Re: setsockopt 228K fails with ERR#55 'No buffer space available'
In the last episode (Sep 11), Steven Hartland said: > The following fails with ERR#55 'No buffer space available' > int buf_size = (228*1024); > int ret = setsockopt( s, SOL_SOCKET, SO_SNDBUF, (void*) &buf_size, sizeof( > buf_size ) ); > > This appears like its controlled by kern.ipc.maxsockbuf which is > default of 256k so the above should work if its per socket and not > total? I think you want net.inet.tcp.sendspace; kern.ipc.maxsockbuf seems to be unused based on a quick grep of the source. Also check the sockbufsize rlimit, although I think that defaults to unlimited. -- Dan Nelson [EMAIL PROTECTED] ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
setsockopt 228K fails with ERR#55 'No buffer space available'
The following fails with ERR#55 'No buffer space available' int buf_size = (228*1024); int ret = setsockopt( s, SOL_SOCKET, SO_SNDBUF, (void*) &buf_size, sizeof( buf_size ) ); This appears like its controlled by kern.ipc.maxsockbuf which is default of 256k so the above should work if its per socket and not total? Does anyone know the implications of increasing this to say 512k and is it per socket or global? Does anyone have any experience with apps setting SNDBUF to large values like 256k? Steve / K This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it. In the event of misdirection, illegible or incomplete transmission please telephone (023) 8024 3137 or return the E.mail to [EMAIL PROTECTED] ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re[4]: Forcing boot to seek for files on other parition then `a'
Re Mike, Sunday, September 11, 2005, 12:56:18 AM, you wrote: > Daniel Gerzo <[EMAIL PROTECTED]> typed: >> Ahoj omyl Mike, >> >> Sunday, September 11, 2005, 12:19:20 AM, si odoslal: >> >> > Daniel Gerzo <[EMAIL PROTECTED]> typed: >> >> How to force boot procedure to automatically seek boot images on >> >> other partition then `a', let's say on `e' one. >> >> > If you google the archives at FreeBSD, you'll find that this is a >> > PITA, assuming it's possible at all. No particular reason for it, just >> > not something anyone has ever wanted to do. Someone may have fixed >> > this since the last time I saw it discussed, but I wouldn't bet on it. >> >> > So the fix is to boot on the fixit CDROM, fire up bsdlabel on your >> > disk, and change the "e" to an "a", and boot off the a partition. >> >> oh, you know that dedicated boxes doesn't have CD-ROM and FreeBSD ISO >> in it, and I don't have a roam KVM available? Also, there's no netboot >> server available :/ > Sorry about that - I didn't read your message carefully enough. >> So maybe I could hack sys/boot/i386/boot2/boot2.c somehow?, but I'm not >> a C knowledgeable and it's little bit mess for me. > Well, if you can get the disk someplace where you could install a > recompiled boot loader, then you can use bsdlabel on it. Well, I actually can install a recombiled boot loader, can't I? I have sources and I can use bsdlabel -B ad0s1 then. but now the question is: which changes do I need to proceed so that the boot(8) will seek for boot images in `e' partition? >
Re: Re[2]: Forcing boot to seek for files on other parition then `a'
In <[EMAIL PROTECTED]>, Daniel Gerzo <[EMAIL PROTECTED]> typed: > Ahoj omyl Mike, > > Sunday, September 11, 2005, 12:19:20 AM, si odoslal: > > > Daniel Gerzo <[EMAIL PROTECTED]> typed: > >> How to force boot procedure to automatically seek boot images on > >> other partition then `a', let's say on `e' one. > > > If you google the archives at FreeBSD, you'll find that this is a > > PITA, assuming it's possible at all. No particular reason for it, just > > not something anyone has ever wanted to do. Someone may have fixed > > this since the last time I saw it discussed, but I wouldn't bet on it. > > > So the fix is to boot on the fixit CDROM, fire up bsdlabel on your > > disk, and change the "e" to an "a", and boot off the a partition. > > oh, you know that dedicated boxes doesn't have CD-ROM and FreeBSD ISO > in it, and I don't have a roam KVM available? Also, there's no netboot > server available :/ Sorry about that - I didn't read your message carefully enough. > So maybe I could hack sys/boot/i386/boot2/boot2.c somehow?, but I'm not > a C knowledgeable and it's little bit mess for me. Well, if you can get the disk someplace where you could install a recompiled boot loader, then you can use bsdlabel on it. http://www.mired.org/consulting.html Independent Network/Unix/Perforce consultant, email for more information. ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re[2]: Forcing boot to seek for files on other parition then `a'
