Linux-Development-Sys Digest #50, Volume #7 Fri, 13 Aug 99 08:14:05 EDT
Contents:
Re: Diamond v770 drivers? (David T. Blake)
Re: Broken Select (Kaz Kylheku)
kernel 2.2.11 slower than 2.2.5?? (System User)
Broken Select ("Al Stanley")
Re: How do I use source RPMs (David Fox)
Re: Stupid question re: sound config ("Ashutosh S. Rajekar")
big-endian Linux on Alpha? ([EMAIL PROTECTED])
Re: Toshiba DVD-RAM and Linux (Christoph Martin)
Re: module configuration (Karl Heyes)
Re: Difference between RH 2.2.5-15 and general 2.2.5 (Villy Kruse)
Re: Kernel upgrades... why? (Greg White)
Re: Group ID maximum (Ralph Geissler)
Re: Group ID maximum (Ralph Geissler)
Question on the timeout for un-ACK-ed SYN's (Tim)
Re: Difference between RH 2.2.5-15 and general 2.2.5 (Robin Becker)
Re: Group ID maximum (Ralph Geissler)
Re: Group ID maximum (Miquel van Smoorenburg)
Difference between RH 2.2.5-15 and general 2.2.5 ("Jackson C. Allen")
Re : mmap corrupting system -- help! (Gilles Alain Pokam Tientcheu)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (David T. Blake)
Subject: Re: Diamond v770 drivers?
Date: 12 Aug 1999 13:39:50 GMT
Reply-To: [EMAIL PROTECTED]
Chris Habgood <[EMAIL PROTECTED]> wrote:
> Anybody know where to get diamond v770 video drivers?
XF86_SVGA
Just set one up two days ago. Just don't select
24 bit color - it likes 32 instead.
--
Dave Blake
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Kaz Kylheku)
Subject: Re: Broken Select
Date: Fri, 13 Aug 1999 04:45:28 GMT
On Thu, 12 Aug 1999 19:02:10 -0500, Al Stanley <[EMAIL PROTECTED]>
wrote:
>The select() functionality in the 2.2.9 kernel appears to be broken. If
>there is nothing readable or writable when select() is initially called, it
>will not return until it times out. The man page says (and the way it is
>normally used) that it will return when something int readfds becomes
>readable or something in writefds becomes writable, and that the timeout is
>only an upper boundary. Our testing indicates that it will not return until
>the timeout expires. Anyone know of a workaround? Or has anyone else
>noticed this problem yet?
I've been using 2.2.9 since it came out, and I haven't seen any strange
behaviors in any networking software. If what you are saying is true, than it
would probably follow that selects with a null struct timeval * argument would
block forever, resulting in a lot of catatonic applications. Have you
investigated what happens if you make the wait infinite?
------------------------------
From: System User <[EMAIL PROTECTED]>
Subject: kernel 2.2.11 slower than 2.2.5??
Date: Thu, 12 Aug 1999 19:25:03 -0500
Hi.. I had 2.2.5-22 installed, and this was my RAID checksum speed:
pII_mmx : 975.741 MB/sec
p5_mmx : 945.261 MB/sec
8regs : 691.134 MB/sec
32regs : 496.443 MB/sec
I just upgraded to 2.2.11-1, and now I get:
pII_mmx : 891.921 MB/sec
p5_mmx : 936.879 MB/sec
8regs : 688.467 MB/sec
32regs : 385.953 MB/sec
What's up with this? Should I downgrade, just to get more speed? we're
running on a celeron 686, and it's choosing now to run on the fastest
function(586).. ??
Please let me know.. thanks
Joe
[EMAIL PROTECTED]
------------------------------
From: "Al Stanley" <[EMAIL PROTECTED]>
Subject: Broken Select
Date: Thu, 12 Aug 1999 19:02:10 -0500
The select() functionality in the 2.2.9 kernel appears to be broken. If
there is nothing readable or writable when select() is initially called, it
will not return until it times out. The man page says (and the way it is
normally used) that it will return when something int readfds becomes
readable or something in writefds becomes writable, and that the timeout is
only an upper boundary. Our testing indicates that it will not return until
the timeout expires. Anyone know of a workaround? Or has anyone else
noticed this problem yet?
