.19
+++ env.1 11 Jan 2015 21:52:32 -
@@ -68,6 +68,16 @@ Causes
to completely ignore the environment it inherits.
.El
.Pp
+To be treated as an assignment, the
+.Ar name
+before the equals-sign must consist only of underscores, digits, and letters,
+and must not start with a digit.
+The first argument to
+.Nm
+that doesn't meet that criteria will be considered the
+.Ar utility
+to execute.
+.Pp
If no
.Ar utility
is specified,
@@ -119,9 +129,3 @@ specification.
The historic
.Fl
option has been deprecated but is still supported in this implementation.
-.Sh BUGS
-.Nm
-doesn't handle commands with equal
-.Pq Sq =
-signs in their
-names, for obvious reasons.
Re: env fix
Philip Guenther guent...@gmail.com writes: On Sun, 11 Jan 2015, Ted Unangst wrote: Even more awesome. How about enforcing the full rule? IIUC the first diff removed '/' from the characters allowed in an environment variable, so that one can run env(1) and a program whose name contains '='... I've never seen such a program name. I think this is not env(1)'s job to tell which letters can be put in a environment variable name; other env(1) implementations don't seem to care at all. btw, in ksh FOO=bar BAZ=quuz ./your=program works around the no '=' in program name rule; I seriously doubt this is a concern though. -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
