Re: Where to put new bus_dmamap_load_mbuf() code
Correction.
This sample:
if (bus_dma_tag_create(pci-parent_dmat, PAGE_SIZE, lim,
BUS_SPACE_MAXADDR, BUS_SPACE_MAXADDR, NULL, NULL, len, 1,
BUS_SPACE_MAXSIZE_32BIT, 0, pci-cntrol_dmat) != 0) {
isp_prt(isp, ISP_LOGERR,
cannot create a dma tag for control spaces);
free(isp-isp_xflist, M_DEVBUF);
free(pci-dmaps, M_DEVBUF);
return (1);
}
You'll need to change the number of segments to match the max
supported by the card (or the max you will ever need). This
example made me realize that the bounce code doesn't deal with
multiple segments being copied into a single page (i.e. tracking
and using remaining free space in a page already allocated for
bouncing for a single map). I'll have to break loose some time
to fix that.
--
Justin
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
Re: header polution
In message [EMAIL PROTECTED], David O'Brien writes: When compiling the `dict' port, one gets: In file included from /usr/include/machine/signal.h:54, from /usr/include/sys/signal.h:178, from /usr/include/signal.h:44, from dict.h:33, from clientparse.y:25: /usr/include/machine/trap.h:105: warning: `T_USER' redefined This seems very wrong. Can't we rename T_USER in the kernel to _T_USER, or wrap it in _KERNEL? The comment in machine/signal.h(x86) says: #include machine/trap.h /* codes for SIGILL, SIGFPE */ but does that mean we must expose the entire contents of trap.h to userland? The problem is T_ is very common in lex source. We most certainly shouldn't. Everything but the needed bits should be #ifdef KERNEL. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 [EMAIL PROTECTED] | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: 4.4-RC NFS panic
I've just done a further test. I've mounted a directory tree from
Vaio to Vaio using localhost (lo driver) and the test has run
smoothly. So chances would be good the bug is in the ep driver.
Unfortunately...
Andre Albsmeier, who's seeing various network problems, is using
the xe driver (also PCMCIA I think), but the problems go away if
he uses an Etherexpress card on the PCI bus of the same machine.
It seems unlikely to be PCMCIA related ('cos it has nothing to do
with the networking itself) it may just be triggered in machines
with slower networking.
David.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
Re: 4.4-RC NFS panic
On Tue, 21-Aug-2001 at 09:35:34 +0100, David Malone wrote:
I've just done a further test. I've mounted a directory tree from
Vaio to Vaio using localhost (lo driver) and the test has run
smoothly. So chances would be good the bug is in the ep driver.
Unfortunately...
Andre Albsmeier, who's seeing various network problems, is using
the xe driver (also PCMCIA I think), but the problems go away if
he uses an Etherexpress card on the PCI bus of the same machine.
As I wrote in my PR (#29845), my problems also happen with
the 3C589 which uses the ep driver. So we can sum up to:
1.) Intel Etherexpress PRO/100 PCMCIA (xe driver) crashes
2.) 3Com 589D EtherLink III PCMCIA (ep driver) crashes
3.) Intel Etherexpress PRO/100+ PCI Card (fxp driver) works perfectly
-Andre
It seems unlikely to be PCMCIA related ('cos it has nothing to do
with the networking itself) it may just be triggered in machines
with slower networking.
David.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
Re: 4.4-RC NFS panic
In message [EMAIL PROTECTED] Andre Albsmeier writes: : As I wrote in my PR (#29845), my problems also happen with : the 3C589 which uses the ep driver. So we can sum up to: : : 1.) Intel Etherexpress PRO/100 PCMCIA (xe driver) crashes : 2.) 3Com 589D EtherLink III PCMCIA (ep driver) crashes : 3.) Intel Etherexpress PRO/100+ PCI Card (fxp driver) works perfectly Interesting. I'm not sure what to make of this. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
unpaired splbio() and splx() in vfs_unmountall()
Dear, I modifid kernel to write to disk directly after unmount all mounted filesystem in boot() kern/kern_shutdown.c. However, I found that my IO requests wouldn't have callback from device interrupt routine. I traced the codes and use gdb to find something out. The interesting is after execute vfs_unmountall() - dounmount() - ffs_unmount() - ffs_flushfiles() - vflush()- ??? the interrupt mask is set by splbio() without splx(), therefore, all my following requests cannot return. Notice that, the situation only happens after heavy IO, for example: cp 30 files at the same time. I use spl0() to solve the prolbem, but I think it's not the right way to do that. Can anyone provide some clues or suggestions. Thanks -- Rex Luo Tel : 886-2-25521814 Ext. 824 Fax : 886-2-25521824 e-mail : [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: 4.4-RC NFS panic
On Tue, 21-Aug-2001 at 03:07:33 -0600, Warner Losh wrote: In message [EMAIL PROTECTED] Andre Albsmeier writes: : As I wrote in my PR (#29845), my problems also happen with : the 3C589 which uses the ep driver. So we can sum up to: : : 1.) Intel Etherexpress PRO/100 PCMCIA (xe driver) crashes : 2.) 3Com 589D EtherLink III PCMCIA (ep driver) crashes : 3.) Intel Etherexpress PRO/100+ PCI Card (fxp driver) works perfectly Interesting. I'm not sure what to make of this. So do I. Ian Dowse already sent me a program to inspect the mbufs in the crashdumps. I don't know a lot about mbufs but the output appears really hosed... I will try it again using another PCMICA card I just got... -Andre To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Zope's performance issues
Hello, I'm using Zope application server (www.zope.org) on FreeBSD
and I'm experiencing some serious performance issues.
