Menu
▾
▴
projectudi-code — Programmer discussion of reference implementation
You can subscribe to this list here.
| 2001 |
Jan
|
Feb
(12) |
Mar
(24) |
Apr
(12) |
May
(13) |
Jun
(67) |
Jul
(9) |
Aug
(11) |
Sep
(13) |
Oct
(7) |
Nov
|
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2002 |
Jan
(56) |
Feb
(58) |
Mar
(9) |
Apr
(16) |
May
(2) |
Jun
(2) |
Jul
|
Aug
(4) |
Sep
(4) |
Oct
|
Nov
|
Dec
|
| 2003 |
Jan
(1) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(2) |
Dec
|
| 2004 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(2) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(1) |
Nov
|
Dec
|
| 2005 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
|
Jun
(1) |
Jul
(11) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Steven B. <sb...@st...> - 2005-07-15 12:26:57
|
Did you try doing this:
cd /usr/src/udiref/env/linux
udisetup -v
I think I recall a problem with specifying the absolute path to the
.udi package file.
--Steve
On Fri, Jul 15, 2005 at 05:50:17AM -0000, prasad roy r wrote:
> hi burk,
> Udisetup in verbos mod give the following information.In tem directory /tmp/udiszB3Ny8
> folder is getting created but why udisetup is looking for /tmp/upropBdpLaz! folder in tem
> directory.
>
> udisetup -v /usr/src/udiref/env/linux/udi.udi
> udisetup: INFO: Package file : udi_brdg.udi
> ++cp udi_brdg.udi /tmp/udiszB3Ny8/udi_brdg/udi_brdg.udi
> ++cd /tmp/udiszB3Ny8/udi_brdg
> ++tar xf udi_brdg.udi
> udisetup: INFO: ..Driver : udi_brdg
> udisetup: INFO: ....Alternative : 3
> udisetup: INFO: ......Component : udi_bridge
> udisetup: ERROR: Cannot open /tmp/upropBdpLaz!
>
> Exiting.
>
> regards
> prasad
> ?
>
>
> On Wed, 13 Jul 2005 Burkhard Daniel wrote :
> >Ok, it looks like the kernel you're using has undergone some significant changes since we last touched the udi ref env. I'm not sure there is much I can do to help you right now, because we're not actively maintaining the ref env at this time.
> >
> >Try using an older kernel from the 2.4 series :-)
> >
> >As to your problem with the /tmp file system, you might want to check which symbols from env_test it doesn't like. But that may not be the cause for your problem at all. If it isn't, your best bet is to step through udisetup and see where exactly it fails. Once you've found that, either the solution is obvious, or you can send your debugging results here and we can see if we can help you.
> >
> >Burk.
> >
> >prasad roy r wrote:
> >>hi burk,
> >> Sorry if i mailed it earlier.In file udi_osdep.c (/udiref/env/linux) i have done the following modification.
> >>spinlock_t sigmask_lock = SPIN_LOCK_UNLOCKED; /* added by prasad */
> >>
> >>void EnterUDIContext() {
> >> long flags;
> >> /* Modify signal mask so we don"t freeze up if they send a signal. */
> >> spin_lock_irqsave(sigmask_lock, flags);/*modified by prasad */
> >> siginitsetinv(?t->blocked, 0); /* disable all signals */
> >>/*sigmask(SIGINT) | sigmask(SIGKILL) | sigmask(SIGSTOP));*/
> >> recalc_sigpending();/*Modified by prasad */
> >> spin_unlock_irqrestore(sigmask_lock, flags);/* MODIFIED BY PRASAD */
> >> _OSDEP_SEM_P(&_udi_MA_init_complete);
> >>}
> >>
> >>Earlier i was getting "sigmask_lock" was not a member of "current".So i modified the
> >>code.I have a write permission in /tmp.
> >>
> >>regards
> >>prasad
> >>
> >>
> >>On Wed, 13 Jul 2005 Burkhard Daniel wrote :
> >> >prasad roy r wrote:
> >> >>Hi all,
> >> >> I am using redhat 9 enterprise to build UDI env with version udiref-ss-2001-09-30.tar.gz.With little modification i was able to build
> >> >
> >> >What kind of modifications?
> >> >
> >> >>the UDI.While i was trying to install the UDI env i am getting the following erros.
> >> >>
> >> >>289 blocks
> >> >>depmod: *** Unresolved symbols in /lib/modules/2.4.21-4.EL/misc/env_test.o
> >> >>udisetup: ERROR: Cannot open /tmp/upropfN8uMY!
> >> >
> >> >Can you check that you have write permission in /tmp?
> >> >
> >> >Burk.
> >> >-- Burkhard Daniel
> >> >Software Technologies Group, Inc.
> >> >bu...@st... * http://www.stg.com
> >> >fon: +49-179-5319489 fax: +49-30-55153591
> >>
> >>
> >>
> >><http://clients.rediff.com/signature/track_sig.asp>
> >
> >
> >-- Burkhard Daniel
> >Software Technologies Group, Inc.
