[Mark Murray]
[Matt Dillon]
Does anyone on this forum have any experience using these sorts of
cards with FreeBSD and could you impart some of your general
knowledge to me? I think the forum would be interested as well.
Mike Smith was doing this as least as long ago as 1999
worth of
work!
(not the) Mike Smith
Dan Nelson wrote:
In the last episode (Jul 26), PSI, Mike Smith said:
Hey all,
Does anyone have any idea what the process 1820 bpbkar is???
Got the infamous page fault trap 12 indicating this is current process.
I cannot find bpbkar ANYWHERE
I'm sure it has to do with something we are augmenting.
And we are using (gasp) 3.2 believe it or not.
Help only if you already know. Not worth any extra effort. Thanks
(Not the) Mike Smith
[EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers
installed in
quite some time so if this is already in place, I apologize.
Just a thought.
(Not THE) Mike Smith
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
that this is 3.2 and thus may be ancient
history.
(not THE) Mike Smith
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
to to strip your system
down to bare bones to eliminate holes. You could bring it up to a fully
capable system at any time it was necessary.
So Is the kernel still running after a halt? If so, has anyone found a
way to take advantage of this?
(Not THE) Mike Smith
To Unsubscribe: send mail
In article [EMAIL PROTECTED],
Dominic Marks [EMAIL PROTECTED] wrote:
On Mon, Feb 04, 2002 at 01:21:25PM -0800, John Polstra wrote:
I'm trying to understand the timecounter code, and in particular the
reason for the microuptime went backwards messages which I see on
just about every
this message has no meaning to you. Oh to be
young again.
Mike Smith
[EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
even remember
where we left our teeth. So by tomorrow, we definitely won't remember
any of this and the list will be boring again. BTW, Never send oldfarts
links to sites you think may interest them that contain words like
VINTAGE.
We are the original hackers!
Mike Smith
[EMAIL PROTECTED
, and undoing them.
Have pity on your successors.
Mike Smith
[EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
What part of searching a path for a matching file is black magic?
Shells have been doing this for decades...
%%%
/*
* Load /boot/kernel/procfs.ko
* XXX: why does this work?
*/
chdir(/);
kldload(procfs);
You should only need the last kldload call. Any other magic is probably
Terry Lambert [EMAIL PROTECTED] writes:
There is so much goo around the module loading these days; there
are incursions into mount and all sorts of other programs that
should not know about module loading.
The kldload(2) interface alone is enough to make me cringe. The way
in which it
I probably just looked in the wrong place. Array rebuild cannot be done
in FreeBSD yet? Or am I wrong about this again? It would be nice to have
this info added to ata(4) I think. Didn't find anything about RAID
there. If a disk fails, is a message logged? Is there any way to view
the status
I've done most of the gruntwork of making AIO a loadable system.
I'd appreciate some feedback and testing, especially since I know
of no programs which use AIO.
Where's the demand-load of the aio module? Are you going to trap ENOSYS
in the libc side of things?
Please, please, please
Has anyone considered adding Pentium 2/3/4/Athlon/Athlon XP support
to the low level string/bytecopy routines? If we just supported SSE (1)
that'd get us (okay, me) a pretty nice performance boost on the P2, P3,
P4 and Athlon XP, all in one hit. These days, the average new computer
I have a Dell Poweredge 2550 with on board SCSI 7899 controller.
That controller is not actually fitted to any drives, only the
tape unit (DDS3).
I actually have a Mylex Acceleraid170 board attached to the four
Scsi drives creating a single drive volume. All seems to work fine
but the
JFWIW, you can build fsx with minimal or no changes on Windows with David
Korn's UWIN kit. All of the other posix-y kits have internal problems
that will cause spurious failures.
If you want to use Windows boxes as test clients (probably a good idea)
this is fairly important...
I gave
On Thu, Dec 13, 2001 at 08:35:08PM -0800, Mike Smith wrote:
Well, eCos is free-as-in-beer.
From a quick glance at the license it looks to be a lot more like
the GPL than a BSD license but I'll dig deeper.
'Any Modification which You create or to which You contribute must
Does anyone know if there is a significant performance gain to maintaining
scatter/gather lists with uio rather than writing the s/g routine yourself.
