Re: fxp driver performance expectations
--- Danial Thom <[EMAIL PROTECTED]> wrote: > You couldn't do 100Mb/s with em on a 100Mb/s line > with min packets, because there are gaps between > packets so its impossible. Thanks for the detailed reply Danial. By 100Mbps I mean line-rate: 148809 packets/sec for 64-byte Ethernet frames + IPG and preamble appended. The 10/100 fxp NIC is on straight PCI-33. The 1000 em NIC is on PCI-Express x1. It can do 100Mbps line-rate (148809pps) in 100Mbps mode. In 1000Mbps mode it can do ~700Kpps, so the bottleneck isn't the FreeBSD IP stack. > Realize that fxp parts are only 32bit/33mhz so > the bus is a factor. Although its a 1Gb/s bus, > thats only when bursting, so its really > substantially less. With shorter packets you > have more setups and I/O and therefore more > overhead on the bus. Yes indeed. > fxp performs similarly to an em controller when > they are both on a 32bit/33mhz bus in Freebsd 4.x. > 5.x is about 20% slower than 4.x, but I expect the > drivers to be about the same for 5.x as well. Thanks for that. I realize that comparing a PCI-33 NIC to a PCI-Express NIC isn't fair. I don't have a PCI-33 Gig NIC - why I need outside info. > Are you using a traffic generator, or are you > relying on some server to return packets? Ixia traffic generator. > We have customers with fxp interfaces on freebsd > 4.x pushing 90Mb/s+ (while doing a lot of other > processing also), so its certainly possible. Yes, I can get 100Mbps using larger frames on the fxp - but it's the performance limit on the smallest frames, with the largest per-frame overhead, that I'm trying to discover. Thanks again for the reply - I appreciate the info. Paul. __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
fxp driver performance expectations
Hello. I am running FreeBSD-5.4 on a 3GHz P4 with two Intel fxp NICs running IP forwarding with polling enabled. For larger packets (e.g. 700 bytes) I am getting 100Mbps throughput port to port. For min-size packets (64 bytes) I am only seeing around 60Mbps. Increasing HZ and the polling parameters does not help. I tried a couple of Gigabit Ethernet cards, using the em driver and they can sustain 100Mbps. For further comparison I tried a recent Linux install and got a little over 20Mbps for 64 byte packets. So much for that. Does anyone know if I should be expecting more performance out of the fxp? Again, the em NICs work beyond 100Mbps with min sized frames. I want to know if there's more performance to be had from the hardware. My working assumption is that the driver is mature, it's a popular NIC, so maybe I'm getting all there is out of it. However, if the driver has been written optimized for larger packets at the expense of smaller packets, then perhaps there's work I can do. The goal, by the way, is to forward min-sized frames at line rate. It's a dumb goal, but one that marketing people care about. I clearly have no clue here, so please feel free to give me one. Thanks, Paul. __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Question regarding bus_dma_tag_create() .
Hi.
The man page for bus_dma_map_create() says this about
the "nsegments" parameter:
Number of discontinuities (scatter/gather seg-
ments) allowed in a DMA mapped region. If there
is no restriction, BUS_SPACE_UNRESTRICTED may be
specified.
BUS_SPACE_UNRESTRICTED is #defined as (~0).
Then, in busdma_map_create() and busdma_mem_alloc():
if (dmat->segments == NULL) {
dmat->segments = (bus_dma_segment_t *)malloc(
sizeof(bus_dma_segment_t) * dmat->nsegments,
M_DEVBUF,
M_NOWAIT);
...
}
I don't understand how this works when
BUS_SPACE_UNRESTRICTED is specified. The malloc will
be fed a -8 (or -12 with PAE) on i386.
I'm reviewing a driver at work that specifies
nsegments = 0, which I think is invalid, but in
reading the man page I came across this ~0 option,
which I've also seen in other FreeBSD drivers and I
just don't get it.
I hope someone can clue me in.
Thanks,
Paul.
__
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Does FreeBSD support sparse kernel crash dumps?
--- Peter Wemm <[EMAIL PROTECTED]> wrote: > On Friday 07 April 2006 12:47 pm, Paul Marciano > wrote: > > Hello. I read a while back about someone working > on > > supporting sparse kernel crash dumps (dumping only > the > > active kernel pages to the dump device as opposed > to > > all physical memory - for machines where the phys > mem > > is greater than the dump dev space.) > > > > Does anyone know the status of that project? Was > it > > committed, or are there plans to commit it? > > I have a working prototype as of last night. There > should be something > committable in the next week or two. > > When I boot my 2GB machine and force a crash dump > from single user mode, the > fully debuggable vmcore file is in the 40-50MB > range. A busy machine with > 12GB ram took about 150MB to dump. > > There are still some things to work out. It was > written for the amd64 kernel, > but can be ported to i386. > > -Peter > That's very timely news Peter. Do you think your code is easily back-portable to 5.4? Are the changes limited to dump_machdep.c or otherwise not dependent on a great deal of -current updates? Do you compress the data stream at all (e.g. gzip)? I have a specific need on a CompactFlash based system. I have a 256MB IDE mode card on a machine with 512MB physical memory. I can probably commit 128MB of the card as a dumpdev but I can't go beyond that. Good luck with it. Regards, Paul. __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Does FreeBSD support sparse kernel crash dumps?
