On Mon, 2006-12-18 at 00:13 -0800, Andreas wrote:

Le dimanche 17 décembre 2006 19:59, Chip Cooper a écrit : It doesn't help you much, but I can comfirm your observations. Doesn't work in openSUSE 10.2, does work on a Windows PC.

A Windows XP Guest on openSUSE 10.2 (vmware) can also connect to the website.

And finally a connection from Fedora Core 6 fails.

Maybe a kernel problem?

Maybe a web browser issue? Which browsers are working and which are not working? -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2006/12/18 09:22 (GMT-0500) Chip Cooper apparently typed:

Well, I've tried konqueror, firefox, mozilla, and opera. None work. Can't get a ping from the terminal. Don't think it is a browser issue. It is interesting Andreas was unable to load the site using Fedora core 6. So maybe a kernel issue/parameter/bug? No idea what is going on. I'm open to ideas!

As many failures as seen in this thread, it has to be something about the site configuration, likely IPV6. What was that URL again? -- "Let your conversation be always full of grace." Colossians 4:6 NIV Team OS/2 ** Reg. Linux User #211409 Felix Miata *** http://mrmazda.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Felix Miata wrote:

On 2006/12/18 09:22 (GMT+-0500) Chip Cooper apparently typed:

...

...

...

http://www.marymount.edu/

Well, I've tried konqueror, firefox, mozilla, and opera. None work. Can't get a ping from the terminal. Out of curiosity (and frustration), I decided to install the 2.6.16 kernel from my 10.1 disks. When I boot with this kernel, the website loads fine. So, it seems something is amiss in 2.6.18 in allowing certain sites to load, but I have no clue what it could be.

Chip -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 12/19/06, Chip Cooper wrote:

Felix Miata wrote:

...

On 2006/12/18 09:22 (GMT+-0500) Chip Cooper apparently typed:

...

...

...

http://www.marymount.edu/

Well, I've tried konqueror, firefox, mozilla, and opera. None work. Can't get a ping from the terminal. Out of curiosity (and frustration), I decided to install the 2.6.16 kernel from my 10.1 disks. When I boot with this kernel, the website loads fine. So, it seems something is amiss in 2.6.18 in allowing certain sites to load, but I have no clue what it could be.

Chip

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Out of curiosity tried it from the office from ThinkPad with OpenSuSE 10.2 (kernel 2.6.18.2-34-default) and site opened in FF without any problem. In both cases (in the office and at home) I'm behind NAT/firewall. Since initial report says ping to this address did not work, it's not browser, but something below. -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Monday 18 December 2006 22:33, Mark Goldstein wrote:

Out of curiosity tried it from the office from ThinkPad with OpenSuSE 10.2 (kernel 2.6.18.2-34-default) and site opened in FF without any problem.

What kind of Nic does the Thinkpad have? My test was on a brand new dell server, with the same kernal as yours (but x86_64), using Broadcom Corporation NetXtreme BCM5721 Gigabit Ethernet PCI Express -- _____________________________________ John Andersen

John Andersen wrote:

On Monday 18 December 2006 22:33, Mark Goldstein wrote:

...

Out of curiosity tried it from the office from ThinkPad with OpenSuSE 10.2 (kernel 2.6.18.2-34-default) and site opened in FF without any problem.

What kind of Nic does the Thinkpad have?

My test was on a brand new dell server, with the same kernal as yours (but x86_64), using Broadcom Corporation NetXtreme BCM5721 Gigabit Ethernet PCI Express

Is it an arch problem? My 10.2 here at home (x86_64) cannot reach this site either, BUT my 10.2 machine at work, which is i386, did work. I would assume Mark's Thinkpad is an i386 arch as well, and John you mentioned yours was x86_64, so the apparent pattern that I am beginning to see is not a 10.2 problem but a 10.2 x86_64 problem. -- Joe Morris Registered Linux user 231871 running openSUSE 10.2 x86_64 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tue, 2006-12-19 at 18:26 +0800, Joe Morris (NTM) wrote:

John Andersen wrote:

...

