Hello!
It seems that during porting to newbus architecture adv_attach was left
unchanged and current version (I mean the current version of -STABLE
sources, I didn't check -CURRENT) returns incorrect values (1 on success, 0 in
all other cases). I've fixed this, the patch is in the attached file.
Andrew Reiter wrote:
You are correct.. However, they were quite useless, imo. This tutorial
will go into detail regarding each step of the skeleton for ading syscalls
and device drivers so that hopefully anyone can quickly learn how to write
them.
I was subtly suggesting you replace
Hi,
In article [EMAIL PROTECTED]
Oleg Sharoiko [EMAIL PROTECTED] writes:
It seems that during porting to newbus architecture adv_attach was left
unchanged and current version (I mean the current version of -STABLE
sources, I didn't check -CURRENT) returns incorrect values (1 on success, 0 in
On May 22, 1:32pm, Matthew Dillon wrote:
} Subject: Re: NFS server problems on 3.4-S, any interest?
} :From the workstation:
} :Name Mtu Network Ipkts IerrsOpktsOerrs Coll Drop
} :fxp0 1500 Link#1 32102492 0 31653667 0 30900 0
30900 collisions is a pretty
At 11:52 +0400 23/5/00, Oleg Sharoiko wrote:
Hello!
It seems that during porting to newbus architecture adv_attach was left
unchanged and current version (I mean the current version of -STABLE
sources, I didn't check -CURRENT) returns incorrect values (1 on success, 0 in
all other cases).
FWIW
On May 22, 3:33pm, Alfred Perlstein wrote:
} Subject: Re: please hep me!
} * David Scheidt [EMAIL PROTECTED] [000522 14:30] wrote:
} dscheidt@shell-2 ~ 536$ ls -al | grep .snapshot
} dscheidt@shell-2 ~ 537$ ls -al .snapshot
} total 60
} drwxrwxrwx 2 root wheel 4096 May
On Tue, 23 May 2000, Bob Bishop wrote:
BB FWIW -CURRENT is the same, but it looks like none of the calls check the
BB return value anyway.
Well. There was a message during the boot of kernel which notified that
"adv0 device prove/attach returned 1" (or smt. like this). So I think the
return code
In [EMAIL PROTECTED], gerald stoller wrote:
I've been looking at the source code for FreeBSD 3.3 and i noticed that a
table-lookUp is called and given the name to look-up and also its hashed
value. Why is the hashed value supplied, why should the caller even be
aware of it (this is
I guess I may have been brain dead or something to not pick up on the last
question in your previous email. I will look at them quickly today and
then work on it saturday (Im leaving to california for a few days starting
tomorrow).
Andrew
On Tue, 23 May 2000, Daniel C. Sobral wrote:
|Andrew
At 16:11 23/05/00 +0400, Oleg Sharoiko wrote:
On Tue, 23 May 2000, Bob Bishop wrote:
BB FWIW -CURRENT is the same, but it looks like none of the calls check the
BB return value anyway.
Well. There was a message during the boot of kernel which notified that
"adv0 device prove/attach returned 1"
Daniel,
First of all, I think you will get a lot more replies if you re-post
your message to freebsd-multimedia (cc'd). Secondly, the author and
maintainer of the new pcm driver in 4-stable and -current is Cameron
Grant ([EMAIL PROTECTED]), and you should talk to him before starting
anything
Is there a document anywhere that explains the newbus function calls?
Technically, i'm supposed to be writing it ;-)
But i am trying to trace some calls and it gets pretty crazy sometimes. I'm
debugging the parallel port zip driver, and i'm not sure which functions are
the entry points. And
Thus spake J McKitrick ([EMAIL PROTECTED]):
Is there a document anywhere that explains the newbus function calls?
I've written a few and am writing some.
Most don't have manual pages.
A good start are the NetBSD manual pages, most stuff is the same.
times, and definitions are scattered all
I think you will get a lot more replies if you re-post
your message to freebsd-multimedia (cc'd).
I've added myself to that list.
Secondly, the author and
maintainer of the new pcm driver in 4-stable and -current is Cameron
Grant ([EMAIL PROTECTED])
That's the subsystem I'm working with.
-*
So, the time has come to think about ripping out the doc distro from
"make release", and to teach sysinstall about these packages instead.
Actually, if the doc distro becomes a set of packages then sysinstall
doesn't need to know about them as a special case at all (which is
good, every
Only seeing 64M on a Dell P550 with 384M of ram
Have tried:
boot -MAXMEM=393216
from loader - any suggestions for a temp fix?
