-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Christian Iversen wrote: | On Wednesday 05 January 2005 12:58, Pedro Venda wrote: | |>David Brownell wrote: |>| On Tuesday 04 January 2005 4:31 am, Pedro Venda wrote: |>|>David Brownell wrote: |>|>| On Monday 03 January 2005 5:05 pm, you wrote: |>|>|>David Brownell wrote: |>|>|>| It's something wierd that started a while back, and so far |>|>|>| it doesn't make sense to me. But see below, and please |>|>|>| try the patch I've attached. If it fails, please send |>|>|>| the full CONFIG_USB_DEBUG output showing the error (as |>|>|>| below) and contents of /sys/class/usb_host/usb4/registers |>|>|>| after the fault. |>|>|> |>|>|>I've tried the patch and it failed. |>| |>| OK, try this slightly modified version. Looks like |>| the Intel chip sets the HALT bit then spontaneously |>| clears it, while the ALI may never set it. |> |>this one has a different output, although I don't understand if it's |>working properly since there are still some fatal errors... there isn't a |>fatal error but a BOGUS error, as introduced by the patch. |> |>[...output...] |> |>EHCI 1.00, hcd state 1 instead of the old value 0 means that the driver's |>working now? | | | Have you tried it? I use a powered hub (with an LED) to tell if ehci is on. If | you have a removeable hard drive or dvd-drive, you should notice a huge | difference in throughput.
I have no powered hub nor hard or dvd drives, at least usb. sorry, I can't test it right now. :-(
|>care to comment? is this a good enough patch for the whole "usb community" |>or does it simply cover this isolated bogus error and by itself is unsafe |>for other environments? | | | Allow me to comment on this. I'm the guy who found out that the controller | reports a bogus fatal error, and I made the first patch that fixed it. When | David Brownell saw my hack (which was to comment out a function call that | resets the chip in the error handling), he came up with the patch you just | tried, which is a more right way to handle it. | | I've been running with this patch for some time now (well, at least 10 days), | and I haven't had any problems of any kind. USB just seems to work for me | now. Of course, that could be pure luck, but since David and I looked at | which changes have been going on in the ehci-driver, and the usb-core in | general at the revisions where it originally broke, we discovered that it | _no_ direct changes were made in the ehci-driver in the breaking revision.
So the driver was working correctly and stopped working... I see.
| That suggests to me that the chip doesn't entirely work within specs, that the | driver is good, and that we need this patch to catch a corner case that is | wrongly handled by the hardware. That could be wrong, of course, but that's | my view of the situation :) | | (it's kind of perplexing that different controllers seem to have somewhat of | the same problem, but I think time will finally tell what's going on here)
yes, manufacturers sometimes tend to create these kinds of problems for one reason or another.
thanks for the comments. will the patch be included in the driver? or will it be tested further in usb devel trees?
regards, pedro venda. - --
Pedro Joćo Lopes Venda email: [EMAIL PROTECTED] http://arrakis.dhis.org
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.6 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFB3AkAeRy7HWZxjWERAiwCAKC0xDST4bnyQN7Hl/XidgnVzV5q+gCggCkn jdpN+DNEWMOk4BaBRmA1upo= =7Kcs -----END PGP SIGNATURE-----
------------------------------------------------------- The SF.Net email is sponsored by: Beat the post-holiday blues Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek. It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