Ed Okerson
Quicknet Technologies, Inc.
------------------------------
From: d s f o x @ c o g s c i . u c s d . e d u (David Fox)
Subject: Re: How do I use source RPMs
Date: 12 Aug 1999 17:12:04 -0700
[EMAIL PROTECTED] writes:
> But if you grab a source RPM and it doesn't build in your environment,
> because it's different than the author's environment, the it's a pain
> in the ass.
Packages that don't build in your environment are a pain in the ass no
matter whether they are RPMs or tarballs. At least with an RPM you
have an idea of how the package got built in *someone's* environment.
Usually a few small modifications or corrections to the spec file are
all that are needed.
--
David Fox http://hci.ucsd.edu/dsf xoF divaD
UCSD HCI Lab baL ICH DSCU
------------------------------
From: "Ashutosh S. Rajekar" <[EMAIL PROTECTED]>
Subject: Re: Stupid question re: sound config
Date: Fri, 13 Aug 1999 12:56:26 -0500
Hello,
On Fri, 13 Aug 1999, root wrote:
> I'm a newbie, so to speak. Any help would be appreciated -- I can't
> seem to get my sound card to work. It's a trident 4d PCI card. No
> ideas.
You will probably have the file 'sndconfig' installed on your box.
At the shell prompt, type sndconfig, and it will probably autodetect your
sound card. If it doesn't, then read the HOW-TO's in /usr/doc/HOWTO/***.
> I tried Lothar - no help. Please help me, oh, the ignorant
> one....
>
>
>
>
>
Thanks,
====================
Ashutosh S. Rajekar
------------------------------
From: [EMAIL PROTECTED]
Subject: big-endian Linux on Alpha?
Date: Fri, 13 Aug 1999 08:24:59 GMT
Sorry if this question has already been posted,
could not find any reference anywhere.
Is it possible to run Linux on an Alpha in
big-endian mode? I know the machine has the
capability of changing sex by use of a single
switch, but is it supported by the Linux kernel?
If it is possible, how? How difficult? Does this
mean you have to recompile everything from scratch?
Which format are the RPMs distributed in?
Thanks for helping
--
Nicolas
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: Christoph Martin <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: Toshiba DVD-RAM and Linux
Date: 13 Aug 1999 10:38:39 +0200
"Hartmut W. Malzahn" <[EMAIL PROTECTED]> writes:
>
> Can you give me a hint how I might reproduce the read problem with the
> drive? I've written a quick (and very dirty :) fix for the write problem
> and the drive appears to work fine now...
If I remember right, it was reading large chunks from a DVD-ROM (not
RAM), like parts of a film, via cp or dd. I try to reproduce it this
weekend and give you the details then.
> I've attached the write patch (patches
> /usr/src/linux/drivers/scsi/scsi.c on kernels 2.2.5 or 2.2.6, probably
> on later ones too) to this message, maybe it helps someone.
Thanks. I try the patch this weekend and let you know, if it work for
me.
--
============================================================================
Christoph Martin, Uni-Mainz, Germany
Internet-Mail: [EMAIL PROTECTED]
------------------------------
From: Karl Heyes <[EMAIL PROTECTED]>
Subject: Re: module configuration
Date: Thu, 12 Aug 1999 13:14:21 +0100
Guilhem Tardy wrote:
> Hi!
>
> Anyone could point me to some better info about how the modules are
> loaded and their configuration (I got some from the depmod / modprobe /
> kerneld man pages)?
>
> I use the initrd command to load the SCSI module (works fine), and this
> is what is loaded before anything else (I believe that only the SCSI
> lines of the file "/etc/conf.modules" are used here). Still, I would
> have some other modules loaded at boot time.
> Using the RH 6.0's Control Panel, I can't edit the modules because I
> installed a new kernel (2.2.11) and didn't figure out how to create a
> new file "module-info-2.2.11"?