> >bu...@st... * http://www.stg.com
> >fon: +49-179-5319489 fax: +49-30-55153591
|
|
From: prasad r. r <rpr...@re...> - 2005-07-15 05:49:02
|
hi burk,=0AUdisetup in verbos mod give the following information.In tem dir=
ectory /tmp/udiszB3Ny8=0Afolder is getting created but why udisetup is look=
ing for /tmp/upropBdpLaz! folder in tem=0Adirectory.=0A =0A udisetup -v /us=
r/src/udiref/env/linux/udi.udi=0A udisetup: INFO: Package file : udi_brd=
g.udi=0A ++cp udi_brdg.udi /tmp/udiszB3Ny8/udi_brdg/udi_brdg.udi =0A =
++cd /tmp/udiszB3Ny8/udi_brdg=0A ++tar xf udi_brdg.udi =0A udisetup: =
INFO: ..Driver : udi_brdg=0A udisetup: INFO: ....Alternative : 3=0A =
udisetup: INFO: ......Component : udi_bridge =0A udisetup: ERROR: Cannot o=
pen /tmp/upropBdpLaz!=0A =0A Exiting.=0A =0A regards=0A prasad=0A=A0=0A=0A=
=0AOn Wed, 13 Jul 2005 Burkhard Daniel wrote :=0A>Ok, it looks like the ker=
nel you're using has undergone some significant changes since we last touch=
ed the udi ref env. I'm not sure there is much I can do to help you right n=
ow, because we're not actively maintaining the ref env at this time.=0A>=0A=
>Try using an older kernel from the 2.4 series :-)=0A>=0A>As to your proble=
m with the /tmp file system, you might want to check which symbols from env=
_test it doesn't like. But that may not be the cause for your problem at al=
l. If it isn't, your best bet is to step through udisetup and see where exa=
ctly it fails. Once you've found that, either the solution is obvious, or y=
ou can send your debugging results here and we can see if we can help you.=
=0A>=0A>Burk.=0A>=0A>prasad roy r wrote:=0A>>hi burk,=0A>> Sorry if i mail=
ed it earlier.In file udi_osdep.c (/udiref/env/linux) i have done the follo=
wing modification.=0A>>spinlock_t sigmask_lock =3D SPIN_LOCK_UNLOCKED; /* a=
dded by prasad */=0A>>=0A>>void EnterUDIContext() {=0A>> long flags;=0A=
>> /* Modify signal mask so we don"t freeze up if they send a signal. *=
/=0A>> spin_lock_irqsave(sigmask_lock, flags);/*modified by prasad */=
=0A>> siginitsetinv(=A4t->blocked, 0); /* disable all signals */=0A>>/*=
sigmask(SIGINT) | sigmask(SIGKILL) | sigmask(SIGSTOP));*/=0A>> recalc_s=
igpending();/*Modified by prasad */=0A>> spin_unlock_irqrestore(sigmask=
_lock, flags);/* MODIFIED BY PRASAD */=0A>> _OSDEP_SEM_P(&_udi_MA_init_=
complete);=0A>>}=0A>>=0A>>Earlier i was getting "sigmask_lock" was not a me=
mber of "current".So i modified the=0A>>code.I have a write permission in /=
tmp.=0A>>=0A>>regards=0A>>prasad=0A>> =0A>>=0A>>On Wed, 13 Jul 2005 Burkha=
rd Daniel wrote :=0A>> >prasad roy r wrote:=0A>> >>Hi all,=0A>> >> I am=
using redhat 9 enterprise to build UDI env with version udiref-ss-2001-09-=
30.tar.gz.With little modification i was able to build=0A>> >=0A>> >What =
kind of modifications?=0A>> >=0A>> >>the UDI.While i was trying to instal=
l the UDI env i am getting the following erros.=0A>> >>=0A>> >>289 blocks=
=0A>> >>depmod: *** Unresolved symbols in /lib/modules/2.4.21-4.EL/misc/en=
v_test.o=0A>> >>udisetup: ERROR: Cannot open /tmp/upropfN8uMY!=0A>> >=0A>=
> >Can you check that you have write permission in /tmp?=0A>> >=0A>> >Bu=
rk.=0A>> >-- Burkhard Daniel=0A>> >Software Technologies Group, Inc.=0A>>=
>bu...@st... * http://www.stg.com=0A>> >fon: +49-179-5319489 fax: +49-3=
0-55153591=0A>>=0A>>=0A>>=0A>><http://clients.rediff.com/signature/track_si=
g.asp>=0A>=0A>=0A>-- Burkhard Daniel=0A>Software Technologies Group, Inc.=
=0A>bu...@st... * http://www.stg.com=0A>fon: +49-179-5319489 fax: +49-30-5=
5153591=0A |
|
From: prasad r. r <rpr...@re...> - 2005-07-14 08:18:32
|