Ahoj omyl Mike, Sunday, September 11, 2005, 12:19:20 AM, si odoslal: > Daniel Gerzo <[EMAIL PROTECTED]> typed: >> How to force boot procedure to automatically seek boot images on >> other partition then `a', let's say on `e' one. > If you google the archives at FreeBSD, you'll find that this is a > PITA, assuming it's possible at all. No particular reason for it, just > not something anyone has ever wanted to do. Someone may have fixed > this since the last time I saw it discussed, but I wouldn't bet on it. > So the fix is to boot on the fixit CDROM, fire up bsdlabel on your > disk, and change the "e" to an "a", and boot off the a partition. oh, you know that dedicated boxes doesn't have CD-ROM and FreeBSD ISO in it, and I don't have a roam KVM available? Also, there's no netboot server available :/ So maybe I could hack sys/boot/i386/boot2/boot2.c somehow?, but I'm not a C knowledgeable and it's little bit mess for me. >
Re: Forcing boot to seek for files on other parition then `a'
In <[EMAIL PROTECTED]>, Daniel Gerzo <[EMAIL PROTECTED]> typed: > How to force boot procedure to automatically seek boot images on > other partition then `a', let's say on `e' one. If you google the archives at FreeBSD, you'll find that this is a PITA, assuming it's possible at all. No particular reason for it, just not something anyone has ever wanted to do. Someone may have fixed this since the last time I saw it discussed, but I wouldn't bet on it. So the fix is to boot on the fixit CDROM, fire up bsdlabel on your disk, and change the "e" to an "a", and boot off the a partition. http://www.mired.org/consulting.html Independent Network/Unix/Perforce consultant, email for more information. ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Forcing boot to seek for files on other parition then `a'
Hello hackers, I'm in the step when I need your help. Let me describe the situation. I've ordered the dedicated server where was installed FreeBSD by default. The main problem was that they created only 4gB swap and one partition mounted under '/'. Well, of course, I didn't liked it, and I decided to repartition disk, so I've swapoff the swap, created there the filesystem, exctracted base from ISO and booted to that partition. Then I've removed the `a' partition and created few more partitions from the free space. So I had a potentional root partition under ad0s1e, the tmp partition under ad0s1f, swap under ad0s1g and rest under ad0s1h as usr partition. I've mounted them all and installed there a base system again and booted the new system, newfs'ed the ad0s1d (there was the original swap, then system) and prepared it as a var partition. Now, what I have: ad0s1e 248mb / ad0s1f 1.5gb /tmp ad0s1g 1.5gb swap ad0s1d 4gb /var ad0s1h rest/usr Ok, I will describe the main problem here; According to the boot(8) manual the automatic boot will attempt to load /boot/loader from partition `a'. But there's no such partition and I have /boot/loader under the `e' partition. I know only one and the only normal way how to force to use `e' partition. This is done throught the boot2 stage of boot process, when one need to manually type 0:ad(0,e)/boot/loader to the console so boot process will be able to use needed utilities. The main question is: How to force boot procedure to automatically seek boot images on other partition then `a', let's say on `e' one. -- Best Regards, Daniel Gerzo
Re: status of ufsj and gjournal
On 9/9/05, Eric Anderson <[EMAIL PROTECTED]> wrote: > [EMAIL PROTECTED] wrote: > > Hi list, > > I wonder whats the status of those summer of code projects. > >>From gjournal we heard that it has been completed but then nothing happens, > >>any > > further information about this? > > Is somebody working on ufsj? Was the summer of code project successful? > > Scott Long is the core person working on ufsj, and I have seen some > stuff worked on in his perforce tree, but I think it's a ways away from > being beta. I'm sure he would welcome help. I was working on the ufsj stuff as a Google SoC project with Scott. It is very close to beta, however this past week involved school starting back up, so I have been unable to do any work at all. However, now that I am settled in at school (for some definition of settled in), I have time to devote to ufsj again. I hope to release a beta "real soon now", so stay tuned. Brian ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
clock software interrupt
Hi folks, I've seen clock software interrupt thread (referring to clk_ithd in kern/kern_intr.c); watching to manpages I read that priority is used as vector for that thread. My question is how can I call this software handler? There's no track in the IDT for the associated handler... (ia32, for better check see i386/i386/exception.s). Thanks, Mark Swetter ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: anoncvs.freebsd.org broken?