Since I hadn't noticed any problems at the past with Linux, I did
some benchmarks to investigate it a little.
I've written a small program that connects to Zope and tries to
retrieve a non-existant URL (source code appended). While Linux
performed almost linearly depending on the number of hits/sec,
FreeBSD (and OpenBSD) performed exponential. I allready sent
a mail to the zope-list but I did not get much info.
Since Zope is a threaded application I was wondering maybe it
was related with the thread implementation (the scheduling maybe?)
I did the following experiment:
I installed a binary distribution of python-1.5.2 (pkg_add or apt-get)
and I compiled Zope-2.3.3+HotFix from sources.
I started running the previous program and after about 1000 hits
(when the server would be a bit loaded) I reloaded a fairly complex
page and measured the time it gets to finish. The results are below:
+-++---+-+
|OS | Hit Rate* | Real hit rate | Reload time |
| | (hits/sec) | (hits/sec)| (sec) |
+-++---+-+
| | 30 | 20 | 13 |
| Linux debian 2.2.19 | 50 | 33 | 16 |
| |100 | 38 | 16 |
+-++---+-+
| | 30 | 20 | 10 |
| FreeBSD 4.3-Release | 50 | 33 | 23 |
| |100 | 40 | 100 |
+-++---+-+
| | 30 | 20 | 10 |
| OpenBSD 2.9-Release | 50 | 33 | 15 |
| |100 | 50 | 285 |
+-++---+-+
*The hit rate is the one passed as an argument to the program, while
real hit rate was calculated according to Zope's log file.
All tests were ran on the same computer, a 600MHz Celeron with
192MB RAM
These tests were ran using the default content. When I use a real
(more heavy) content I even get time-outs when I try to access
some pages.
Has anyone in this list had any previous experience with zope? and
how do tou explain the big differences between the OSs?
Thank you in advance
Giorgos Verigakis
-- 8 --
#include stdio.h
#include stdlib.h
#include unistd.h
#include sys/types.h
#include sys/socket.h
#include netinet/in.h
#include netdb.h
int main(int argc, char *argv[])
{
struct sockaddr_in sa;
struct hostent *hp;
int s;
int i, t, len;
unsigned long sleeptime;
char request[128];
char *url=/foo;
if (argc != 4) {
printf(Usage: %s host port rate\n, argv[0]);
printf(rate in hits/sec\n);
exit(1);
}
if ((hp = gethostbyname(argv[1])) == NULL) {
printf(error looking up host\n);
exit(1);
}
sprintf(request, GET %s HTTP/1.0\015\012\015\012, url);
len = strlen(request);
sleeptime = (1 / atof(argv[3])) * 100;
bzero(sa, sizeof(sa));
bcopy(hp-h_addr, (char *)sa.sin_addr, hp-h_length);
sa.sin_family = hp-h_addrtype;
sa.sin_port = htons(atoi(argv[2]));
for (i = 0;; i++) {
if (( s = socket(hp-h_addrtype, SOCK_STREAM, 0)) 0) {
printf(Socket error\n);
continue;
}
if (connect(s, (struct sockaddr *) sa, sizeof sa) 0) {
printf(Connection error\n);
close(s);
continue;
}
t = write(s, request, len);
printf(i=%d: %d bytes\n, i, t);
usleep(sleeptime);
close(s);
}
}
-- 8 --
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
Re: 4.4-RC NFS panic
On Tue, 21-Aug-2001 at 03:07:33 -0600, Warner Losh wrote: In message [EMAIL PROTECTED] Andre Albsmeier writes: : As I wrote in my PR (#29845), my problems also happen with : the 3C589 which uses the ep driver. So we can sum up to: : : 1.) Intel Etherexpress PRO/100 PCMCIA (xe driver) crashes : 2.) 3Com 589D EtherLink III PCMCIA (ep driver) crashes : 3.) Intel Etherexpress PRO/100+ PCI Card (fxp driver) works perfectly Interesting. I'm not sure what to make of this. We can now add: 4.) D-Link DFE-650 PCMCIA (ed driver)freezes :-( Warner, I have seen your mails regarding pcic-44rc1.diff.1. My box has a TI PCI-1225 chip... I will try the patch... -Andre To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Zope's performance issues
On Tue, 21 Aug 2001, Giorgos Verigakis wrote: Hello, I'm using Zope application server (www.zope.org) on FreeBSD and I'm experiencing some serious performance issues. Since I hadn't noticed any problems at the past with Linux, I did some benchmarks to investigate it a little. I've written a small program that connects to Zope and tries to retrieve a non-existant URL (source code appended). While Linux performed almost linearly depending on the number of hits/sec, FreeBSD (and OpenBSD) performed exponential. I allready sent a mail to the zope-list but I did not get much info. Since Zope is a threaded application I was wondering maybe it was related with the thread implementation (the scheduling maybe?) I suspect that the reload time spiking is due to our initial sequence number generation scheme, which is currently shared with OpenBSD. This will be changing RSN, which will likely change how your benchmarks look. (They should look better after the change.) I'm curious about the reload time, though. Is it really in seconds? What is it measuring? Oops, I just noticed the source below... I'll read it and try out the test to see if my above theory is correct later today. Mike Silby Silbersack To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Zope's performance issues
