Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.

Dismiss

freebsd-hackers-digest V5 #730

0 views

Skip to first unread message

owner-freebsd-...@freebsd.org

unread,

Feb 26, 2003, 8:00:10 AM2/26/03

freebsd-hackers-digest Wednesday, February 26 2003 Volume 05 : Number 730

In this issue:
HOWTO track resource leaks in kernel modules ?
Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix needs to be committed.
Re: booting from Promise tx2000: FIXED
Re: booting from Promise tx2000: FIXED
=?Big5?B?p0G3UcX9rmGkSLlMp/Ombqq6pc2sobbcPw==?=
Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix needs to be committed.
Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix needs to be committed.
Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix need
Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix need
Touchpad program
FreeBSD 5.0 roadmap
Isn't today Troll Tuesday?
Isn't today Troll Tuesday?
i386 tinderbox failure
IA-64 tinderbox failure
Newest Arrivals
Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix need (fwd)
new freebsd distribution...
Jail seperation patch
Re: Jail seperation patch
Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix need (fwd)
ESS1868 sound card and the *infamous* play interrupt timeout, channel dead:)

----------------------------------------------------------------------

Date: Tue, 25 Feb 2003 04:09:44 -0800
From: Daxbert <daxber...@dweebsoft.com>
Subject: HOWTO track resource leaks in kernel modules ?

Hi -

I was thinking about making some changes to if_de.c to support a tulip card
which isn't being recognized properly.

To begin this process, I need to make if_de.c export a detach function so it
could be unloaded from the kernel to make debugging of the module a little
easier. I've implemented a detach function, which at first glance seems ok.

I've run an extremely crude test by kldload / kldunload in an infinte loop for
about an hour... and the system hasn't panic-ed. And memory utilization doesn't
appear to have grown. However, I'd like something a bit more accurate.

Where would I look for resource leaks? Is there a library or toolkit to track
such things for kernel modules?

Sorry for my ignorance, but most of my Unix development experience has been
centered around apache modules.

Thanks,

- --daxbert

------------------------------

Date: Tue, 25 Feb 2003 13:10:25 +0100 (CET)
From: Martin Blapp <m...@imp.ch>
Subject: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix needs to be committed.

Hi all,

To tell the short story. Linux-mozilla works like a charm as root, but it
doesn't as a user if you have a java-applet.

Some solutions are mentioned for jdk4 but they applly to jdk13 too:

http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&th=5dfc5db60b48af3f&rnum=1
http://docs.freebsd.org/cgi/getmsg.cgi?fetch=132728+0+archive/2002/freebsd-java/20020714.freebsd-java

Started as user:

55909 java_vm RET mprotect 0
55909 java_vm CALL mprotect(0x28b21000,0x9000,0x5)
55909 java_vm RET mprotect 0
55909 java_vm CALL linux_brk(0x8050000)
55909 java_vm RET linux_brk 134545408/0x8050000
55909 java_vm CALL linux_sched_getscheduler(0xda65)
55909 java_vm RET linux_sched_getscheduler RESTART
-> BOOM

Started as root:

55836 java_vm RET mprotect 0
55836 java_vm CALL mprotect(0x28b21000,0x9000,0x5)
55836 java_vm RET mprotect 0
55836 java_vm CALL linux_brk(0x8050000)
55836 java_vm RET linux_brk 134545408/0x8050000
55836 java_vm CALL linux_sched_getscheduler(0xda1c)
55836 java_vm RET linux_sched_getscheduler 0
55836 java_vm CALL sched_getparam(0xda1c,0xbfbf9444)
55836 java_vm RET sched_getparam 0
55836 java_vm CALL linux_sched_getscheduler(0xda1c)
55836 java_vm RET linux_sched_getscheduler 0
55836 java_vm CALL sched_getparam(0xda1c,0xbfbf9444)
55836 java_vm RET sched_getparam 0

[...]

Martin

------------------------------

Date: Tue, 25 Feb 2003 06:22:56 -0600
From: Len Conrad <LCo...@Go2France.com>
Subject: Re: booting from Promise tx2000: FIXED

>Len Conrad <LCo...@Go2France.com> writes:
> > .... while waiting for Soeren Schmidt to get the Promise SX4000 driver
> done!
>
>I was under the impression that the SX4000 and SX6000 were already
>supported? I know that phk has an SX6000 which he says works fine.
>OTOH, it's possible that this hasn't percolated down to -STABLE yet.