> I am ready to edit the file "/etc/conf.modules" by hand, but I have no
> clue about the syntax of the options either (e.g. IRQ).
>
> Well, all this sounds like a real lack of info! Hope you'll have some
> links for me...
> Guilhem.
use modprobe -c to see the current configuration. it will give you an
example of the syntax for the /etc/conf.modules file.
any parameters will go on the option lines (eg irq=10). There are howto's
under /usr/doc/HOWTO on this and a Documentation directory under the
kernel source tree (try /usr/src/linux). The actual arguments to modules
depends on the module in question.
Also kerneld is now redundant under 2.2. now a thread (kmod) is started
which invokes modprobe.
karl.
------------------------------
From: [EMAIL PROTECTED] (Villy Kruse)
Subject: Re: Difference between RH 2.2.5-15 and general 2.2.5
Date: 13 Aug 1999 09:31:27 +0200
In article <[EMAIL PROTECTED]>,
Jackson C. Allen <[EMAIL PROTECTED]> wrote:
> My question is what is the difference? How does RH's patches track with
>the general kernel patches? I have submitted this question to RH 3 days
>ago, but have not gotten any response other than thank you for
>submitting your question.
If you realy want to know, grap the kernel source package from
the SRPMS directory and unpack it into /usr/src/redhat/SOURCES and
/usr/src/redhat/SPECS. The diferences will be reflected in the patch
files found in /usr/src/redhat/SOURCES.
Alot of download, though.
Villy
------------------------------
From: Greg White <[EMAIL PROTECTED]>
Subject: Re: Kernel upgrades... why?
Date: Fri, 13 Aug 1999 04:03:59 GMT
bobrien wrote:
>
> I can understand making the jump from 1.x or 2.1 or 2.2 but it is really a
> necessity or even useful to upgrade each release... 2.2.5-15, 2.2.5-20,
> 2.2.10, 2.2.11 ... etc... I'm still quite new to this and curious why
> everyone seems to jump right into each tiny upgrade. I can't see a reason
> to upgrade until my soundcard has built in support in the kernel. Or I can
> plug in a module for it.
>
> just curious,
> Blake
1. You just _have_ to try the newest kernel, to see what's new in it, to
see if it is
faster, etc.
2. You have more than one box installed, maybe in other places,
customer's machines, or
whatever, and you're testing a new kernel feature, or waiting until a
beta feature works
to your satisfaction.
3. Current support for your existing hardware is constantly improving.
4. You are actually directly affected by a fix included in that
particular release.
I always belong to group one and two, and currently am in group 3,
anxiously awaiting the
native support for my ALi Aladdin controllers to migrate down from the
2.3.x tree. (Tried the 2.3.x tree for quite a while, the native drivers
_are_ faster, but there is too much filesystem support development going
on in that tree for me to be comfortable with it).
GW
------------------------------
From: Ralph Geissler <[EMAIL PROTECTED]>
Subject: Re: Group ID maximum
Date: Fri, 13 Aug 1999 11:29:56 +0200
Chris Gregory wrote:
>
> On Wed, 11 Aug 1999 01:47:27 +0200, Ralph Geissler <[EMAIL PROTECTED]> wrote:
>
> This is not really a good way to do this, because users generally own their
> own directories. If one of them thinks to chmod go-rwx ~ then your plan will
> fail. It's better just to do the read as root.
I maybe should have mentioned that the application/user that needs to
read the files is the Apache webserver. So the users better make their
directories group-readable, so that their homepages are displayed. It is
in their own interest.
But I don't want Apache to run as root..
Thanks for the answer, anyway.
Bye
Ralph
------------------------------
From: Ralph Geissler <[EMAIL PROTECTED]>
Subject: Re: Group ID maximum
Date: Fri, 13 Aug 1999 11:47:37 +0200
Stephen Satchell wrote:
>
> >On Wed, 11 Aug 1999 01:47:27 +0200, Ralph Geissler <[EMAIL PROTECTED]>
> >wrote:
> >
> >>I have a user account on a Linux box (Intel) that has to be able to
> >>read a hundert user's directories without the users reading the
> >>directories of other users. I don't want my user account to be root.