Hi steve,=0A=0AUdisetup in verbos mod give the following information.In tem= directory /tmp/udiszB3Ny8 folder is getting created but why udisetup is lo= oking for /tmp/upropBdpLaz! folder in tem directory.=0A=0Audisetup -v /usr/= src/udiref/env/linux/udi.udi=0Audisetup: INFO: Package file : udi_brdg.u= di=0A ++cp udi_brdg.udi /tmp/udiszB3Ny8/udi_brdg/udi_brdg.udi =0A ++c= d /tmp/udiszB3Ny8/udi_brdg=0A ++tar xf udi_brdg.udi =0Audisetup: INFO: .= .Driver : udi_brdg=0Audisetup: INFO: ....Alternative : 3=0Audisetup:= INFO: ......Component : udi_bridge =0Audisetup: ERROR: Cannot open /tmp/u= propBdpLaz!=0A=0AExiting.=0A=0Aregards=0Aprasad=0A=0A=0A=0A=0AOn Wed, 13 Ju= l 2005 Steven Bytnar wrote :=0A>On Wed, Jul 13, 2005 at 02:08:03PM +0200, B= urkhard Daniel wrote:=0A> > As to your problem with the /tmp file system, y= ou might want to check which=0A> > symbols from env_test it doesn't like. B= ut that may not be the cause for=0A> > your problem at all. If it isn't, yo= ur best bet is to step through udisetup=0A> > and see where exactly it fail= s. Once you've found that, either the solution=0A> > is obvious, or you can= send your debugging results here and we can see if=0A> > we can help you.= =0A>=0A>Before stepping through the code, you could first isolate where the= =0A>problem is at by executing udisetup with the verbose option:=0A>udisetu= p -v=0A>=0A>--Steve=0A>=0A |
|
From: Steven B. <sb...@st...> - 2005-07-13 15:36:39
|
On Wed, Jul 13, 2005 at 02:08:03PM +0200, Burkhard Daniel wrote: > As to your problem with the /tmp file system, you might want to check which > symbols from env_test it doesn't like. But that may not be the cause for > your problem at all. If it isn't, your best bet is to step through udisetup > and see where exactly it fails. Once you've found that, either the solution > is obvious, or you can send your debugging results here and we can see if > we can help you. Before stepping through the code, you could first isolate where the problem is at by executing udisetup with the verbose option: udisetup -v --Steve |
|
From: Burkhard D. <bu...@st...> - 2005-07-13 12:08:42
|
Ok, it looks like the kernel you're using has undergone some significant
changes since we last touched the udi ref env. I'm not sure there is much I can
do to help you right now, because we're not actively maintaining the ref env at
this time.
Try using an older kernel from the 2.4 series :-)
As to your problem with the /tmp file system, you might want to check which
symbols from env_test it doesn't like. But that may not be the cause for your
problem at all. If it isn't, your best bet is to step through udisetup and see
where exactly it fails. Once you've found that, either the solution is obvious,
or you can send your debugging results here and we can see if we can help you.
Burk.
prasad roy r wrote:
> hi burk,
>
> Sorry if i mailed it earlier.In file udi_osdep.c (/udiref/env/linux) i
> have done the following modification.
> spinlock_t sigmask_lock = SPIN_LOCK_UNLOCKED; /* added by prasad */
>
> void EnterUDIContext() {
> long flags;
> /* Modify signal mask so we don"t freeze up if they send a signal. */
> spin_lock_irqsave(sigmask_lock, flags);/*modified by prasad */
> siginitsetinv(¤t->blocked, 0); /* disable all signals */
> /*sigmask(SIGINT) | sigmask(SIGKILL) | sigmask(SIGSTOP));*/
> recalc_sigpending();/*Modified by prasad */
> spin_unlock_irqrestore(sigmask_lock, flags);/* MODIFIED BY PRASAD */
> _OSDEP_SEM_P(&_udi_MA_init_complete);
> }
>
> Earlier i was getting "sigmask_lock" was not a member of "current".So i
> modified the
> code.I have a write permission in /tmp.
>
> regards
> prasad
>
>
>
> On Wed, 13 Jul 2005 Burkhard Daniel wrote :
> >prasad roy r wrote:
> >>Hi all,
> >> I am using redhat 9 enterprise to build UDI env with version
> udiref-ss-2001-09-30.tar.gz.With little modification i was able to build
> >
> >What kind of modifications?
> >
> >>the UDI.While i was trying to install the UDI env i am getting the
> following erros.
> >>
> >>289 blocks
> >>depmod: *** Unresolved symbols in
> /lib/modules/2.4.21-4.EL/misc/env_test.o
> >>udisetup: ERROR: Cannot open /tmp/upropfN8uMY!
> >
> >Can you check that you have write permission in /tmp?
> >
> >Burk.
> >-- Burkhard Daniel
> >Software Technologies Group, Inc.