Could I use kernel_options in /boot/defaults/loader.conf - if so what
syntax?
I have RT'd every FM by the way ;-)
I intend to rebuild kernel first chance I get ...
Doug Barton [EMAIL PROTECTED] wrote:
Hrrrmm... I just took a look at the settings for each card. I did not
specify full duplex in the fxp0 ifconfig line, since autoselect has
always worked before.
Autonegotiation is prone to problems. It only really works if both
ends have the same setting
hi all,
i am working on a project to do a zero copy tranfer from a netwrok
buffer to a file. i need to know about buffer allocation in freeBSD. ive
read the source but there are a few queries:
i) what is meant by the B_MALLOC flag?? it seems if this flag is set the
buffers memory is malloced
On Tue, 23 May 2000, Tony Finch wrote:
Doug Barton [EMAIL PROTECTED] wrote:
Hrrrmm... I just took a look at the settings for each card. I did not
specify full duplex in the fxp0 ifconfig line, since autoselect has
always worked before.
Autonegotiation is prone to problems. It only
Answer straight from the manual:
Actually, the .snapshot directory in the mount point is "real" to make
the pwd command work, whereas the .snapshot directories in all other
directories are "magic"; that is, can be accessed when they are
referenced by name but do not show up in a directory
On Tue, 23 May 2000, Don Lewis wrote:
On May 22, 1:32pm, Matthew Dillon wrote:
} Subject: Re: NFS server problems on 3.4-S, any interest?
} :From the workstation:
} :Name Mtu Network Ipkts IerrsOpktsOerrs Coll Drop
} :fxp0 1500 Link#1 32102492 0 31653667 0
hi all,
do disk addresses in struct dinode ( di_db[] array) address block
addresses or fragment addresses?? the comment in dinode.h says they are
block addresses. but fs.h says addresses are capable of addressing
fragments. will somebody please explain this??
thanx all
joy
To Unsubscribe:
Hi all
In a wireless NIC driver should one drain the output queue when the interface is
stopped? I've been perusing /sys/dev/awi.c and the output queue is drained in
that driver.
Duncan
---
Duncan Barclay | God
I've added kernel support for the extra keys on my MS Internet Keyboard Pro
(all 19 of them) to atkbd.c, using keycodes 0x6d-0x7f. (I can submit the
patches to /usr/src/sys/dev/kbd/atkbd.c, /usr/src/sys/dev/kbd/kbdtables.h,
and /usr/share/syscons/keymaps/us.iso.kbd if people are interested, but
Alexander Langer wrote:
Thus spake J McKitrick ([EMAIL PROTECTED]):
Is there a document anywhere that explains the newbus function calls?
I've written a few and am writing some.
Most don't have manual pages.
A good start are the NetBSD manual pages, most stuff is the same.
If a fragment address refers to a fragment at a block boundary, then it is
also called a block address.
-Zhihui
hi all,
do disk addresses in struct dinode ( di_db[] array) address block
addresses or fragment addresses?? the comment in dinode.h says they are
block addresses. but fs.h
Hi,
We have a pre-release of the new modular syslog system is available
for download.
It has been tested under OpenBSD 2.6 and 2.7, and RedHat 6.1, but we'd
like to hear from you.
It's the predecessor of the secure syslog. Here's a brief list of
features:
* Input and output was
Hi,
I apologize beforehand if this topic has already been discussed at
length here or elsewhere.
More and more commerical sites are providing software packages that contain
binaries for Linux. While FreeBSD does provide Linux emulation, this is often
flaky and breaks down more often
I am trying to monitor (eg use bpf tcpdump to see packets) a FDDI
ring with a Digital DEFPA FDDI NIC (and the fpa driver) via a fiber
tap. Therefore the interface only gets to receive, not transmit.
This doesn't work, I assume because the firmware on the DEFPA
wants/needs to participate in the
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
This is definately a matter of opinion.
First of all, FreeBSD's kernel can use both the Linux and FreeBSD
interfaces at once. Although it might be possible to remove the FreeBSD
interfaces, then all you have are the Linux ones - it sounds like a
That sounds a lot like the DaemonLinux project:
http://synack.net/daemonlinux/
Except it appears to have died stillborn.
And not without reason. Their proposal aimed to replace FSF utilities with
BSD equivalents - I don't think they are considering the kernel as a utility.