On Tue, 21 Aug 2001, Mike Silbersack wrote: On Tue, 21 Aug 2001, Giorgos Verigakis wrote: Hello, I'm using Zope application server (www.zope.org) on FreeBSD and I'm experiencing some serious performance issues. Since I hadn't noticed any problems at the past with Linux, I did some benchmarks to investigate it a little. I've written a small program that connects to Zope and tries to retrieve a non-existant URL (source code appended). While Linux performed almost linearly depending on the number of hits/sec, FreeBSD (and OpenBSD) performed exponential. I allready sent a mail to the zope-list but I did not get much info. Since Zope is a threaded application I was wondering maybe it was related with the thread implementation (the scheduling maybe?) I suspect that the reload time spiking is due to our initial sequence number generation scheme, which is currently shared with OpenBSD. This will be changing RSN, which will likely change how your benchmarks look. (They should look better after the change.) I don't understand. How is it related with the initial sequence number? I'm curious about the reload time, though. Is it really in seconds? What is it measuring? Yes it's in seconds. What I wanted to measure was the time a page needs to be loaded, because the results I was getting were unexpectable. Oops, I just noticed the source below... I'll read it and try out the test to see if my above theory is correct later today. Mike Silby Silbersack To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
inet6 address host - network order
What's the most portable way of converting an IPv6 address from host to network order? Cheers, -- Jacques Vidrine / [EMAIL PROTECTED] / [EMAIL PROTECTED] / [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: inet6 address host - network order
On Tue, Aug 21, 2001 at 09:37:49AM -0500, Jacques A. Vidrine wrote: What's the most portable way of converting an IPv6 address from host to network order? Nothing like following up your own post. RFC 2553 indicates that the IPv6 address is stored in the `struct in6_addr' in network byte order, so there shouldn't be any need to swap bytes. Cheers, -- Jacques Vidrine / [EMAIL PROTECTED] / [EMAIL PROTECTED] / [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Zope's performance issues (cont.)
To check if it was indeed a thread problem, I repeated my tests using Zope-2.4.0 and python 2.1.1 First I did a default python compilation (using libc_r) and then using pth (GNU portable threads) Using the program I sent on my previous email and a hit rate of 100 I got libc_r: 243 sec, 28 real hits/sec pth: 15 sec, 36 real hits/sec So, I guess there is some problem on libc_r Giorgos PS: I did the same on OpenBSD and had similar results To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: wireless nic recommendations
Dan Cox wrote: Here is my situation. My house can't get DSL or cable. Our neighbors who live 20-30 feet away do have DSL and have agreed to share the connection. To make a long story short we have successfully set up a wireless LAN for the two houses, we've been using a windows laptop to test the connection. I would like to find out what wireless NIC's are compatible with freeBSD or some recommendations. I have a wireless ISP, my connection to the internet is a Cisco/Aironet: an0: Aironet PCI4800 port 0x6700-0x673f,0x6600-0x667f mem 0xe1001000- 0xe100107f irq 12 at device 13.0 on pci0 I've had no problems with this card, it is currently reaching about 6.5 miles without an amplifier, using a 24 dB fruit basket antenna on the roof of my house. The throughput is sufficient to hit my 300 Kbps band- width limitation enforced at the access point 24x7. The initial speed tests performed by the ISP showed speed consistent above 1,000 Kbps. You should have little problem with speed going between two houses. The Cisco cards are pretty expensive compared to other brands. -- Where am I, and what am I doing in this handbasket? Wes Peters Softweyr LLC [EMAIL PROTECTED] http://softweyr.com/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
notebook HW
Is there a maiing list that deal especially with FreeBSD on notebook issues? If not, could it be created or is this covered by another list? (hardware?) -- Chris Christoph P. U. Kukulies [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
XMM[0-7] preserved across context switch?
A quick peek at swtch.s seems to show that the SSE registers (XMM0-7) aren't being preserved across context switches. Am I missing somewhere that's doing this, or are they really not being saved now? -- Kevin Day [EMAIL PROTECTED] - [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: notebook HW
On Tue, Aug 21, 2001 at 07:26:05PM +0300, Peter Pentchev wrote: On Tue, Aug 21, 2001 at 06:24:10PM +0200, Christoph Kukulies wrote: Is there a maiing list that deal especially with FreeBSD on notebook issues? If not, could it be created or is this covered by another list? (hardware?) Try [EMAIL PROTECTED] :) Argh. How could I overlook that :-) Thanks. -- Chris Christoph P. U. Kukulies [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: XMM[0-7] preserved across context switch?