> >bu...@st... * http://www.stg.com
> >fon: +49-179-5319489 fax: +49-30-55153591
>
>
>
> <http://clients.rediff.com/signature/track_sig.asp>
--
Burkhard Daniel
Software Technologies Group, Inc.
bu...@st... * http://www.stg.com
fon: +49-179-5319489 fax: +49-30-55153591
|
|
From: prasad r. r <rpr...@re...> - 2005-07-13 11:29:22
|
hi burk,=0A =0A Sorry if i mailed it earlier.In file udi_osdep.c (/udiref=
/env/linux) i have done the following modification.=0A spinlock_t sigmask_l=
ock =3D SPIN_LOCK_UNLOCKED; /* added by prasad */=0A =0A void EnterUDIConte=
xt() {=0A long flags;=0A /* Modify signal mask so we don"t freeze u=
p if they send a signal. */=0A spin_lock_irqsave(sigmask_lock, flags);/=
*modified by prasad */=0A siginitsetinv(¤t->blocked, 0); /* disab=
le all signals */=0A /*sigmask(SIGINT) | sigmask(SIGKILL) | sigmask(SIGSTOP=
));*/=0A recalc_sigpending();/*Modified by prasad */=0A spin_unlock=
_irqrestore(sigmask_lock, flags);/* MODIFIED BY PRASAD */=0A _OSDEP_SEM=
_P(&_udi_MA_init_complete);=0A }=0A =0A Earlier i was getting "sigmask_lock=
" was not a member of "current".So i modified the=0Acode.I have a write per=
mission in /tmp.=0A =0A regards=0A prasad=0A =A0=0A=0A=0AOn Wed, 13 Jul 200=
5 Burkhard Daniel wrote :=0A>prasad roy r wrote:=0A>>Hi all,=0A>> I am usi=
ng redhat 9 enterprise to build UDI env with version udiref-ss-2001-09-30.t=
ar.gz.With little modification i was able to build=0A>=0A>What kind of modi=
fications?=0A>=0A>>the UDI.While i was trying to install the UDI env i am g=
etting the following erros.=0A>>=0A>>289 blocks=0A>>depmod: *** Unresolved =
symbols in /lib/modules/2.4.21-4.EL/misc/env_test.o=0A>>udisetup: ERROR: Ca=
nnot open /tmp/upropfN8uMY!=0A>=0A>Can you check that you have write permis=
sion in /tmp?=0A>=0A>Burk.=0A>-- Burkhard Daniel=0A>Software Technologies G=
roup, Inc.=0A>bu...@st... * http://www.stg.com=0A>fon: +49-179-5319489 fax=
: +49-30-55153591=0A |
|
From: prasad r. r <rpr...@re...> - 2005-07-13 08:42:31
|
hi Daniel,=0A =0AIn file udi_osdep.c (/udiref/env/linux) i have done the f=
ollowing modification.=0Aspinlock_t sigmask_lock =3D SPIN_LOCK_UNLOCKED; /*=
added by prasad */=0A=0Avoid EnterUDIContext() {=0A long flags;=0A /=
* Modify signal mask so we don't freeze up if they send a signal. */=0A =
spin_lock_irqsave(sigmask_lock, flags);/*modified by prasad */=0A sigini=
tsetinv(¤t->blocked, 0); /* disable all signals */=0A/*sigmask(SIGINT=
) | sigmask(SIGKILL) | sigmask(SIGSTOP));*/=0A recalc_sigpending();/*Mod=
ified by prasad */=0A spin_unlock_irqrestore(sigmask_lock, flags);/* MOD=
IFIED BY PRASAD */=0A _OSDEP_SEM_P(&_udi_MA_init_complete);=0A}=0A=0AEar=
lier i was getting "sigmask_lock" was not a member of "cutternt".So i modif=
ied the code.I have a write permission in /tmp.=0A=0Aregards=0Aprasad=0A=0A=
On Wed, 13 Jul 2005 Burkhard Daniel wrote :=0A>prasad roy r wrote:=0A>>Hi a=
ll,=0A>> I am using redhat 9 enterprise to build UDI env with version udir=
ef-ss-2001-09-30.tar.gz.With little modification i was able to build=0A>=0A=
>What kind of modifications?=0A>=0A>>the UDI.While i was trying to install =
the UDI env i am getting the following erros.=0A>>=0A>>289 blocks=0A>>depmo=
d: *** Unresolved symbols in /lib/modules/2.4.21-4.EL/misc/env_test.o=0A>>u=
disetup: ERROR: Cannot open /tmp/upropfN8uMY!=0A>=0A>Can you check that you=
have write permission in /tmp?=0A>=0A>Burk.=0A>-- Burkhard Daniel=0A>Softw=
are Technologies Group, Inc.=0A>bu...@st... * http://www.stg.com=0A>fon: +=
49-179-5319489 fax: +49-30-55153591=0A |
|
From: Burkhard D. <bu...@st...> - 2005-07-13 06:50:51
|
prasad roy r wrote: > Hi all, > > I am using redhat 9 enterprise to build UDI env with version > udiref-ss-2001-09-30.tar.gz.With little modification i was able to build What kind of modifications? > the UDI.While i was trying to install the UDI env i am getting the > following erros. > > 289 blocks > depmod: *** Unresolved symbols in /lib/modules/2.4.21-4.EL/misc/env_test.o > udisetup: ERROR: Cannot open /tmp/upropfN8uMY! Can you check that you have write permission in /tmp? Burk. -- Burkhard Daniel Software Technologies Group, Inc. bu...@st... * http://www.stg.com fon: +49-179-5319489 fax: +49-30-55153591 |
|
From: prasad r. r <rpr...@re...> - 2005-07-13 05:49:32
|