To be more specific:
I currently have a routine that takes a memory segments and places them in a
contiguous memory buffer. We need to
Hard. Lots of stuff relies on mmap, which basiclly requires an MMU or
other tricks. The other tricks can be somewhat expensive...
That's sort of what I figured. WRS and possible others use BSD
as a basis for their embedded OS on archs without an MMU, right?
No. Many embedded
If you want an OS, consider eCOS or one of the real embedded systems,
don't make too much work for yourself trying to take a system that's
entirely unsuited to the task and butchering it...
In eCOS are you talking about this one?
http://sources.redhat.com/ecos/
Yes.
I'd
I never understood why the PCI-bus is not a interupt vector design.
The way the story goes, it wasn't until the last minute that the initial
PCI working group realised that they'd forgotten to do something about
interrupts, and so the gross hack that we're now stuck with was
implemented.
It
I should point out that FSX can be used against any filesystem, and
that there are reports locally (at Apple) that it's great for killing
FreeBSD machines. I wasn't able to reproduce this when I tried, but I
may not have let it run long enough.
Oooh. Very cool! I'll start messing with
hi,
in trying to make a klm out of the meteor driver, and in
the process im stumbling onto some problems - major understatment :-)
for starters, how can i get an unshared irq? at the moment the irq is
shared among the video, scsi and the meteor.
im using
A PCI slot has 4 irq lines named INTA to INTD.
On PCs all slots share the same 4 physical irqs and the lines are
hardwired on the board in alternating order to each slot.
Not necessarily. This is the classic barber-pole or PCI swizzle;
it's not mandated for motherboard designs though (but it
Sorry, unbelievably bad at explaining myself. Per-open data is what i
meant. The reason I'm interested is it would make a full nvidia driver
port quite a bit easier.
Sorry, I know of no current plans which adress this.
The issue is non-trivial to fix because we currently don't
Sorry, unbelievably bad at explaining myself. Per-open data is what i
meant. The reason I'm interested is it would make a full nvidia driver
port quite a bit easier.
Sorry, I know of no current plans which adress this.
The issue is non-trivial to fix because we currently don't
The chip driver is entierly bogus and only claims things that nothing
else does. This generally means that your probe code isn't working
correctly because if it were your driver would have claimed this device.
As I am installing my driver after boot time (using KLD), the chip
driver
It's still there. See the code in /sys/boot, in particular,
the bios code in boot2 and boot 3.
You're thinking of the bioscall interface in BTX, I guess, which uses
v86 mode, not real mode.
As John said, actually, really going back to real mode is hard. It would
be easier to just reboot
This makes me believe the devclass_t structure defined in a driver is
never used. Is there another code path I'm missing?
Yes; there is a twisty maze of macros which ultimately results in the
driver_module_data structure ending up in a linker set. The devclass
structure is, as Warner
are when the system freezes. Is there any way
to kick the system in the butt somehow to at least let the debugger do
it's thing?
On the same line, can anyone tell me where I can find out everything a
program does when executed BEFORE the first line of code is reached
(gcc).
not THE Mike Smith
What happend at Intel? Their driver is even released under the
BSD license! (and the Linux one under the GPL)
Many Intel software products are released under a BSD-like license.
Consider the ACPI CA codebase we use.
The driver will be committed to -CURRENT first and MFC'ed to
-STABLE
The Intel driver will be the preferred driver for these cards.
That still is under discussion.
More to the point, let the respective drivers stand on their own merits.
There is no need to decide for one or the other.
This is, unfortunately, not entirely true. One of them is going
Take it easy on the poor guy! There have been many an occasion when all
I could say was help...actually help m plse. Of
course never when dealing with FreeBSD!
(not the) Mike Smith
Geoff Mohler wrote:
Uhh..with what?
On Tue, 27 Nov 2001, Andrey Pugachev wrote
G'day.
I'm sorry to say that 4.4 has a regression related to systems with 4GB
of memory, mostly related to poor test coverage.