+-le 10/09/2005 02:15 -0500, R. Tyler Ballance écrivait : | I was trying to get the most recent RELENG_6 code from one of the AnonCVS | servers here in America, so I was using the USA mirrors, and they don't | seem to be properly configured anymore... | | | %sudo cvs co -rRELENG_6 -P src | cvs [checkout aborted]: cannot write /home/ncvs/CVSROOT/val-tags: | Permission denied | % | % echo $CVSROOT | [EMAIL PROTECTED]:/home/ncvs | % | | Keep in mind, anoncvs1.freebsd.org works perfectly fine. | | Ho hum, guess we only get one working USA mirror ;) use cvs -R co -rRELENG_6 -P src :-) -- Mathieu Arnold ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Software suspend on FreeBSD
On Fri, Sep 09, 2005 at 07:25:31PM -0400, Bill Vermillion wrote: > Earlier in the linear time track, on approximately Fri, Sep 09, > 2005 at 23:01 , [EMAIL PROTECTED] divulged this > public information: > > > > From: Bruno Ducrot <[EMAIL PROTECTED]> > > Subject: Re: Software suspend on FBSD. > > To: Pranav Peshwe <[EMAIL PROTECTED]> > > Cc: freebsd-hackers@freebsd.org > > Message-ID: <[EMAIL PROTECTED]> > > Content-Type: text/plain; charset=us-ascii > > > On Fri, Sep 09, 2005 at 07:52:46PM +0530, Pranav Peshwe wrote: > > > Hello, > > > Does FreeBSD have 'software suspend' like linux ? or > > > maybe something similar... > > > No. > > Yes. It depends upon which shell you are using. For shells > that support it you just suspend it with control-Z. > > A restart is issued with 'fg' - foreground. > > Unless of course the OP means something entirely different. Unless I mistaken of course, the question of the OP is: does FreeBSD support "software suspend" as for Linux, and this means for me: "does FreeBSD support a generic mechanism in under to suspend to disk without support from BIOS (or with very minimal help from the BIOS)". The answer to this question is no. Of course FreeBSD support suspend to disk via APM or via ACPI S4bios, but this is another story. Cheers, -- Bruno Ducrot -- Which is worse: ignorance or apathy? -- Don't know. Don't care. ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: FreeBSD 6.0-BETA4 panics when when configuring pfsync0
On Saturday 10 September 2005 00:00, Max Laier wrote: > On Friday 09 September 2005 19:19, Max Laier wrote: > > On Friday 09 September 2005 15:05, Dominic Marks wrote: > > > Dominic Marks wrote: > > > > Hello, > > > > > > > > I am setting up a redundant firewall setup for our company, the > > > > systems are identically configured Dell servers running FreeBSD > > > > 6.0-BETA4. Software being used is pf, carp, pfsync and altq. > > > > > > > > When I attempt to configure the pfsync0 interface the systems > > > > panic. The systems currently have HTT enabled, I've googled but > > > > I didn't find anything to suggest pfsync was not SMP/HTT > > > > friendly. Could this be a configuration problem? > > > > > > > > interface: > > > > > > > > # ifconfig em2 > > > > em2: flags=8843 mtu > > > > 1500 options=b > > > > inet 172.16.254.2 netmask 0xff00 broadcast > > > > 172.16.254.255 ether 00:04:23:bd:7a:ef > > > > media: Ethernet autoselect (100baseTX ) > > > > status: active > > > > > > > > command: > > > > > > > > # ifconfig pfsync0 syncdev em2 syncpeer 172.16.254.1 > > > > > > Following a little more testing: > > > > > > If I don't specify a syncpeer I don't get the panic. > > > > Thanks for the dump. Seems like we were leaking a lock from > > pfsyncioctl(). Can you please try this patch and report back. > > Thanks in advance. AWESOME!! I'll test this on Monday when I'm back at work and report on the outcome. > And the patch ... sorry. :) -- Dominic Marks ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
RE: "Smart" Hubs
On Friday, September 09, 2005 3:40 PM, Ryan P. Sommers <> unleashed the infinite monkeys and produced: > PS If anyone knows of a hub that's "easy" to find and still is an actuall > good 'ol hub, let me know. Not a hub, but a different solution - a network "tap". They're designed to do exactly what you're looking for - allow sniffing of traffic from a link. Most taps require you to sniff the traffic on 2 ports, one for each direction. However NetOptics (and probably others) do a range of taps that aggregate the traffic onto a single cable. -- Rob | Oh my God! They killed init! You bastards! ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"
anoncvs.freebsd.org broken?
I was trying to get the most recent RELENG_6 code from one of the AnonCVS servers here in America, so I was using the USA mirrors, and they don't seem to be properly configured anymore... %sudo cvs co -rRELENG_6 -P src cvs [checkout aborted]: cannot write /home/ncvs/CVSROOT/val-tags: Permission denied % % echo $CVSROOT [EMAIL PROTECTED]:/home/ncvs % Keep in mind, anoncvs1.freebsd.org works perfectly fine. Ho hum, guess we only get one working USA mirror ;) Cheers, -R. Tyler Ballance ___ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "[EMAIL PROTECTED]"