Hi all,=0A =0AI am using redhat 9 enterprise to build UDI env with version= udiref-ss-2001-09-30.tar.gz.With little modification i was able to build t= he UDI.While i was trying to install the UDI env i am getting the following= erros.=0A=0A289 blocks=0Adepmod: *** Unresolved symbols in /lib/modules/2.= 4.21-4.EL/misc/env_test.o=0Audisetup: ERROR: Cannot open /tmp/upropfN8uMY!= =0A=0AExiting.=0Amake[2]: [udisetup-mapper] Error 1 (ignored)=0Audisetup: E= RROR: Cannot open /tmp/upropyZMmo8!=0A=0AExiting.=0Amake[2]: [udisetup] Err= or 1 (ignored)=0Audisetup: ERROR: Cannot open /tmp/upropCMg0ql!=0A=0AExitin= g.=0Amake[2]: [udisetup] Error 1 (ignored)=0Audisetup: ERROR: Cannot open /= tmp/upropGvmoKv!=0A=0AExiting.=0Amake[2]: [udisetup] Error 1 (ignored)=0A/u= sr/src/udiref/tools/bin/udisetup: line 3: 10000 Segmentation fault sh = -c "$LOC/udisetup-`sh $LOC/../../env/common/Make_conf -E` $*"=0Amake[2]: [u= disetup] Error 139 (ignored)=0A=0AKindly help me to solve this problem.=0A= =0Aregards=0Aprasad |
|
From: Eutropio S. <Ste...@gc...> - 2005-07-07 10:56:53
|
Hello, destination, why our course is what it is?plainly written on his enormous = yellowish countenance. At his side,it considered the self-contained = buccaneer.their absence, I think. They are wealthy men that can = ransomcomplete possession of the town, glutting themselves hideously = upongiving me your parole under stress, and afterwards playing me = false.overwrought by the suspense, mademoiselle flung herself upon = herhave it! - the Santiago sprang her mainmast; and so I was glad = tosince a slave had made an attack upon him and all but strangled = himLevasseur and Blood, and signed by them and - as was usual - by = theThe Deputy-Governor was flung into panic. He lost his temper, andShe = cried out at that, and clutched her breast whose calm wasthe meantime, = as an hors d'oeuvre to his vindictive appetite, heM. de Rivarol's hard = eyes considered him a moment. Perhaps it willThat crossed the swords = between them, and it was against Blood'sMeanwhile, the Frenchmen going = about, gave the like reception to |
|
From: Marge D. <Ma...@ga...> - 2005-07-01 14:01:52
|
Hello, wondered, what was afoot? Her next sentence resolved his doubt.Sure, now, = I am not constraining you at all. I'm giving you acompassion and = charity are lost upon you, and therefore I will sayThe Royal Mary - the = vessel bearing that ingenious, tolerablyMajesty's Government, which you = would not find ungrateful. YouNoting the pause and the look, Esteban = bounded forward, his facethat I consider just and therefore uphold, you = may look for trouble,even within his embrace, which would not be denied; = a look of dreadDon Miguel was bawling orders wildly, and peering ever = and anonpiquant face, which seemed here so out of place, and finding = hisailing. Reaching the summit of a gentle, grassy slope, she met aDon = Francisco shook his head. That must remain my affair, heremained to = ascertain the precise identity of these mysteriousbought and sold was a = new one, and I was hardly in the mood to loveface suddenly white and = drawn.those others who came crowding round him. It's not his way to be |
|
From: Hameed L. <Ham...@jg...> - 2005-06-28 18:00:05
|
Hello, left Barbados, and swearing to follow him loyally whithersoever heit, and = there ran into Pitt, alone, toiling with a wooden spade uponin a red = cotton kerchief, whose blue shirt hung open at the waist.it was blent. = Oh, I don't believe the tale. I won't believe it!back.sensation. Spain = and England were variously and unpleasantlydesire to leave = her.tears.miniature village enclosed in a stockade which the plantation = slavespropose an enterprise of a purely buccaneering character; = whilstThere could be no doubt that he lay in the great cabin of his = ownto France or Holland. But in the long weeks of waiting for a = shipBlood were both guilty of treason, the first for having = harbouredfavour. But sink me now if I'd rot myself in rum on account = ofand what's left of the fort until we put to sea.I intend has been = already accomplished. And lest it increase your |
|
From: Robert L. <ro...@sc...> - 2004-10-04 23:20:51
|
This was a great source of entertainment to find. If
_OSDEP_DEV_NODE_DEINIT happened need to grab the mutex on the
_udi_dev_node of the region it was tearing down (as it might if it was
calling _udi_instance_attr_get which simulates being synchronous with an
insanely complex scheme involving locks and timeouts to retry the fetch)
hilarity would ensue since we'd just called _MUTEX_DEINIT on that very
mutex. Bummer.