You can work around this by reducing the amount of physical memory that
the kernel will use by setting a loader tunable. At the 10-second countdown
before the
I changed the disklabels on a few SCSI disks and now
I keep getting these BTX halted messages every time
I reboot.
The most likely cause of this is that you're messing up the disks to the
point that your BIOS (probably your SCSI controller BIOS) is crashing when
it tries to read them.
I was hoping its easy enough to crack this myself (with some online
tips references) but John Baldwin convinced me otherwise :-)
The evidence suggests that my original analysis is correct:
Error message when the SCSI disk is attached to
the AIC-7896 SCSI BIOS v2.20s1B1
...
cs=c800
to be called during a program running, can
anyone give a clue as to what stupid programming error could cause this
to happen?
Thanks
Mike Smith (not THE Mike Smith)
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
is if anyone else has had problems using C++ (in
general) crashing the kernel during subsequent initialization of the
same program or specifically with AF_ISO family (-liso) sockets. Nothing
more than that.
Thanks again,
Mike Smith (again, not THE Mike Smith)
[EMAIL PROTECTED]
To Unsubscribe: send
, and may
in fact demonstrate never-really-worked-the-first-time syndrome.
Mike Smith (again, not THE Mike Smith)
You're no more or less Mike Smith than I am, dude. 8)
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
Don't do this; write your own ppbus driver and just claim the ppbus
directly when you are 'active'.
Hi hackers,
I have to hijack lpt_intr() from lpt device driver
(/sys/dev/ppbus/lpt.c) to measure latency.
Is this right =
BUS_TEARDOWN_INTR(old interrupt handler);
And why is that ? for RAID's with the same parameters, I'd expect them
to perform equal, maybe with a sligtht advantage to the ATA driver since it
saves a few cycles by being tighter integrated.
Agreed. I don't think the RAIDs will have the same parameters. Most
of these cheap
I've just updated the ACPI CA subsystem to the Intel 20011018 snapshot.
This primarily fixes a couple of bugs in the ACPI interpreter; see the
changelog at
http://developer.intel.com/technology/iapc/acpi/downloads/CHANGES.txt
for full details.
These changes may fix some situations where:
Well, the question is rather simple... i am running some experiments
on system with severe load on the PCI bus (basically a router with 4 interfac
es
trying to forward 2..4 streams of 64-byte packets at 100Mbit/s (i.e. 144kpps
on each stream), and from low level timing i notice that the
. Yea, I know it's ancient but we have added
significant kernel hacks to support specialized ATN and X.25 protocols
and don't envision upgrading until we get our modifications completed.
Any insight would be appreciated.
Mike Smith (but not THE Mike Smith)
To Unsubscribe: send mail to [EMAIL
Could you please tell me where in the source the KLD location is
determined? I found out that the KLD can be loaded under / in 4.3-Release
even if I have a newer copy under /modules. This gives me a lot of
trouble debugging.
It shouldn't give you a lot of trouble because you should never
You're a) misrepresenting the project, b) dismissing the opinions and
statements of others that are arguably more in touch with the project,
and c) you won't let this stupid thread die.
I do not feel that this thread is stupid.
This much is obvious. You are, however, largely alone in this
Well, honestly, FreeBSD makes the life of the developers of third-party
binary-only drivers fairly difficult.
It does? On the whole, actually, I'd say we do a pretty good job of
making it easy.
The reason is that there
are a lot of API changes happening between the releases (take
Julian
Doug; I would recommend against falling for Ted's flamebait here, since
that's really all it is.
That's silly, what did you find in it that's flamebait? I think you didn't
read it.
You're a) misrepresenting the project, b) dismissing the opinions and
statements of others that are
Doug, in the entire history of the FreeBSD project, when given a choice
between a better driver or code that is closed source, and a worse
driver that has open source, the FreeBSD community has never chosen the
driver or code with closed source. In fact I can only remember ONCE
that
Is there any chance off implementing syntax like
kernel=${kernel:-/kernel}
which is obviously sh-compilant?
I don't much like either of these proposals.
My principal objection is that they're trying to solve the wrong problem.
The original poster is setting $kernel in the DHCP client
Mike Smith wrote:
Is there any chance off implementing syntax like
kernel=${kernel:-/kernel}
which is obviously sh-compilant?