Hello. I read a while back about someone working on supporting sparse kernel crash dumps (dumping only the active kernel pages to the dump device as opposed to all physical memory - for machines where the phys mem is greater than the dump dev space.) Does anyone know the status of that project? Was it committed, or are there plans to commit it? Thanks, Paul. __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Getting EBUSY upon unmount ATA disk...
I said: > # mount /dev/ad0s1a /flash > # : > /flash/foobar > # umount /flash > umount: unmount of /flash failed: Device busy Then I said: > No processes are camping on the mount point. I didn't realize /bin/sh doesn't close the fd. Man, I feel so dumb. Paul. __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Getting EBUSY upon unmount ATA disk...
(originally sent to freebsd-fs, but that looks like a developer list). I am running FreeBSD-5.4 with root on md0. I am seeing the following problem: # df Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/md0 17370 15810 17299%/ devfs 1 1 0 100%/dev /dev/md1369416 3384 0%/var /dev/md2 31470 8 28946 0%/tmp /dev/md3.uzipc 20808 18890 25499%/usr # mount /dev/ad0s1a /flash # : > /flash/foobar # umount /flash umount: unmount of /flash failed: Device busy Waiting doesn't help. I can repeat the unmount command after a time and it still reports EBUSY. ad0 is a SanDisk Compact Flash card. So far I've traced the unmount op to ffs_flushfiles() in ffs_vfsops.c: /* * Flush all the files. */ if ((error = vflush(mp, 0, flags, td)) != 0) return (error); Still trying to track it down, but I'm not an fs guy and so I'm asking for a little help. Some more background: My system uses a Compact Flash card to store configuration information. It's seldom accessed and it would be nice to support controlled removing/replacing during operation. It looks like it is possible using atacontrol to detach the channel and then reattach when the card is inserted. I can: insert, attach, mount, read, umount, remove all day long without problems. But now as soon as I do one write the drive doesn't want to unmount. No processes are camping on the mount point. manually typing "sync" doesn't help. unmount -f succeeds. I hope someone can advise me. Thanks, Paul. __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Looking for a USB Compact Flash card reader.
Hi, On 5.4-RELEASE I'm having problems with my AVB card reader: I often get "Synchronize Cache" and "CSW Tag" errors. Does anyone have a solid recommendation for a USB Compact Flash card reader? Thanks, Paul. __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Is remote gdb debugging supported on FreeBSD-5.4?
I have a need to remotely debug an application on a remote FreeBSD-5.4 machine that is devoid of utilities and source code. I previously asked a question about gdbserver, but it looks like it's not available so I'll ask the more general question. I could login in to the remote machine, mount a source and utility NFS partition and run gdb locally, but I would rather attach a gdb session from my development machine to the target machine. Is remote debugging with gdb supported at all on FreeBSD (5.4)? Thanks, Paul. __ Yahoo! DSL Something to write home about. Just $16.99/mo. or less. dsl.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: How do I get gdbserver for FreeBSD 5.4?
--- Lowell Gilbert <[EMAIL PROTECTED]> wrote: > Paul Marciano <[EMAIL PROTECTED]> writes: > > What I want to do is attach gdb from a development > > machine to a process running on an remote embedded > > target (stripped binaries, no source). I've only > > recently heard of gdbserver. I'm happy to use an > > alternative, if one exists. > > Oh; I thought you wanted to run gdbserver on > *FreeBSD*. ACK! Wires crossing! Both the target and the client are FreeBSD-5.4, but the target has limited flash and no local utilities so I can't debug on it natively (I'd rather not mount NFS utility partitions, but that's an option if there's nothing more elegant). So I do need a FreeBSD target gdbserver-like thing! Paul. __ Yahoo! DSL Something to write home about. Just $16.99/mo. or less. dsl.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: How do I get gdbserver for FreeBSD 5.4?
--- Lowell Gilbert <[EMAIL PROTECTED]> > gdbserver isn't currently ported for FreeBSD. If > you really need it (as opposed to some other form of > remote debugging), you could probably best get started > by asking obrien for hints. Thanks for the info Lowell. What do you mean by "some other form of remove debugging"? What I want to do is attach gdb from a development machine to a process running on an remote embedded target (stripped binaries, no source). I've only recently heard of gdbserver. I'm happy to use an alternative, if one exists. Thanks, Paul. __ Yahoo! for Good - Make a difference this year. http://brand.yahoo.com/cybergivingweek2005/ ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
How do I get gdbserver for FreeBSD 5.4?