With this in place, all the "spank me" memory allocators are happy once
again. (And a fast MP system doesn't come crashing down around your
head if you're lucky and just silently corrupting core if you're not...)
RJL
Index: env/common/udi_MA.c
===================================================================
RCS file: /cvsroot/projectudi/udiref/env/common/udi_MA.c,v
retrieving revision 1.2.2.6
diff -p -u -u -1 -0 -r1.2.2.6 udi_MA.c
--- env/common/udi_MA.c 7 Feb 2003 22:52:48 -0000 1.2.2.6
+++ env/common/udi_MA.c 4 Oct 2004 23:14:19 -0000
@@ -1473,29 +1473,30 @@ _udi_buf_path_list_release(_udi_dev_node
STATIC void
_udi_MA_remove_node(_udi_dev_node_t *node)
{
#if _UDI_PHYSIO_SUPPORTED
int i;
#endif
_udi_instance_attr_release(node);
_udi_buf_path_list_release(node);
UDI_QUEUE_REMOVE(&node->sibling);
- _OSDEP_MUTEX_DEINIT(&node->lock);
+
if (node->attr_address_locator) {
_OSDEP_MEM_FREE(node->attr_address_locator);
}
if (node->attr_physical_locator) {
_OSDEP_MEM_FREE(node->attr_physical_locator);
}
_OSDEP_DEV_NODE_DEINIT(node);
+ _OSDEP_MUTEX_DEINIT(&node->lock);
#if _UDI_PHYSIO_SUPPORTED
for (i = 0; i < node->n_pio_mutexen; i++) {
_OSDEP_MUTEX_DEINIT(&node->pio_mutexen[i].lock);
}
if (node->pio_mutexen) {
_OSDEP_MEM_FREE(node->pio_mutexen);
}
#endif /* _UDI_PHYSIO_SUPPORTED */
|
|
From: Robert L. <ro...@sc...> - 2004-05-11 15:48:28
|
Ooops. Wrong list. Sorry. RJL |
|
From: Robert L. <rob...@us...> - 2004-05-11 15:45:44
|
Rick Richardson wrote:
> Missing long name output from the attached cache .loc file.
>
> $ gpsbabel -igeo -f gc55a5.loc
> 57.458600N 12.753333E GC55A5/ 0.000000
>
> I think its the accented A that is tripping it up. I wouldn't mind
> if it were stipped, but right now the whole name is tossed.
Ron gets another "told you so" realized. This was wrong the day I
checked it in...
$ cvs diff util.c
Index: util.c
===================================================================
RCS file: /cvsroot/gpsbabel/gpsbabel/util.c,v
retrieving revision 1.62
diff -p -u -r1.62 util.c
--- util.c 16 Apr 2004 15:37:30 -0000 1.62
+++ util.c 11 May 2004 15:42:52 -0000
@@ -772,7 +772,7 @@ char * str_utf8_to_ascii( const char * s
if ( *cur & 0x80 ) {
int bytes;
int value;
- char *strvalue = "";
+ char *strvalue = NULL;
utf8_to_int( cur, &bytes, &value );
switch (value) {
case 0x2026: strvalue = "..."; break;
|
|
From: Robert L. <ro...@sc...> - 2003-11-06 19:13:01
|
Robin Farine wrote: > Before going further, I would like to submit the changes for review > and feed-back on things that I should fix/improve/change (for Lovely. Please submit the proposed changes in the form of a patch ala http://projectudi.sourceforge.net/contribute.php for review and incorporation. Thanx, RJL |
|
From: Robin F. <rob...@te...> - 2003-11-06 18:15:03
|
Finally, I found some spare time to invest in this task. The current state of things allows one to build and run cross tools on a GNU/Linux machine (IA32) to generate a kernel environment and driver modules that run on another GNU/Linux architecture. My test target is a StrongARM based embedded system without PCI bus so I only performed limited tests with the pseudo driver. Before going further, I would like to submit the changes for review and feed-back on things that I should fix/improve/change (for instance, it was not clear to me how to define the ABI for this architecture). In case someone involved in the UDI project, like me, uses the Arch(tla) version control system, I can make these changes available for pulling through HTTP. Otherwise, I could generate diffs and email the tar file to where it is appropriate. -- rnf |
|
From: Robert L. <ro...@sc...> - 2003-01-25 00:03:34
|
Once upon a time the sprops code was believed to be totally OS-specific. The way it was passed in and out of the MA was in the udi_driver_t in most places and inside the _OSDEP_DRIVER_T in others. This caused much flail esp. as we moved metas into the fold since the didn't otherwise NEED an _OSDEP_DRIVER_T and, in many cases, were being created in places where we couldn't conveniently get one. While it's tempting to maintain the ability to jack in new sprops code, we've proven that what we have now has a pretty nice abstraction and is used in every environment in the tree. I have on my desk a problem that I'm trying to solve and keeping this separation is proving costly. I'd like to be able to pass around an _udi_std_sprops * in more places but currently cannot since that's technically an OS choice. The realy problem is that the abstraction line was drawn in the wrong place and the interaction between modules, drivers, metas, and