>
> I've been thinking about this requirement for quite some time, and there
> are other ways to do the same thing. Here are a couple:
>
> 1) If the file base is small, you can write a setuid script that will
> mirror the filesets of your hundred users into a tree belonging to your
> supervising user. That script would need to: (a) copy the file, (b)
> change the owner of the copies to the owner, and (c) log the event for
> your review. To do this quickly and easily, consider using "tar" or
> "cpio" to perform the collection and tree creation. This means that your
> supervising user doesn't have to be running as root, and you can put all
> sorts of authentication into the script. Watch out for how you handle
> signals so you don't provide a means of gaining root access.
The files aren't that large, but I don't want such a redundancy. I
generally don't like setuid *scripts*. You mention signal handling, can
I make script usage safer by handling signals in the script? How do I do
that?
<snip>
> Just remember that your hundred users can use "chmod" to make any file
> visible to others, simply by setting the world-readable bit...
If they *want* their files to be world-readable, I have nothing against
that, but most of them *don't* want this.
Thanks for the suggestions.
Bye
Ralph
------------------------------
From: Tim <[EMAIL PROTECTED]>
Subject: Question on the timeout for un-ACK-ed SYN's
Date: Thu, 12 Aug 1999 08:45:46 -0700
I am new to the Linux world and I am looking where the timeout is set
for SYN's that have not been completed. The number of SYN's that can be
backlogged can be set in the tcp_max_syn_backlog file in
/proc/sys/net/ipv4 but where is the timeout set for the SYN's once they
are in the backlog??
Any help is appreciated.
Tim
Tim Brandt
[EMAIL PROTECTED]
------------------------------
From: Robin Becker <[EMAIL PROTECTED]>
Subject: Re: Difference between RH 2.2.5-15 and general 2.2.5
Date: Fri, 13 Aug 1999 10:39:03 +0100
In article <[EMAIL PROTECTED]>, Jackson C. Allen
<[EMAIL PROTECTED]> writes
> I have an application that requires kernel 2.2.10. Well I have RH 6.0
>which is kernel 2.2.5-15. I went looking for the patch files to bring it
>up to 2.2.10, found then, and down loaded them. When I tried to apply
>the first patch I got a lot of rejects. I checked the files the rejects
>were for and found that what the patch was trying to change I already
>had. So i went back to the RH WEB page and found the upgrade for kernel
>2.2.5-22.
>
> My question is what is the difference? How does RH's patches track with
>the general kernel patches? I have submitted this question to RH 3 days
>ago, but have not gotten any response other than thank you for
>submitting your question.
>
>Thanks in advance for any help.
>Jack Allen
Look at the diffs they've done a lot of their own patching in all sorts
of places. I just recompiled with stock 2.2.11 and it works for me.
remember to add an extraversion string in makefile and copy the relevant
files/modules into /boot.
I've also been asking simple questions like why are you trying to link
in the kernel versioning files in /boot when /boot isn't yet mounted etc
etc. Of course I paid for the CD so I can at least ask.
--
Robin Becker
------------------------------
From: Ralph Geissler <[EMAIL PROTECTED]>
Subject: Re: Group ID maximum
Date: Fri, 13 Aug 1999 11:47:39 +0200
Ralph Geissler wrote:
>
> I have a user account on a Linux box (Intel) that has to be able to read
> a hundert user's directories without the users reading the directories
> of other users. I don't want my user account to be root.
I should have mentioned that the user that needs to access all the files
is the Apache webserver user account.
I have found a different solution for this problem. In the new setup,
the files and directories belong the users, but belong to the group my
Apache user ('www' in this case). The files and directories have to be
group-readable. My user can read all the files but the users can't
access the files of the other users. The users belong to the 'users'
group, so there is no need for a group for every user.
The only problem with this setup is that I have to keep the files in the
'www' group. I fiddled a little with scripts that chown the files, but
there is a simpler solution: make the directories setgid 'www' with
'chmod g+s www'! New files and directories created in those directories
will automatically belong to the 'www' group and my problem is solved.