On Tue, Aug 21, 2001 at 11:27:38AM -0500, Kevin Day wrote: A quick peek at swtch.s seems to show that the SSE registers (XMM0-7) aren't being preserved across context switches. Am I missing somewhere that's doing this, or are they really not being saved now? SSE support has recently been added to -current and -stable. David. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: unpaired splbio() and splx() in vfs_unmountall()
* Rex Luo [EMAIL PROTECTED] [010821 04:14] wrote: Dear, I modifid kernel to write to disk directly after unmount all mounted filesystem in boot() kern/kern_shutdown.c. However, I found that my IO requests wouldn't have callback from device interrupt routine. I traced the codes and use gdb to find something out. The interesting is after execute vfs_unmountall() - dounmount() - ffs_unmount() - ffs_flushfiles() - vflush()- ??? the interrupt mask is set by splbio() without splx(), therefore, all my following requests cannot return. Notice that, the situation only happens after heavy IO, for example: cp 30 files at the same time. I use spl0() to solve the prolbem, but I think it's not the right way to do that. Can anyone provide some clues or suggestions. That makes no sense. All you need to do is find the where the splbio() is that doesn't have a corresponding splx(). You can use the SPLASSERT functions to do that. -- -Alfred Perlstein [[EMAIL PROTECTED]] Ok, who wrote this damn function called '??'? And why do my programs keep crashing in it? To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Zope's performance issues (cont.)
* Giorgos Verigakis [EMAIL PROTECTED] [010821 10:55] wrote: To check if it was indeed a thread problem, I repeated my tests using Zope-2.4.0 and python 2.1.1 First I did a default python compilation (using libc_r) and then using pth (GNU portable threads) Using the program I sent on my previous email and a hit rate of 100 I got libc_r: 243 sec, 28 real hits/sec pth: 15 sec, 36 real hits/sec So, I guess there is some problem on libc_r Giorgos PS: I did the same on OpenBSD and had similar results Do you think you could link it using profiling? 'cc -pg' Then run zope, then use gprof to see what's going on? -- -Alfred Perlstein [[EMAIL PROTECTED]] Ok, who wrote this damn function called '??'? And why do my programs keep crashing in it? To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
tunable support for ata interrupt sharing
Attached it a patch to make sharing of the main ata control interrupts
dependent on a tunable, hw.ata.shared_irqs. This is required for my new
HP Omnibook 500 to use the CMD 648 in the expansion base to work as it
appears hardwired to interrupt 15 (which is fairly logical given that
there is no where to attach devices to the secondary channel.) If this
looks ok and you don't have time to deal with it, I'd be happy to commit
it myself.
-- Brooks
--
Any statement of the form X is the one, true Y is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4
Index: sys/dev/ata/ata-pci.c
===
RCS file: /usr/cvs/src/sys/dev/ata/ata-pci.c,v
retrieving revision 1.6
diff -u -r1.6 ata-pci.c
--- sys/dev/ata/ata-pci.c 8 Jun 2001 09:51:33 - 1.6
+++ sys/dev/ata/ata-pci.c 21 Aug 2001 05:08:32 -
@@ -49,6 +49,15 @@
#include pci/pcireg.h
#include dev/ata/ata-all.h
+/* internal vars */
+static int shared_irqs = 0;
+TUNABLE_INT(hw.ata.shared_irqs, shared_irqs);
+
+/* systcl vars */
+SYSCTL_DECL(_hw_ata);
+SYSCTL_INT(_hw_ata, OID_AUTO, shared_irqs, CTLFLAG_RD, shared_irqs, 0,
+ Share PCI IRQs?);
+
/* misc defines */
#define IOMASK 0xfffc
#define ATA_MASTERDEV(dev) ((pci_get_progif(dev) 0x80) \
@@ -515,7 +524,8 @@
return BUS_ALLOC_RESOURCE(device_get_parent(dev), child,
SYS_RES_IRQ, rid,
- irq, irq, 1, flags ~RF_SHAREABLE);
+ irq, irq, 1, flags
+ (shared_irqs ? ~0 : ~RF_SHAREABLE));
#endif
}
else {
Index: share/man/man4/ata.4
===
RCS file: /usr/cvs/src/share/man/man4/ata.4,v
retrieving revision 1.20
diff -u -r1.20 ata.4
--- share/man/man4/ata.414 Jul 2001 19:40:46 - 1.20
+++ share/man/man4/ata.421 Aug 2001 19:19:56 -
@@ -71,6 +71,9 @@
.It Va hw.ata.tags
set to 1 to enable Tagged Queuing support (default is disabled)
(only IBM DPTA and DTLA drives support that)
+.It Va hw.ata.shared_irqs
+set to 1 to allow sharing interrupts with the main controler
+(default is disabled)
.El
.Sh DESCRIPTION
This driver provides access to disk drives, ATAPI CD-ROM and DVD drives,
PGP signature
Re: tunable support for ata interrupt sharing
It seems Brooks Davis wrote: Attached it a patch to make sharing of the main ata control interrupts dependent on a tunable, hw.ata.shared_irqs. This is required for my new HP Omnibook 500 to use the CMD 648 in the expansion base to work as it appears hardwired to interrupt 15 (which is fairly logical given that there is no where to attach devices to the secondary channel.) If this looks ok and you don't have time to deal with it, I'd be happy to commit it myself. I have just today committed always sharing all irq's to -current, the consensus is that if the BIOS allows sharing it should work. This makes sense, the MB maker is the only one that knows if this is working, if they blow it in thier BIOS, well -Søren To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Zope's performance issues