Hello, I hope someone can help me. I am using FreeBSD 5.4-RELEASE and would like to use gdbserver, but I can't find it. The source tree has gdbserver in two places: /usr/src/contrib/gdb/gdb/gdbserver and /usr/src/gnu/usr.bin/binutils/gdbserver There's also a gdb6 in /usr/ports/devel. The installed gdb says its version 6.1.1, and it looks like the /usr/src/contrib/gdb tree is also 6.1.1, but there's no configure script or Makefile. Is this tree a patch for the usr/ports/devel/gdb6? I'm a bit lost right now. I hope someone can help me find my way. Thanks, Paul. __ Yahoo! for Good - Make a difference this year. http://brand.yahoo.com/cybergivingweek2005/ ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
5.4-RELEASE reset with /boot.config -m
I wonder if someone can help me. I want to run my 5.4-RELEASE system with a serial console and muted cons. My /boot.config is: -h -m The system loads the kernel but during startup the PC resets. No messages, just a reset. If I use -h by itself it is fine. If I use -m by itself it similarly resets. I must be doing something dumb. If you have any insights please let me know. Thanks, Paul. __ Yahoo! DSL Something to write home about. Just $16.99/mo. or less. dsl.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Will USB serial ever be fixed?
--- "M. Warner Losh" <[EMAIL PROTECTED]> wrote: > You asked for honesty, and I'll give you some more: > You are being a jerk and acting like you are > entitled to having your particular > problem fixed without pitching in and helping in > some way. You're funny... but actually I was only asking for an assessment... asking if it was going to be fixed. If not, that's my problem. No demands explicit or implied. I'm neither demanding nor expecting anything except the truth. For you, the truth is that it isn't a problem you're interested in (besides calling me names, which let's face it, is fun). That's not intended as a slant on you, mearly an interpretation of your position. That there's an open bug says it hasn't been a problem on the top of anyone elses list. If the final truth is that there are bigger fish to fry and so it's going to be a while before the heroic USB developers get around to this, then that's fine and it's something I'll have to deal with. However, if someone is a week away from fixing it, then that would be useful to know. Thanks, Paul. __ Click here to donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Will USB serial ever be fixed?
--- Frank Mayhar <[EMAIL PROTECTED]> wrote: > Well, I've got a handful of such devices that appear > to work just fine. > Now, mine are based on the Prolific PL-2303* chips > but you don't mention > any specific hardware. All I know is that I'm using > them for consoles > (and am running getty on them) with no hangs or > crashes of any kind. Thanks Frank. I'm using a converter based on the FTDI232BM, on FreeBSD 5.3-RELEASE. ucom0: FTDI USB to Serial Cable, rev 1.10/4.00, addr 2 I am seeing the ucom0 work with normal input/output, but not with getty. With getting it prompts for the login then, upon hitting return, freezes. Without getty I can send and receive characters all day long without problems. Once the freeze happens, I can stty -f ucom0 -hup and data starts again. Referencing the following: http://lists.freebsd.org/pipermail/freebsd-questions/2004-June/049433.html http://www.freebsd.org/cgi/query-pr.cgi?pr=65769 --- Warner Losh <[EMAIL PROTECTED]> wrote: > Personal opinion: You overstate the probelems, and > have a bad attitude. Perhaps, at least on one count. > I use umodem + ucom all the time for connecting to > the internet on my laptop. I've had exactly 0 > problems doing this for the past 4 years. I'm happy for you. I've been using FreeBSD since it was called 386BSD + patches and there are many things I've never had problems with too. It seems we're both happy people... mostly. > There may be other problems in the serial subsystem > intersected with usb [...] That's precisely the area I'm interested in. > Even to the point that it will be hard for you to > effect change because people don't take you > seriously. I don't want to effect change, I just want a working system. If you're serious about supporting an operating system you have to support all your uses, even the assholes. Thanks, Paul. __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Will USB serial ever be fixed?
I have a general question on the status of USB to serial support in FreeBSD (specifically the ucom and ftdi drivers). I need an extra serial port on my system and for reasons outside the scope of this email I can't just plug in a PCI card. I'm trying to use USB, but it seems that ucom is broken and no one is fixing it. A tcflush() bug posted last year reports that USB serial ports hang on a call to tcflush(). This can be observed by using getty on one. Another poster reported problems running PPP (including a kernel crash). Given the lack of movement in these bugs, it looks like the FreeBSD developers don't care about ucom... certainly not enough to fix it. Now I could get up on a soapbox and say how unacceptable it is for a fundamental peripheral like serial ports (even though they're over USB) to be so obviously broken for so long... but I won't, because I'm not a developer and I understand that it's a volunteer thing and overall FreeBSD is otherwise excellent. So without wanting to offend (whilst secretly being pretty frustrated with the corner I find myself in) I would like to know, hopefully from someone responsible for the subsystem, if ucom/ftdi is likely to be fixed in the next six months. Sigh. I really don't want to have to move to Linux for something this small, but USB serial is a system component I can't change. Be honest. Be brutal. I need to know where FreeBSD stands on this. Thanks, Paul. __ Click here to donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