By
Ralph
------------------------------
From: [EMAIL PROTECTED] (Miquel van Smoorenburg)
Subject: Re: Group ID maximum
Date: 13 Aug 1999 12:00:36 +0200
In article <[EMAIL PROTECTED]>,
Ralph Geissler <[EMAIL PROTECTED]> wrote:
>Chris Gregory wrote:
>>
>> On Wed, 11 Aug 1999 01:47:27 +0200, Ralph Geissler <[EMAIL PROTECTED]> wrote:
>>
>
>> This is not really a good way to do this, because users generally own their
>> own directories. If one of them thinks to chmod go-rwx ~ then your plan will
>> fail. It's better just to do the read as root.
>
>I maybe should have mentioned that the application/user that needs to
>read the files is the Apache webserver. So the users better make their
>directories group-readable, so that their homepages are displayed. It is
>in their own interest.
Let the users open their base homedirs for chdir by setting the 'x' bit,
chmod o+x ~username. Then make their public_html directory or whatever
you are calling it owned by the user, group 'www' (or whatever group
apache runs as) and mode 750.
Mike.
--
... somehow I have a feeling the hurting hasn't even begun yet
-- Bill, "The Terrible Thunderlizards"
------------------------------
From: "Jackson C. Allen" <[EMAIL PROTECTED]>
Subject: Difference between RH 2.2.5-15 and general 2.2.5
Date: Fri, 13 Aug 1999 04:18:34 GMT
I have an application that requires kernel 2.2.10. Well I have RH 6.0
which is kernel 2.2.5-15. I went looking for the patch files to bring it
up to 2.2.10, found then, and down loaded them. When I tried to apply
the first patch I got a lot of rejects. I checked the files the rejects
were for and found that what the patch was trying to change I already
had. So i went back to the RH WEB page and found the upgrade for kernel
2.2.5-22.
My question is what is the difference? How does RH's patches track with
the general kernel patches? I have submitted this question to RH 3 days
ago, but have not gotten any response other than thank you for
submitting your question.
Thanks in advance for any help.
Jack Allen
------------------------------
From: Gilles Alain Pokam Tientcheu <[EMAIL PROTECTED]>
Subject: Re : mmap corrupting system -- help!
Date: 13 Aug 1999 11:05:11 GMT
In article <[EMAIL PROTECTED]> you wrote:
> I'm writing a simple driver for a PCI board that only supports
> open, close, and mmap[*]. It has 3 minor numbers that mmap
> different 3 differnt things:
> A page of memory allocated by the driver for use as a DMA buffer
>
> The third one sort-of works (the userprogram sees the right page of memory),
> but it's corrupting things (processes are croaking mysteriously,
> and the memory management system complained about that page when shutting
> down.)
>
> Here's the mmap code:
>
> static unsigned long scratch_mmap_nopage(struct vm_area_struct * area,
> unsigned long address,
> int error_code)
>{
> int pos;
> address &= PAGE_MASK;
> pos = address - area->vm_start + area->vm_offset;
> printk("mm nopage start=%08lx end=%08lx offs=%08lx\n",
> area->vm_start, area->vm_end, area->vm_offset);
> if (pos < 0 || pos >= PAGE_SIZE)
> {
> printk("invalid position = %d\n", pos);
> return 0; /* seg fault user process */
> }
>
> printk("mm nopage returning %08x\n",scratchPage);
> return scratchPage;
> }
Looking at your code i think you may have missed somethink. The nopage method like
described
in Rubini's book should return the physical address of the missing page. The "address"
argument in your function scratch_mmap_nopage() is the virtual address of the missing
page. I think what you need to do is to retrieve the real physical address to which
this virtual address is mapped. What your doing instaed is to allocate a new physical
page in memory that doesn't necessary corresponds to the desired page. If you use a
DMA buffer that has been allocated with _get_free_pages or similar to hold your data
in kernel space, you should then increment the usage count of each page in your
nopage method.
Regards,
[EMAIL PROTECTED]
[EMAIL PROTECTED]
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.development.system) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Development-System Digest
******************************