On Tue, 21 Aug 2001, Giorgos Verigakis wrote: On Tue, 21 Aug 2001, Mike Silbersack wrote: I suspect that the reload time spiking is due to our initial sequence number generation scheme, which is currently shared with OpenBSD. This will be changing RSN, which will likely change how your benchmarks look. (They should look better after the change.) I don't understand. How is it related with the initial sequence number? I'm curious about the reload time, though. Is it really in seconds? What is it measuring? Yes it's in seconds. What I wanted to measure was the time a page needs to be loaded, because the results I was getting were unexpectable. The current scheme causes problems with TIME_WAIT recycling, which may cause long delays in establishing new connections if you're connection to the same host rapidly enough to cause TIME_WAIT recycling to be an issue. This is why there's a huge spike only when you get to 1000 hits/second. There could be other reasons, of course, but this will overshadow the others. I tested your test program with apache, and the change is noticeable. So, hold off on further testing until later this week. Mike Silby Silbersack To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Zope's performance issues
On Tue, 21 Aug 2001, Mike Silbersack wrote: The current scheme causes problems with TIME_WAIT recycling, which may cause long delays in establishing new connections if you're connection to the same host rapidly enough to cause TIME_WAIT recycling to be an issue. This is why there's a huge spike only when you get to 1000 hits/second. No, maybe I wasn't clear. I didn't get 1000 hits/sec, what I said was that I measured the time a page needs to be loaded after the program had made 1000 hits to Zope (so that it would be a bit a loaded). Also I don't know if this is caused by TIME_WAIT because when I do a netstat I don't see a lot of connections on TIME_WAIT Actually I think it has to do with the threads library (see my other mail on that) There could be other reasons, of course, but this will overshadow the others. I tested your test program with apache, and the change is noticeable. So, hold off on further testing until later this week. Mike Silby Silbersack To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: Zope's performance issues
On Tue, 21 Aug 2001, Mike Silbersack wrote: On Tue, 21 Aug 2001, Mike Silbersack wrote: On Tue, 21 Aug 2001, Giorgos Verigakis wrote: The current scheme causes problems with TIME_WAIT recycling, which may cause long delays in establishing new connections if you're connection to the same host rapidly enough to cause TIME_WAIT recycling to be an issue. This is why there's a huge spike only when you get to 1000 hits/second. There could be other reasons, of course, but this will overshadow the others. I tested your test program with apache, and the change is noticeable. So, hold off on further testing until later this week. Mike Silby Silbersack There's a second problem with your program too: It causes TIME_WAITs to pile up on the client side, which is not a well-handled situation for us. Use apachebench for future testing, it will automatically generate the stats you are looking for. Yes tou are right on that, it was that on my system (the client) the connections were closing after a few seconds and I hadn't noticed it. Mike Silby Silbersack To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Any ATAPI gurus out there?
Okay, here's the deal. My friend, whom I am trying to teach unix [4.3-stable] was having some problems on his secondary ATAPI bus. A Creative CD-ROM was the master, a HP burner was the slave. I'm a SCSI guy meself, and I think the following may be the problem, but as such, I don't know for sure... The CD-ROM was having data dropouts, and was due to old age, no doubt. Under Winblowz, the HP burner would only get 4x when it was able to burn. it's a 12x/8x/32x drive. Under BSD, the CD-ROM would be usable, but would be prone to dropouts. Under BSD, the HP burner came up with sense errors on boot, prior to a proper probe reply. It would hang for the duration of several timeouts at the point before it got the probe. Under BSD, the HP burner would cause a terminal wait state upon any access [such as a mount request, or a burncd command]. Red-button time... This last Friday, he bought a new CD-ROM, and a new burner, as we both thought both drives had issues. He gave me the burner [after a promise to pay him $50 for the under-a-year-old burner, assuming I could fix it]... He put his new drives in his box. Everything works now, under Winblowz AND BSD. I put the the HP burner in MY box, and voila! Nothing is wrong with it... Questions: 1). Can a flaky ATAPI Master cause a good ATAPI Slave to APPEAR [however incorrectly] that the Slave has a problem? 2). If #1 is true, then, why? One added benefit, I've been pushing him to go SCSI [with the good eBay prices on reliable used and new drives, as well as the independance of the drives, the speed, etc... This has gotten him one step closer to committing to SCSI, so not all is lost in this story.. He got two new drives, I got a helluva deal on a good burner, and he may convert to SCSI... I am relatively ignorant on ATAPI matters, and under SCSI the devices are independant, but we do have this question concerning the Master/Slave thing... jim -- ET has one helluva sense of humor! He's always anal-probing right-wing schizos! _ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: ld -X == important or not?