Last autumn, SS and I tried to get SX4000 docs from Promise and I was told
by Promise .tw that the "Promise doesn't support FreeBSD", and SS has lost
his earlier contact person at Promise. So it's a happy surprise that I
learn this week that SS now has coop from Promise. He didn't say the
driver was ready.

Len

------------------------------

Date: Tue, 25 Feb 2003 13:45:10 +0100 (CET)
From: Soeren Schmidt <s...@spider.deepcore.dk>
Subject: Re: booting from Promise tx2000: FIXED

It seems Dag-Erling Smorgrav wrote:
> Len Conrad <LCo...@Go2France.com> writes:
> > .... while waiting for Soeren Schmidt to get the Promise SX4000 driver done!
>
> I was under the impression that the SX4000 and SX6000 were already
> supported? I know that phk has an SX6000 which he says works fine.
> OTOH, it's possible that this hasn't percolated down to -STABLE yet.

The SX6000 is supported, the SX4000 is quite a different animal and
is not supported yet. However I'm working with Promise to write
support for it...

- -S鷨en

------------------------------

Date: Tue, 25 Feb 2003 22:22:08 +0800
From: =?Big5?B?p9qmYrRNp+QuLi4uLi4u?= <pop9...@yahoo.com.tw>
Subject: =?Big5?B?p0G3UcX9rmGkSLlMp/Ombqq6pc2sobbcPw==?=

<html>

<body>

深夜1點了
躺在床上的我不斷的思索...無法入睡
腦中出現的是父母親疲憊的身體...
我睡不著...睡不著.......
曾經...我誇下海口要讓他們過好日子
可是微薄薪水卻無法讓我履行承諾
幾個月前...我接觸到了一片光碟...
短短的40幾分鐘...我看到了希望

我一步一步的在對的地方為家努力

而我的承諾也將兌現

如果你也想讓家人過更好的生活

請留下資料,我將把這片光碟寄給你

我保證,只要你看懂了,一定會為你開啟亮麗的人生

姓名 
姓別 
年齡 
家中電話 
行動電話 
郵寄地址 
郵遞區號 
請寄到<a href="mailto:pop9...@yahoo.com.tw">pop9...@yahoo.com.tw</a>

</body>

</html>

------------------------------

Date: Tue, 25 Feb 2003 06:56:17 -0800
From: Murray Stokely <mur...@FreeBSD.ORG>
Subject: Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix needs to be committed.

- --NQTVMVnDVuULnIzU
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Tue, Feb 25, 2003 at 01:10:25PM +0100, Martin Blapp wrote:
> To tell the short story. Linux-mozilla works like a charm as root, but it
> doesn't as a user if you have a java-applet.

This has been a problem for years and it affects other Linux
appications such as LabView. There is some commented out code in
linux_sched_getscheduler() function that provides the proper
functionality. It's unclear to me why the code was commented out in
the first place, but it would be really nice to get that resolved once
and for all.

- Murray

- --NQTVMVnDVuULnIzU
Content-Type: application/pgp-signature
Content-Disposition: inline

- -----BEGIN PGP SIGNATURE-----

iD8DBQE+W4QRtNcQog5FH30RAi+hAJ4isrArLzRudfiKH7lPQbdNHefWgACgkvBv
D0po4VwnRANdMaMwa9LnWRU=
=dJPN
- -----END PGP SIGNATURE-----

- --NQTVMVnDVuULnIzU--

------------------------------

Date: Tue, 25 Feb 2003 15:39:38 +0000
From: Bob Bishop <r...@gid.co.uk>
Subject: Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix needs to be committed.

At 14:56 25/2/03, Murray Stokely wrote:
>This has been a problem for years and it affects other Linux
>appications such as LabView. There is some commented out code in
>linux_sched_getscheduler() function that provides the proper
>functionality. It's unclear to me why the code was commented out in
>the first place, but it would be really nice to get that resolved once
>and for all.

According to kern/40611, the problem is in posix4/p1003_1b.c not in the
linux wrapper. The updated patch in the PR audit trail seems to work.

- --
Bob Bishop +44 (0)118 977 4017
r...@gid.co.uk fax +44 (0)118 989 4254

------------------------------

Date: Tue, 25 Feb 2003 11:20:34 -0500 (EST)
From: John Baldwin <j...@FreeBSD.org>
Subject: Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix need

On 25-Feb-2003 Bob Bishop wrote:
> At 14:56 25/2/03, Murray Stokely wrote:
>>This has been a problem for years and it affects other Linux
>>appications such as LabView. There is some commented out code in
>>linux_sched_getscheduler() function that provides the proper
>>functionality. It's unclear to me why the code was commented out in
>>the first place, but it would be really nice to get that resolved once
>>and for all.
>
> According to kern/40611, the problem is in posix4/p1003_1b.c not in the
> linux wrapper. The updated patch in the PR audit trail seems to work.