I don't much like either of these proposals.
My principal objection is that they're trying to solve the wrong problem.
The original poster
What would be the best way to allocate:
1) a VM page whose physical address falls within a certain boundary, and
2) a VM object whose pages are contiguous in physical address space?
Background:
The !@*%^*!#^%*!#^$!@ Intel 810/815 graphics controller requires its
instruction and hardware
When fgrepping a huge file (say 10GB) for a non-existing string,
fgrep's memory size skyrockets. At a certain point in time its SIZE was 391M
(RSS was about 30MB) and the system got rather unreponsive. The
string was about 12 bytes big, and we fail to see why grep would
need so much.
Is
I've been trying to find out some information on programming the
fb/vesa interface, eg. set_video_mode() and friends.
Try the Vesa 3.0 document (google is your friend).
From the few examples I've seen, it appears that you have to
muck about with banks rather than use a pointer to linear
Hello,
| In short, which program gives enough knowledge to the microprocessor (?)
| and allow him to use kern.flp mfsroot.flp in order to boot and make the
| operating system running.
your BIOS reads the first sektor from your floppy which consists
of a boot loader, which usually loads
On Thu, Sep 13, 2001 at 04:09:57PM -0700, Mike Smith wrote:
Is there a single blessed way to define packed structures
for use in drivers? I suspect that using #pragma pack(1)
will lead to alignment errors in non-Intel architectures.
Any form of packing is going to cause
A quote from the end of the boot_i386.8 manual page..
IMPORTANT NOTE: Because of limitations imposed by the conventional disk
interface provided by the BIOS, all boot-related files and structures
(including the kernel) that need to be accessed during the boot phase
must reside on
So.. if I read you right, booting correctly for 1024 cylinders works
if boot0 knows about it. Isn't boot0 the one in the MBR, not in the fbsd
slice? Does this mean that boot1 and boot2 should work just fine if they
are loaded by another kind of MBR loader (say, Grub), and they find out
In message [EMAIL PROTECTED] Mike Smith writes:
: Any form of packing is going to cause problems for items that are
: located in illegal fashions.
It is almost more appropriate to use bus_space_{read,write} than using
memory mapped structures. The bus macros will work for otherwise
Hi there,
Is there a single blessed way to define packed structures
for use in drivers? I suspect that using #pragma pack(1)
will lead to alignment errors in non-Intel architectures.
Any form of packing is going to cause problems for items that are
located in illegal fashions.
Having
Also, which routing specifically implements the probe calls to drivers?
device_probe_and_attach
Another option is to probe the wired device first explicitly, and then skip
it in the normal probe scan. In linux there is a clearly defined routine
that does this, but i havent found it
I'd be OK with this being done as a hack for now. I think the bridge
code needs to be a bit kinder about allowing stupid things to be done
if they're set up by the BIOS.
I'd like to propose committing the following change which adds a new
undocumented option in the spirit of
On Sat, 25 Aug 2001, Mike Smith wrote:
I/O space is easy, but memory space is hard. Userspace access to
physical memory is a big no-no in the *nix world.
I want to disagree just a bit. If you look at myrinet, or the many fpga
cards, it's the standard modus operandi. You have to do
The memory is not freed until you unmount (and then, the memory is
only free'd for use by other cfs mounts -- the process size does not, of
course, shrink).
It doesn't? Does it just use malloc for these structs? When you say of
course, you kinda imply you're thinking of the bad old
This is *totally* not how to do interrupts with Mach. You should be
creating a thread for each interrupt source, and keeping a worklist of
interrupt handlers registered against that source.
Unfortunately, I can't help you with Mach-related things, since I work
for Apple and that'd put me in
, it could have been intended to be += and you've found a
heretofore undiscovered bug! All you have to do is press Shift at the
wrong time (not that I've ever done that).