In article 3B7FE5C2.18273.16C3288@localhost, Dan Langille [EMAIL PROTECTED] wrote: How important is the -X option on ld? -X Delete all temporary local symbols. For most tar- gets, this is all local symbols whose names begin with `L'. I ask because I'm porting something to Solaris and it seems rather odd that the solaris ld doesn't have this option. It's not important. It just makes the output file smaller. I wouldn't be surprised if the Solaris linker did this by default. John -- John Polstra [EMAIL PROTECTED] John D. Polstra Co., Inc.Seattle, Washington USA Disappointment is a good sign of basic intelligence. -- Chögyam Trungpa To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: secure Filesystem
On Thu, Aug 16, 2001 at 12:48:59PM -0700, Darryl Okahata wrote: A bigger problem is that doing anything with a file uses up 1-2KB PER FILE. If you want to see cfsd grow *really big*, do a find . of any large cfs-controlled hierarchy with lots of files. I'd really like to put my MH mail messages under cfs, but I've got too many files (I can't afford having a 200+MB cfsd). I don't seem to have that problem... voyager ~ du -sk crypt 1342056 crypt voyager ~ find /crypt/obonilla /dev/null voyager ~ du -sk crypt 1342056 crypt am i missing something? should i just be happy ;) regards, -oscar -- pgp fingerprint: BC64 2E7A CAEF 39E1 9544 80CA F7D5 784D FB46 16C1 To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
vhost problems
Hi everyone, For the last week I have been trying everything I know to try an work out my problem. Since it seems I have exausted all my resources, I figured I'd try here. I'm trying to export vhosts on my system and it isnt working. I'm using FreeBSD 4.4-RC, along with ipnat/ipf and my connection is pppoe. I have 1 assigned ip I connect with(static), the one tun0 obtains. I also have a block of static external addressed assigned on my external interface rl0, also obtained from my isp. My connection ip isn't from the same ip block as my /29 obtained from my ISP. It seems all my incoming connections to each different ip on the block worka fine. I am fully able to ipnat the ip to another internal box via sis0 the internal interface. The problem is all my outbound connections are all seen as from tun0 interface ip, even when I export an ip or hostname locally. How is it possible to be able to export hosts, and have one of my ip block addresses seen on the internet, as opposed to having all outbound traffic seen from the 1 tun0 connection ip? Jason [EMAIL PROTECTED]
Re: unpaired splbio() and splx() in vfs_unmountall()
Dear Alfred, I have tried to add several asserts to verify its interrupt mask, however, the abnormal behaviour disappered if I did that. That's really rediculous and I don't know why? I would continue to find out what's the reason, and feedback if something new. Anyway, I really appreciate your help and kindness. Further question, what kind of books, or news, information I can get to study about pc's interrupt handling and architecture. Thanks, -- Rex Luo Alfred Perlstein [EMAIL PROTECTED] said: * Rex Luo [EMAIL PROTECTED] [010821 04:14] wrote: Dear, I modifid kernel to write to disk directly after unmount all mounted filesystem in boot() kern/kern_shutdown.c. However, I found that my IO requests wouldn't have callback from device interrupt routine. I traced the codes and use gdb to find something out. The interesting is after execute vfs_unmountall() - dounmount() - ffs_unmount() - ffs_flushfiles() - vflush()- ??? the interrupt mask is set by splbio() without splx(), therefore, all my following requests cannot return. Notice that, the situation only happens after heavy IO, for example: cp 30 files at the same time. I use spl0() to solve the prolbem, but I think it's not the right way to do that. Can anyone provide some clues or suggestions. That makes no sense. All you need to do is find the where the splbio() is that doesn't have a corresponding splx(). You can use the SPLASSERT functions to do that. -- -Alfred Perlstein [[EMAIL PROTECTED]] Ok, who wrote this damn function called '??'? And why do my programs keep crashing in it? -- Rex Luo Tel : 886-2-25521814 Ext. 824 Fax : 886-2-25521824 e-mail : [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: 4.4-RC NFS panic
On Tue, 21-Aug-2001 at 11:45:12 -0600, Warner Losh wrote:
In message [EMAIL PROTECTED] David Malone writes:
: Andre Albsmeier, who's seeing various network problems, is using
: the xe driver (also PCMCIA I think), but the problems go away if
: he uses an Etherexpress card on the PCI bus of the same machine.
:
: It seems unlikely to be PCMCIA related ('cos it has nothing to do
: with the networking itself) it may just be triggered in machines
: with slower networking.