I
We already have a pretty complete implementation of the Linux kernel ABI -
most of the problems with running Linux binaries on FreeBSD comes from
userland stuff: missing libraries, etc. It's not "Linux emulation" - see
http://www.freebsd.org/handbook/x18949.html
Yes, which is why I'd
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Here's a Slashdot article about Debian/FreeBSD. It has links and a LOT of
angry comments :).
http://slashdot.org/bsd/99/11/23/1939210.shtml
-- Dan Feldman
Hacker, webmaster and computer connoisseur
Out of sight, out
On Tue, May 23, 2000 at 08:56:18PM -0500, Mohit Aron wrote:
Hi, I apologize beforehand if this topic has already been discussed at
length here or elsewhere.
More and more commerical sites are providing software packages
that contain binaries for Linux. While FreeBSD does provide Linux
On Tue, 23 May 2000, Mohit Aron wrote:
Yes, which is why I'd rather use GNU utilities running on FreeBSD than spend
hours figuring out how to make a Linux binary work. As someone pointed out,
Debian is making some effort in this direction. I'll check that out.
Oh I see, you're looking for a
In article local.mail.freebsd-hackers/[EMAIL PROTECTED] you
write:
We already have a pretty complete implementation of the Linux kernel ABI -
most of the problems with running Linux binaries on FreeBSD comes from
userland stuff: missing libraries, etc. It's not "Linux emulation" - see
I think you need to keep in mind that FreeBSD is more than just the
kernel, in contrast to Linux. The tight integration and control of all
of the userspace tools makes system management much easier IMHO with
FreeBSD than any Linux distribution I have used. This is especially true
when you
First of all, FreeBSD's kernel can use both the Linux and FreeBSD
interfaces at once. Although it might be possible to remove the FreeBSD
interfaces, then all you have are the Linux ones - it sounds like a loss
in functionality to me. Second, the Linux emulator is actually extremely
good
On Tue, 23 May 2000, Mohit Aron wrote:
And not without reason. Their proposal aimed to replace FSF utilities with
BSD equivalents - I don't think they are considering the kernel as a utility.
I don't really any benefit from this.
The binaries being distributed for Linux make use of Linux
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
The thing is, using Linux applications on Linux isn't all that easy. There
definately needs to be a better packaging system, but that won't happen
anytime soon.
On the other hand, commercial apps usually have very precise instructions
as to what
On Tue, 23 May 2000, Mohit Aron wrote:
Well, I'm not about to give up FreeBSD running on my desktop, but at times
it is frustrating to not being able to use so much stuff out there that's
meant to work for Linux but doesn't work for FreeBSD for one small reason
or another. I think the user
On the other hand, commercial apps usually have very precise instructions
as to what one should do. If you're a little creative you can generally
break the rules to get these to run on FreeBSD. For instance, I installed
StarOffice by timing the length of the binary first-stage installer
But seriously, I think the problem can be fixed with a more transparent
interface for Linux programs. Rather than requiring Linux libraries to be put
in /compat/linux, it would be much easier if everything could be put in
/usr/lib. Which probably means having the SAME interface as Linux.
On 24-May-00 Mohit Aron wrote:
But seriously, I think the problem can be fixed with a more transparent
interface for Linux programs. Rather than requiring Linux libraries to be
put
in /compat/linux, it would be much easier if everything could be put in
/usr/lib. Which probably means
Perhaps we should go just a bit further with that approach and make
things _write_ into that hierarchy first as well, e.g. if you run
/compat/linux/bin/bash and then install something with rpm, it will
install (as far as it's concerned) into /usr/bin, /usr/lib, etc. but
really be chrooted
On 24-May-00 Mohit Aron wrote:
Yes, that looks promising. That'll possibly enable one to install rpms
easily on FreeBSD.
You can try this too..
rpm --ignoreos --root /compat/linux --dbbath /var/lib/rpm --nodeps
--replacepkgs foo.rpm
I did this to install Linux gtk libs a while ago and
On Tue, 23 May 2000, Mohit Aron wrote:
I believe even to make netscape plugins (for Linux) work, you need to
use the linux version of netscape - not the FreeBSD one (at least this
used to be true some time back). All these nifty things really scare
any new users away from FreeBSD.
You can't
: But seriously, I think the problem can be fixed with a more transparent
: interface for Linux programs. Rather than requiring Linux libraries to be put
: in /compat/linux, it would be much easier if everything could be put in
: /usr/lib. Which probably means having the SAME interface as
Consider a little test program:
#include stdio.h
extern int func_a( int, int );
int main()
{
int i;
for( i = 0; i 100; i++ ) {
printf( "%d\n", func_a( i, i+2 ) );
}
return 0;
}
Part of assembler code built by gcc 2.7.2.3 (on FreeBSD-3.4-stable):
pushl %ebp
49 matches
Mail list logo