Mike Smith
(but not THE Mike Smith)
Rob wrote:
My first post on hackers, so please don't flame me too bad :) I think
I really can't buy the idea that interrupt threads are a good
idea for anything that can flood your bus or interrupt bandwidth,
or have tiny/non-existant FIFOs, relative to the speeds they are
being pushed; right now that means might be OK for disks; not OK
for really fast network
I'm developing some code running in kernel that use a lot of stack. And it
seems i run into stack overflow. This results in some proc structure
related parts overwrite (particulary p-p_stats-p_timer[ITIMER_PROF]) and
unexpected signals. (Otherwise, it usually page faults inside
This is system-specific. Typically, systems only clear memory on
cold-boot, but the behaviour is not standardised.
As far as I understand, this feature works only if the machine does not
clear its memory upon reboot. AT compatibles clear memory during the
BIOS POST, thus, we don't see
It is important for you to send plain-text messages to public lists.
In general a address in a process is just a linear address which refer to
physical address indirectly by page directory. This is reasonable in
user space. However is it necessary to do such thing in kernel? It is sure
to
It also has the unfortunate property of locking us into virtual
wire mode, when in fact Microsoft demonstrated that wiring down
interrupts to particular CPUs was good practice, in terms of
assuring best performance. Specifically, running in virtual
wire mode means that all your CPUs get hit
I tried your suggestion below, and for some reason its still assigning the
same interrupt (whichever one I pick) to both the network card and the
wavelan card, and interstingly enough even if I remove one of them, its
still trying to get a routeable interrupt and the wavelan still doesnt
In message [EMAIL PROTECTED] lists w
rites:
: I tried your suggestion below, and for some reason its still assigning the
: same interrupt (whichever one I pick) to both the network card and the
: wavelan card, and interstingly enough even if I remove one of them, its
: still trying to get a
Hello,
I have a question. I'm trying to make a module for a PCI card. My
problem is my 'detach' function never get's called when I unload the
module. Al my other functions get called correctly. Here is my source:
Detach != unload.
The detach method is called by your parent to kick you
Hi Mike, ok my pci-pcmcia bridge is in slot 0, my network card is in slot
3, below are the dmesg outputs from both oldcard and newcard,
Ok; this is different from the linked dmesg you were showing before,
and what it's highlighting is the weakness in the algorithm that we use
for picking an
Yes, I understand that. I'm just trying to find out why Mike keeps
saying we cannot determine the processor cache characteristics at
runtime.
Because I believed we couldn't. It appears I'm wrong. 8)
The only question left really then is whether it's worth actually trying
to tune for cache
, Mike Smith wrote:
Tried the patch, interesting thing, for some reason or other its always
routing the IRQ to the same IRQ as the realtek network card I have in
here, and with the patch in (before nothing worked at all on the pccbb),
now if the network card is in slot0 it doesnt work
The costs involved in doing DMA to/from the memory region
above 4G will be incredible, unless the address space is
both exported, and known, to the PCI bus; even then, it
could only work for 64 bit cards, since 32 bith cards will
only be able to address the first 4G of physical memory.
No
The space is linear in physical space and if you have PCI/64
capable devices they can access it all too.
(In fact 64 bit addresses have been supported even in 32 bit wide PCI
since day 1).
OK, then what was that whole paging thing everyone was talking about, I
thought that
Tried the patch, interesting thing, for some reason or other its always
routing the IRQ to the same IRQ as the realtek network card I have in
here, and with the patch in (before nothing worked at all on the pccbb),
now if the network card is in slot0 it doesnt work, and the wavelan does,
if
If I added this to a man page would I be telling the truth :).
Note, these are my notes and not the exact text that I would
add, and I have not bother with anything to do with object
coloring etc. I just want to make sure I've got this part
down.
It looks about right, but page colouring
You should format your messages in ascii to send to this list.
I know PIII can support 64G physical memory. In FreeBSD how can I visit
such range memory(4G-64G) ?
You can't. Those memory ranges are strictly off-limits to non-US
citizens.
--
... every activity meets with opposition,
I was wondering whether someone could shed some light on this for me: I've i
nstalled FreeBSD 4.3, Debian Linux 2.2r2 and windows 98 on my laptop. Everyt
hing is fine except that after using FreeBSD if I try to go into windows, the
system locks up. If I turn the power off and back on
I want to use the function inet_aton() in the kernel code. However, I =
found no kernel equivalent of this function int the freebsd sources. I =
could find inet_ntoa(), but not inet_aton(). Is it named by some other =
name or how can I locate it?