After talking with Ian Dowse, I think that we've hammered out what may
cause this. Basically, the problem is
code in net doing splnet()
interrupt here - pcic_pci_intr - netcard_intr - network code.
And we've interrupted the critical section, broken all kinds of
invariants.
Warner
P.S. I think that with Ian's other interrupt changes, we can do the
following w/o problems. This should fix the network problems, I
think.
Runs perfectly for about 10 minutes now under full load. It didn't
survive 10 seconds before :-)
I still have the hangs on a warm reboot but this is a different
story...
Thanks a lot for the quick help!
-Andre
Index: pcic_pci.c
===
RCS file: /cache/ncvs/src/sys/pccard/pcic_pci.c,v
retrieving revision 1.54.2.7
diff -u -r1.54.2.7 pcic_pci.c
--- pcic_pci.c2001/08/21 09:06:25 1.54.2.7
+++ pcic_pci.c2001/08/21 17:18:06
@@ -515,15 +515,6 @@
* in the CD change.
*/
sp-getb(sp, PCIC_STAT_CHG);
-
- /*
- * If we have a card in the slot with an interrupt handler, then
- * call it. Note: This means that each card can have at most one
- * interrupt handler for it. Since multifunction cards aren't
- * supported, this shouldn't cause a problem in practice.
- */
- if (sc-cd_present sp-intr != NULL)
- sp-intr(sp-argp);
}
/*
@@ -784,36 +775,6 @@
return (0);
}
-static int
-pcic_pci_setup_intr(device_t dev, device_t child, struct resource *irq,
-int flags, driver_intr_t *intr, void *arg, void **cookiep)
-{
- struct pcic_softc *sc = (struct pcic_softc *) device_get_softc(dev);
- struct pcic_slot *sp = sc-slots[0];
-
- if (sp-intr) {
- device_printf(dev,
-Interrupt already established, possible multiple attach bug.\n);
- return (EINVAL);
- }
- sp-intr = intr;
- sp-argp = arg;
- *cookiep = sc;
- return (0);
-}
-
-static int
-pcic_pci_teardown_intr(device_t dev, device_t child, struct resource *irq,
-void *cookie)
-{
- struct pcic_softc *sc = (struct pcic_softc *) device_get_softc(dev);
- struct pcic_slot *sp = sc-slots[0];
-
- sp-intr = NULL;
- sp-argp = NULL;
- return (0);
-}
-
static device_method_t pcic_pci_methods[] = {
/* Device interface */
DEVMETHOD(device_probe, pcic_pci_probe),
@@ -829,8 +790,8 @@
DEVMETHOD(bus_release_resource, bus_generic_release_resource),
DEVMETHOD(bus_activate_resource, pcic_activate_resource),
DEVMETHOD(bus_deactivate_resource, pcic_deactivate_resource),
- DEVMETHOD(bus_setup_intr, pcic_pci_setup_intr),
- DEVMETHOD(bus_teardown_intr,pcic_pci_teardown_intr),
+ DEVMETHOD(bus_setup_intr, bus_generic_setup_intr),
+ DEVMETHOD(bus_teardown_intr,bus_generic_teardown_intr),
/* Card interface */
DEVMETHOD(card_set_res_flags, pcic_set_res_flags),
--
BSD, from the people who brought you TCP/IP.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-hackers in the body of the message
Re: 4.4-RC NFS panic
In message [EMAIL PROTECTED] Andre Albsmeier writes: : I still have the hangs on a warm reboot but this is a different : story... Eh? what kind of hangs and when? Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
Re: 4.4-RC NFS panic
On Tue, 21-Aug-2001 at 23:44:40 -0600, Warner Losh wrote: In message [EMAIL PROTECTED] Andre Albsmeier writes: : I still have the hangs on a warm reboot but this is a different : story... Eh? what kind of hangs and when? Attached below is the dmesg... It hangs only when warm booting; after a power toggle everything is OK... Copyright (c) 1992-2001 The FreeBSD Project. Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved. FreeBSD 4.4-RC #23: Wed Aug 22 07:21:34 CEST 2001 [EMAIL PROTECTED]:/src/obj-4/src/src-4/sys/schlappy Calibrating clock(s) ... TSC clock: 30160 Hz, i8254 clock: 1193146 Hz Timecounter i8254 frequency 1193146 Hz CPU: Pentium II/Pentium II Xeon/Celeron (366.66-MHz 686-class CPU) Origin = GenuineIntel Id = 0x66a Stepping = 10 Features=0x183f9ffFPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,MMX,FXSR real memory = 134152192 (131008K bytes) Physical memory chunk(s): 0x1000 - 0x0009efff, 647168 bytes (158 pages) 0x00325000 - 0x07febfff, 130838528 bytes (31943 pages) avail memory = 127590400 (124600K bytes) bios32: Found BIOS32 Service Directory header at 0xc00f6230 bios32: Entry = 0xfd790 (c00fd790) Rev = 0 Len = 1 pcibios: PCI BIOS entry at 0x225 pnpbios: Found PnP BIOS data at 0xc00f6260 pnpbios: Entry = f:a34e Rev = 1.0 pnpbios: Event flag at 4b4 Other BIOS signatures found: ACPI: 000f61f0 Preloaded elf kernel kernel at 0xc02ff000. Pentium Pro MTRR support enabled pci_open(1):mode 1 addr port (0x0cf8) is 0x8000384c pci_open(1a): mode1res=0x8000 (0x8000) pci_cfgcheck: device 0 [class=06] [hdr=00] is there (id=71908086) Using $PIR table, 7 entries at 0xc00fdf50 apm0: APM BIOS on motherboard apm: found APM BIOS v1.2, connected at v1.2 npx0: math processor on motherboard npx0: INT 16 interface pcib0: Intel 82443BX (440 BX) host to PCI bridge on motherboard found- vendor=0x8086, dev=0x7190, revid=0x03 class=06-00-00, hdrtype=0x00, mfdev=0 subordinatebus=0secondarybus=0 map[10]: type 1, range 32, base f800, size 26 found- vendor=0x8086, dev=0x7191, revid=0x03 class=06-04-00, hdrtype=0x01, mfdev=0 subordinatebus=1secondarybus=1 found- vendor=0x8086, dev=0x7110, revid=0x02 class=06-80-00, hdrtype=0x00, mfdev=1 subordinatebus=0secondarybus=0 found- vendor=0x8086, dev=0x7111, revid=0x01 class=01-01-80, hdrtype=0x00, mfdev=0 subordinatebus=0secondarybus=0 map[20]: type 1, range 32, base fcd0, size 4 found- vendor=0x8086, dev=0x7112, revid=0x01 class=0c-03-00, hdrtype=0x00, mfdev=0 subordinatebus=0secondarybus=0 intpin=d, irq=9 map[20]: type 1, range 32, base fce0, size 5 found- vendor=0x8086, dev=0x7113, revid=0x02 class=06-80-00, hdrtype=0x00, mfdev=0 subordinatebus=0secondarybus=0 map[90]: type 1, range 32, base 2180, size 4 found- vendor=0x104c, dev=0xac1c, revid=0x01 class=06-07-00, hdrtype=0x02, mfdev=1 subordinatebus=0secondarybus=0 intpin=a, irq=10 found- vendor=0x104c, dev=0xac1c, revid=0x01 class=06-07-00, hdrtype=0x02, mfdev=1 subordinatebus=0secondarybus=0 intpin=b, irq=11 pci0: PCI bus on pcib0 pcib1: Intel 82443BX (440 BX) PCI-PCI (AGP) bridge at device 1.0 on pci0 found- vendor=0x10c8, dev=0x0005, revid=0x12 class=03-00-00, hdrtype=0x00, mfdev=1 subordinatebus=0secondarybus=0 intpin=a, irq=10 map[10]: type 1, range 32, base f600, size 24 map[14]: type 1, range 32, base fe40, size 22 map[18]: type 1, range 32, base feb0, size 20 found- vendor=0x10c8, dev=0x8005, revid=0x12 class=04-01-00, hdrtype=0x00, mfdev=1 subordinatebus=0secondarybus=0 intpin=b, irq=11 map[10]: type 1, range 32, base f780, size 22 map[14]: type 1, range 32, base fea0, size 20 pci1: PCI bus on pcib1 pci1: NeoMagic MagicMedia 256AV SVGA controller (vendor=0x10c8, dev=0x0005) at 0.0 irq 10 chip1: NeoMagic MagicMedia 256AX Audio controller mem 0xfea0-0xfeaf,0xf780-0xf7bf irq 11 at device 0.1 on pci1 isab0: Intel 82371AB PCI to ISA bridge at device 7.0 on pci0 isa0: ISA bus on isab0 atapci0: Intel PIIX4 ATA33 controller port 0xfcd0-0xfcdf at device 7.1 on pci0 ata0: iobase=0x01f0 altiobase=0x03f6 bmaddr=0xfcd0 ata0: mask=03 status0=50 status1=50 ata0: mask=03 ostat0=50 ostat2=50 ata0-slave: ATAPI probe a=14 b=eb ata0-master: ATAPI probe a=00 b=00 ata0: mask=03 status0=50 status1=00 ata0-master: ATA probe a=01 b=a5 ata0: devices=09 ata0: at 0x1f0 irq 14 on atapci0 ata1: iobase=0x0170 altiobase=0x0376 bmaddr=0xfcd8 ata1: mask=00 status0=ff status1=ff ata1: probe allocation failed pci0: Intel 82371AB/EB (PIIX4) USB controller (vendor=0x8086, dev=0x7112) at
Re: 4.4-RC NFS panic
In message [EMAIL PROTECTED] Andre Albsmeier writes: : Attached below is the dmesg... It hangs only when warm booting; after : a power toggle everything is OK... ... : pcic0: Event mask 0xf stat 0x3419 : ### : ### Now it hangs until poweroff/poweron ### : ### OK. Looks like maybe an interrupt storm on warm boot. I'll have to check into this. Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-hackers in the body of the message