drivers was not characterized and defined as well as we should have. (Of course, those are some of the oldest structures in the code base and predate sprops completely, so that's not terribly suprising.) I can provide reasonable wrappers to plaster over the _udi_driver_t vs. _udi_std_sprops * in most cases, but in unwinding this, I have a few cases that I'd like to just stop pretending they're OS-dependent and making up udi_driver_t's just so I can grope the sprops of a meta. Do any maintainers of environments have objections to me committing changes to loosen this up a little? If there are and they're too expensive for me to deal with, I'll just do this for my own environments. RJL |
|
From: <Fra...@ao...> - 2002-09-09 20:16:38
|
Thanks for all your help. I finally got it working :). Now it's time to start coding that Realtek 8139 driver... Cheers, Frank Millea |
|
From: Bryan S. <cor...@hy...> - 2002-09-08 19:46:48
|
I have noticed this problem when doing an install under Debian Linux as well. Basically, depmod is not finding all of the dependant modules, even though they are installed. The module list it uses is passed to it on the command line so that it doesn't sort through ALL of the installed Linux modules. This was done so that the install doesn't take a huge amount of time on a distro like Red Hat that installs all of the possible kernel modules. To get around this run 'make -i install' to ignore those errors. You only have to do this for the install target, so you will still catch build errors. It would also be wise to redirect the output to a log file so you can look through it to verify that there aren't any other errors. Bryan Smant On Sun, 2002-09-08 at 09:35, Fra...@ao... wrote: > Sorry if I double posted this, I wasn't sure if it got through. > > I installed libelf 0.8.2 per the README instructions. I am running on > Slackware Linux 8.1, with a 2.4.18 kernel. I untarred and unzipped the UDI > reference implementation for Linux I got from STG, and ran 'make build' from > the top-level directory. This worked fine. Just plain 'make' also works > fine. My problem is running the next command in the series, 'make install'. > It crashes out after a screen or two, and gave me this: > > --- Making install in linux --- > make[2]: Entering directory `/root/udi/udiforlinux/env/linux' > /usr/bin/install -d /lib/modules/2.4.18/misc; > /usr/bin/install -c ./env_test.o /lib/modules/2.4.18/misc; > /usr/bin/install -c ./mod_test.o /lib/modules/2.4.18/misc; > /sbin/depmod /lib/modules/2.4.18/misc/env_test.o; > depmod: *** Unresolved symbols in /lib/modules/2.4.18/misc/env_test.o > /lib/modules/2.4.18/misc/env_test.o: > > # module id=string > # pci module vendor device subvendor subdevice class > class_mask driver_data > # isapnp module cardvendor carddevice driver_data vendor function > ...# usb module match_flags idVendor idProduct bcdDevice_lo > bcdDevice_hi bDeviceClass bDeviceSubClass bDeviceProtocol bInterfaceClass > bInterfaceSubClass bInterfaceProtocol driver_info > # module pattern > # ieee1394 module match_flags vendor_id model_id specifier_id version > # module id > make[2]: *** [install] Error 1 > make[2]: Leaving directory `/root/udi/udiforlinux/env/linux' > make[1]: *** [install] Error 1 > make[1]: Leaving directory `/root/udi/udiforlinux/env' > make: *** [install] Error 1 > root@general:~/udi/udiforlinux# > > Any help would be appreciated. > > Cheers, > Frank Millea |
|
From: <Fra...@ao...> - 2002-09-08 14:35:43
|
Sorry if I double posted this, I wasn't sure if it got through. I installed libelf 0.8.2 per the README instructions. I am running on Slackware Linux 8.1, with a 2.4.18 kernel. I untarred and unzipped the UDI reference implementation for Linux I got from STG, and ran 'make build' from the top-level directory. This worked fine. Just plain 'make' also works fine. My problem is running the next command in the series, 'make install'. It crashes out after a screen or two, and gave me this: --- Making install in linux --- make[2]: Entering directory `/root/udi/udiforlinux/env/linux' /usr/bin/install -d /lib/modules/2.4.18/misc; /usr/bin/install -c ./env_test.o /lib/modules/2.4.18/misc; /usr/bin/install -c ./mod_test.o /lib/modules/2.4.18/misc; /sbin/depmod /lib/modules/2.4.18/misc/env_test.o; depmod: *** Unresolved symbols in /lib/modules/2.4.18/misc/env_test.o /lib/modules/2.4.18/misc/env_test.o: # module id=string # pci module vendor device subvendor subdevice class class_mask driver_data # isapnp module cardvendor carddevice driver_data vendor function ...# usb module match_flags idVendor idProduct bcdDevice_lo bcdDevice_hi bDeviceClass bDeviceSubClass bDeviceProtocol bInterfaceClass bInterfaceSubClass bInterfaceProtocol driver_info # module pattern # ieee1394 module match_flags vendor_id model_id specifier_id version # module id make[2]: *** [install] Error 1 make[2]: Leaving directory `/root/udi/udiforlinux/env/linux' make[1]: *** [install] Error 1 make[1]: Leaving directory `/root/udi/udiforlinux/env' make: *** [install] Error 1 root@general:~/udi/udiforlinux# Any help would be appreciated. Cheers, Frank Millea |