If you are trying to parse an ascii internet
This hunk is needed for lint(1) to recognize special comments.
Don't remove it.
The '/*-' part? What does lint do special with those?
It's actually a signal to indent(1) to leave the comment's formatting
alone. See the manpage.
--
... every activity meets with opposition, everyone who
I cvs'ed the current version of btx by cvs co btx and tried to build it on
my FBSD-4.0 box and here is what I got:
...
bash-2.04$ as --version
GNU assembler 2.11
...
What should I do?
Uninstall your custom binutils:
ziplok:~uname -r
4.3-STABLE
ziplok:~as --version
GNU assembler
Dear Friends
I'm incorporating the Real Time Protocol RTP (rfc 1889) to
FreeBSD 4.0 kernel.
Months ago, I compiled successfully the RTP Library API developed
by Lucent into the FreeBSD kernel with the right logical and technical
adjustments for the BSD kernel of course (copyin,
Well, this BTX thing is amazing: all this effort, (btxld, run-time
library crt0.o, loader, etc.) seems to just to provide a 32-bit
protected and possibly paging-enabled environment to start the
kernel/loader(and to confuse a new-comer like me.) What are the
other gains? Where can I
Hello
I am experimenting with kernel modules and am trying to write to a file.
This is the syscall function (sorry of my terminology is messed up)
static int write_file(struct proc *p, void *arg) {
struct write_args *wstructure;
struct open_args *ostructure;
I call this function with (curproc, PATH_MAX+1), and everything is fine
when I have just a few local variables defined in the caller (it all
works on MOD_LOAD only). However, if I have 2 buffers, 4096 bytes each,
as local variables and then try to allocate userspace memory the same
Make sense. But there are other things in the UPAGES.
Yes; in reality you have about 7k.
It's plenty of space for a deep call stack, just not for large locals.
--
... every activity meets with opposition, everyone who acts has his
rivals and unfortunately opponents also. But not because
3) Steal an idea from Linux (gasp!), and have module dependencies. ie,
load ipfw.ko and then before we load up natd, we check to see if
ipdivert.ko is loaded and load it. Alternatively, loading ipdivert.ko
(before loading ipfw.ko), will automagically load ipfw.ko since ipfw is
This sounds great; nothing else I can think of. Compaq are happy about
this. 8)
I need to MFC changes in ida driver, which start backround
firmware processing on Integrated SmartArray controllers
(this allows automatic on-line rebuild of failed drives).
I am going to do it in next few
I need to MFC changes in ida driver, which start backround
firmware processing on Integrated SmartArray controllers
(this allows automatic on-line rebuild of failed drives).
I am going to do it in next few days. I understood that I shall
avoid all changes for interrupt-entropy
hi,
In FreeBSD i want to compile the module dev which should reflect the
kernel also, without compiling the kenel completely.
You can't do this.
and also i want to
include one more directory in /usr/src/sys/dev . so to compile that
directory files shall i have to made changes in
We support ATAPI devices and has been for a long time (also CD burners)...
I believe I forgot to do a group reply on my previous reply
to Søren.
OK, it seems a misunderstanding of the term ATAPI.
The author of cdrecord, Joerg Schilling, told me - I will translate:
Citation:
You
In fact there are other considerations as well, namely the power supply,
which is typically not all that beefy in 1U systems. For both of these
reasons and others, our current 1U offering, although using an MP motherboard
,
can only be ordered with one CPU. We've recently improved the
In the default case, it should attempt to obtain a DHCP lease,
and, failing that, ask the user to give it settings, or let
them do IPv4 stateless autoconfiguration. Ad Hoc networking
should always just work.
If anyone is taking a vote, I disagree. I do not want any system
ever
For those of you that have too much time on your hands.
http://people.freebsd.org/~msmith/subr_figlet.diff
You'll need to load a font before booting;
load -t figlet_font somefile.flf
Now, for the complication; it's wrapping lines at a single character. If
someone wants to work out why,
1 - 100 of 779 matches
Mail list logo