On Monday 18 December 2006 22:33, Mark Goldstein wrote:

...

Out of curiosity tried it from the office from ThinkPad with OpenSuSE 10.2 (kernel 2.6.18.2-34-default) and site opened in FF without any problem.

What kind of Nic does the Thinkpad have?

My test was on a brand new dell server, with the same kernal as yours (but x86_64), using Broadcom Corporation NetXtreme BCM5721 Gigabit Ethernet PCI Express

Is it an arch problem? My 10.2 here at home (x86_64) cannot reach this site either, BUT my 10.2 machine at work, which is i386, did work. I would assume Mark's Thinkpad is an i386 arch as well, and John you mentioned yours was x86_64, so the apparent pattern that I am beginning to see is not a 10.2 problem but a 10.2 x86_64 problem.

My laptop is i386 and cannot reach the site, so no it is not an arch problem. -- Ken Schneider UNIX since 1989, linux since 1994, SuSE since 1998 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 12/18/06, Peter Van Lone wrote:

On 12/17/06, Chip Cooper wrote:

...

This has me scratching my head. After a clean install of 10.2, I am now unable to access my work website (www.marymount.edu). All other websites load up just fine. At first, I thought it was a problem with this specific website, but

really weird ... I can get to it from suse10 using ff or konq -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

No problem here: Old IBM Aptiva PC with OpenSuSE 10.2 just installed, FF and Konqueror both opened the site OK. -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Nick Zentena wrote:

On Monday 18 December 2006 12:47, Mark Goldstein wrote:

...

No problem here: Old IBM Aptiva PC with OpenSuSE 10.2 just installed, FF and Konqueror both opened the site OK.

I can't get to www.marymount.edu either. For that matter I can't get to www.keh.com.

OTOH both work from my old 9.2 box. Both boxes connect to the same router.

Nick

I can't get www.keh.com either. I am jealous of mark--wonder why his 10.2 system can access those sites? I disabled ipv6, but it didnt' solve the problem. Interesting Fedora 6 has same problem. That kernel is 2.6.18-1xxx (I believe), and my present kernel is 2.6.18.2-34 Chip

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Hi, On Tuesday 19 December 2006 07:50, Randall R Schulz wrote:

Chip,

...

With all that, only the 10.2 box cannot access http://www.marymount.edu/!

I should have made some note of my hardware: ASUS P5B Deluxe Core 2 Duo 32-bit Linux install % /sbin/lspci | grep -i ether 02:00.0 Ethernet controller: Marvell Technology Group Ltd. Unknown device 4364 (rev 12) 05:04.0 Ethernet controller: Marvell Technology Group Ltd. 88E8001 Gigabit Ethernet Controller (rev 14) So it's not Intel Ethernet NIC that's an issue, it would seem. So I started up Wireshark (nee Ethereal) and tried to spot any problems, and my first hunch, as improbable as it seems, was borne out: A bad checksum!! According to Wireshark, the IP packet that was synthesized to convey the HTTP "GET / HTTP/1.0" message had a bad checksum: Checksum: 0x15a8 [incorrect, should be 0xa4bb (maybe caused by checksum offloading?)] Funky, eh? Since this symptom is occurring on different NICs (and different hardware vendor's NICs), the "checksum offloading?" hypothesis doesn't seem right. In fact, the person who pointed the finger at the 10.2 TCP/IP stack as the culprit appears to have been on the right track. Randall Schulz -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tuesday 19 December 2006 12:43, Randall R Schulz wrote:

I don't know what could be causing this and why it's (so far) manifested only when trying to access this particular site / address, but clearly the checksum generation is the problem.

Perhaps it is in some way associated with the destination IP address or range of addresses?