There is a much simpler patch one can do:

Index: p1003_1b.c
===================================================================
RCS file: /usr/cvs/src/sys/posix4/p1003_1b.c,v
retrieving revision 1.5.2.1
diff -u -r1.5.2.1 p1003_1b.c
- --- p1003_1b.c 3 Aug 2000 01:09:59 -0000 1.5.2.1
+++ p1003_1b.c 25 Feb 2003 16:17:55 -0000
@@ -62,7 +62,7 @@
* only root can do this.
*/

- -#if 0
+#if 1
/*
* This is stolen from CANSIGNAL in kern_sig:
*

Basically, it changes p31b_proc() to not always return an error
for non-root. If rwaston@ signs off on the security implications
(should be minimal, basically means that you can change your own
scheduling params and can change the params of other processes
you own) then I would prefer this patch.

I don't know why the check was turned off. The entire #if 0 /
#else / #endif seems to have been around since revision 1.1.

- --

John Baldwin <j...@FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve!" - http://www.FreeBSD.org/

------------------------------

Date: Tue, 25 Feb 2003 08:44:19 -0800
From: Murray Stokely <mur...@FreeBSD.org>
Subject: Re: Jdk13/14 still hangs in 4.8 Prerelease. Outstanding Fix need

- --t4apE7yKrX2dGgJC
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Tue, Feb 25, 2003 at 11:20:34AM -0500, John Baldwin wrote:
> Basically, it changes p31b_proc() to not always return an error
> for non-root. If rwaston@ signs off on the security implications
> (should be minimal, basically means that you can change your own
> scheduling params and can change the params of other processes
> you own) then I would prefer this patch.

Yes this is the one I was thinking about when I said the commented out
code. I've run a kernel with this patch for months, and I didn't see
any problems with it in my (unqualified) review. ;)

- Murray

- --t4apE7yKrX2dGgJC
Content-Type: application/pgp-signature
Content-Disposition: inline

- -----BEGIN PGP SIGNATURE-----

iD8DBQE+W51itNcQog5FH30RAlBDAJ4vZ8XvLSafyXhmjJ2ykjHwqJ3rCACeNwSo
7hSh5DCxEYEFic/pYIyq/KU=
=M7i7
- -----END PGP SIGNATURE-----

- --t4apE7yKrX2dGgJC--

------------------------------

Date: Tue, 25 Feb 2003 12:09:05 -0600
From: roc...@galileo.edu
Subject: Touchpad program

I want to make a program (or pseudo-device) to emulate the scrolling wheel in my
laptop touchpad (alps), the same way it is done in the windows drivers: if you
drag your finger along the right side it will act as the wheel scrolling. I had
been looking into psm, atkbd and atkbdc but I still don't know where to really
start. If some one could give me a basic help to where to start looking or to
what files I shoul look at, it would be very helpfull. Thanks in advance.

Please CC to my address as I am not currently subscribed to this list.

Rodrigo F.

------------------------------

Date: Tue, 25 Feb 2003 10:29:52
From: "elcott" <elc...@mailfreeway.com>
Subject: FreeBSD 5.0 roadmap

Fellow committers, let's have a look at the 5.0 planned roadmap:

1) KSE

KSE is a joke at best. Is one of those over-engineered ideas that will never be finished. Too bad, it looked good some months ago.

2) GEOM

GEOM is another ego trip for Poul-Henning Kamp. He won't let anyone touch it or improve it. Another piece of code that will rot, like phkmalloc and devd.

3) devfs and devd

Like Bruce Evans, I don't see what the advantage of having this is. Another ego trip Poul?

4) UFS2

Unlike other parts of FreeBSD, UFS2 is now a reality, kudos to you Mr. McKusick.

5) gcc + toolchain

Until Troll Glass brings us TenDRA, we'll have to do with Gah! Nu's proprietary software. Thankfully, we have Mr. O'Brien on board to take care of it. Also Mr. Kabaev as well.

6) Ports

The ports people have done an excellent job, kudos to them.

7) PowerPC port

Little progress in this area, come on Benno, you can do better.

8) IA-64

Mr. LNUX Torvalds thinks it's not good, so it must be a heck of a cpu. As we all know, Linux is pure hore sh*t.

9) RAIDframe

Pathetic!!!! Those RAIDframes are *crap*. Scott, not only you fscked up with the release, but your patches are crap!

Sincerely,
Elcott Song, RE