|
From: <Fra...@ao...> - 2002-09-08 14:12:19
|
I installed libelf 0.8.2 per the README instructions. I am running on Slackware Linux 8.1, with a 2.4.18 kernel. I untarred and unzipped the UDI reference implementation for Linux I got from STG, and ran 'make build' from the top-level directory. This worked fine. Just plain 'make' also works fine. My problem is running the next command in the series, 'make install'. It crashes out after a screen or two, and gave me this: --- Making install in linux --- make[2]: Entering directory `/root/udi/udiforlinux/env/linux' /usr/bin/install -d /lib/modules/2.4.18/misc; /usr/bin/install -c ./env_test.o /lib/modules/2.4.18/misc; /usr/bin/install -c ./mod_test.o /lib/modules/2.4.18/misc; /sbin/depmod /lib/modules/2.4.18/misc/env_test.o; depmod: *** Unresolved symbols in /lib/modules/2.4.18/misc/env_test.o /lib/modules/2.4.18/misc/env_test.o: # module id=string # pci module vendor device subvendor subdevice class class_mask driver_data # isapnp module cardvendor carddevice driver_data vendor function ...# usb module match_flags idVendor idProduct bcdDevice_lo bcdDevice_hi bDeviceClass bDeviceSubClass bDeviceProtocol bInterfaceClass bInterfaceSubClass bInterfaceProtocol driver_info # module pattern # ieee1394 module match_flags vendor_id model_id specifier_id version # module id make[2]: *** [install] Error 1 make[2]: Leaving directory `/root/udi/udiforlinux/env/linux' make[1]: *** [install] Error 1 make[1]: Leaving directory `/root/udi/udiforlinux/env' make: *** [install] Error 1 root@general:~/udi/udiforlinux# Any help would be appreciated. Cheers, Frank Millea |
|
From: Steven B. <sb...@st...> - 2002-08-21 14:20:22
|
On Wed, Aug 21, 2002 at 08:25:31AM +0200, Burkhard Daniel wrote: > Burkhard Daniel wrote: > >Robert Lipe wrote: > > > >>Have I hosed things up in my targets or has the trunk started to fail to > >>build becuase of propfile not being defined? > > propfile, yes. I forgot that I had to manually remove the comments > around the declaration of propfile in spparse.h. Steve commented this > our. Steve, can you verify why this was necessary? There are two ways to fix this: 1) Add "extern char *propfile;" to tools/common/spparse.h. I had removed this. 2) Change propfile references to udi_global.propfile. The only time propfile seems to be different from udi_global.propfile is when the parser is parsing 'message_file <files>' declarations. --Steve |
|
From: Burkhard D. <bu...@st...> - 2002-08-21 06:25:49
|
Burkhard Daniel wrote: > Robert Lipe wrote: > >> Have I hosed things up in my targets or has the trunk started to fail to >> build becuase of propfile not being defined? propfile, yes. I forgot that I had to manually remove the comments around the declaration of propfile in spparse.h. Steve commented this our. Steve, can you verify why this was necessary? > I have seen problems with udisetup on Linux. I am investigating. If > that's hwat's bothering you, go back to the tools as of 2002-06-30, and > they'll work fine. > > What's happening is that udisetup will mysteriously find more files to > work on than just the correct .udi file. Only problem is, the other > "files" have garbage for a name. > > Burk. -- Burkhard Daniel Software Technologies Group, Inc. bu...@st... * http://www.stg.com fon: +49-179-5319489 fax: +49-179-335319489 |
|
From: Burkhard D. <bu...@st...> - 2002-08-21 04:47:11
|
Robert Lipe wrote: > Have I hosed things up in my targets or has the trunk started to fail to > build becuase of propfile not being defined? I have seen problems with udisetup on Linux. I am investigating. If that's hwat's bothering you, go back to the tools as of 2002-06-30, and they'll work fine. What's happening is that udisetup will mysteriously find more files to work on than just the correct .udi file. Only problem is, the other "files" have garbage for a name. Burk. -- Burkhard Daniel Software Technologies Group, Inc. bu...@st... * http://www.stg.com fon: +49-179-5319489 fax: +49-179-335319489 |
2 messages has been excluded from this view by a project administrator.