No it's also happening with www.keh.com for me. I haven't noticed it on any other site but considering the number of sites out there it's likely hitting others. Nick -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Sigh... Apparently the checksum problem is not the issue. Wireshark is reporting this for many (as far as I can tell, all) packets. None of these (except for the marymount.edu site) exhibit the original symptom. I'm not sure why Wireshark is reporting this error spuriously. Perhaps it has something to do with the checksum offloading. Sorry for the false alarm. Randall Schulz -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

John, On Tuesday 19 December 2006 14:09, John Andersen wrote:

On Tuesday 19 December 2006 08:11, Randall R Schulz wrote:

...

ASUS P5B Deluxe Core 2 Duo 32-bit Linux install

How did That happen? Core 2 is x86_64 by default...

Says who? What "default?" And why would I want to incur all that overhead when I have no applications that come close to needing the address space of a 64-bit architecture? % uname -a Linux smiley 2.6.18.2-34-bigsmp #1 SMP Mon Nov 27 11:46:27 UTC 2006 i686 i686 i386 GNU/Linux

...

According to Wireshark, the IP packet that was synthesized to convey the HTTP "GET / HTTP/1.0" message had a bad checksum:

Lots of, (dare I say most?) Gigabit nics have offloading turned on by default. I think there is a way to suppress same, but have to dig it out.

It's clear that the skge driver using the Marvell 8838001 Gigabit Ethernet Controller is doing offloaded checksumming. It's equally clear that the sky2 driver on the SysKonnect controller is not. From what I could find digging around on the Web, drivers that offer the option of controlling offloaded checksum generation (such as the 3c59x) do so via modprobe (see the 6th line of output): % modinfo -p 3c59x debug:3c59x debug level (0-6) options:3c59x: Bits 0-3: media type, bit 4: bus mastering, bit 9: full duplex global_options:3c59x: same as options, but applies to all NICs if options is unset full_duplex:3c59x full duplex setting(s) (1) global_full_duplex:3c59x: same as full_duplex, but applies to all NICs if full_duplex is unset hw_checksums:3c59x Hardware checksum checking by adapter(s) (0-1) flow_ctrl:3c59x 802.3x flow control usage (PAUSE only) (0-1) enable_wol:3c59x: Turn on Wake-on-LAN for adapter(s) (0-1) global_enable_wol:3c59x: same as enable_wol, but applies to all NICs if enable_wol is unset rx_copybreak:3c59x copy breakpoint for copy-only-tiny-frames max_interrupt_work:3c59x maximum events handled per interrupt compaq_ioaddr:3c59x PCI I/O base address (Compaq BIOS problem workaround) compaq_irq:3c59x PCI IRQ number (Compaq BIOS problem workaround) compaq_device_id:3c59x PCI device ID (Compaq BIOS problem workaround) watchdog:3c59x transmit timeout in milliseconds global_use_mmio:3c59x: same as use_mmio, but applies to all NICs if options is unset use_mmio:3c59x: use memory-mapped PCI I/O resource (0-1) When I use modinfo on the skge and sky2 modules no such options are listed, so it seems they don't offer software control over this capability: % modinfo -p sky2 debug:Debug level (0=none,...,16=all) copybreak:Receive copy threshold disable_msi:Disable Message Signaled Interrupt (MSI) idle_timeout:Watchdog timer for lost interrupts (ms) % modinfo -p skge debug:Debug level (0=none,...,16=all) Randall Schulz -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tuesday 19 December 2006 13:48, John Andersen wrote:

On Tuesday 19 December 2006 13:31, Randall R Schulz wrote:

...

John,

On Tuesday 19 December 2006 14:09, John Andersen wrote:

...

On Tuesday 19 December 2006 08:11, Randall R Schulz wrote:

...

ASUS P5B Deluxe Core 2 Duo 32-bit Linux install

How did That happen? Core 2 is x86_64 by default...

Says who? What "default?"

Well, Randall, since your past responses suggest you are a big wikipedia fan, check here:

http://en.wikipedia.org/wiki/Core_2_duo

Its a 65bit processor. It will emulate 32bit.

Doh! 64bit of course. Extra bit inserted by wayward digits. -- _____________________________________ John Andersen

On Tuesday 19 December 2006 14:08, Randall R Schulz wrote:

If you don't need a 64-bit architecture, you definitely should not use one. I don't need one.

so why did you buy one? And now that you've discovered, (but not yet admitted) that you installed the wrong kernel, why not do it over and prove or disprove your elaborate theory? -- _____________________________________ John Andersen

On Tuesday 19 December 2006 14:28, Randall R Schulz wrote:

Why on earth must you be so rude?

Pot/Kettle. But it wasn't my intent.

And how, exactly, is the kernel that's running my system just fine, "wrong?"

Its not the native kernel for that hardware. As for your theory, a quick look at the clock times will reveal that the 64bit registers load in the same number of clocks as the 32. And the 64bit address registers, pointer registers, etc also load in the same number of clocks as the 32 bit registers. So moving stuff around in memory is faster doing so in 64bit chunks. The reason for 64bit processors is NOT simply address space. Drive controllers move data to memory at what ever speed and chunk size they are designed to use, but your access to that data in 32bit chunks that take just as long to load as 64bit chunks is slower because it takes more clocks. Look, I've done this exact same thing, loaded the i386 kernel on a Pentium D (also a 64bit dual well processor, in spite of the Pentium name). Since it was a new server with no mission data installed I re-installed using the correct kernel, and it was significantly faster with the 64bit kernel. I did not bench long running applications, or time long compiles, just running yast, kde, and some database loading. Since it runs at init 3 most of the time the 32bit would have worked, and might have gone un-noticed. But it was faster launching applications in 64bit, and the time to load the sql database from raw files was also faster, 37 minutes compared to 52. -- _____________________________________ John Andersen

On 2006-12-19 18:38, Randall R Schulz wrote:

On Tuesday 19 December 2006 15:49, John Andersen wrote:

...

<snip> <snip> *ahem*, guys....

-- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Tuesday 19 December 2006 15:38, Randall R Schulz wrote:

...

The reason for 64bit processors is NOT simply address space.

If one has no need for 64 bit anything, it's all a waste.

Not really. Its faster. How can that be a waste?

...

Drive controllers move data to memory at what ever speed and chunk size they are designed to use, but your access to that data in 32bit chunks that take just as long to load as 64bit chunks is slower because it takes more clocks.

Driver controllers are DMA. The processor is not involved in each individual data word transfered.

True, but not relevant to the point of my paragraph. Moving data in memory in 64 bit chunks is faster than moving in 32 bit chunks.

...

Look, I've done this exact same thing, loaded the i386 kernel on a Pentium D (also a 64bit dual well processor, in spite of the Pentium name).

Why would results for an ancient Pentium-D imply anything for a Core 2 Duo, which is an entirely different implementation of the x86 architecture?

Penitum-D is not ancient. It was only announced in 2005 and the Presler 915 was released in the 1st quarter of 2006. Its a curret processor with emt64 and it works faster with a x86_64 kernel. http://en.wikipedia.org/wiki/Pentium_D

Which leaves us with the original accusation--that I made a 32/64 choice. I just stuck the disc in and installed.

Just exactly what disk did you stick in? Unless you have the boxed set (and I know you don't), you downloaded EITHER a 32bit OR a 64 bit distro. When downloading you can't get both on one DVD. So you made the choice at download time. The installer did not have the choice of picking the optimal kernel. Given a choice (such as when your boxed set arrives with its double layer disk) yast would choose the x86_64 kernel. Randall, run what you want, its your machine. The only reason I commented was you seemed unaware that the core 2 was a 64bit cpu. If you knew that already why not just say you consciously made a choice to stay with 32bits instead of getting all defensive about it? -- _____________________________________ John Andersen

On 2006-12-20 03:42, Verner Kjærsgaard wrote:

<snip> Now, what if the server side has a subtle bug that only manifests itself in certain, very specific situations?? If one reads through all of this thread, what conclusively excludes this possibility? Absolutely nothing.

-- The best way to accelerate a computer running Windows is at 9.81 m/s² -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org