FS#44841 - [filesystem] remove usb-load-ehci-first.conf

Attached to Project: Arch Linux
Opened by Tom Yan (tom.ty89) - Monday, 04 May 2015, 07:48 GMT
Last edited by Doug Newgard (Scimmia) - Wednesday, 13 May 2015, 21:45 GMT
Task Type General Gripe
Category Packages: Core
Status Closed
Assigned To Dave Reisner (falconindy)
Tom Gundersen (tomegun)
Architecture All
Severity Very Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:
The file is not approriate anymore since now there's also the xhci modules. I can't even trace back the reason it exists. Probably a workaround for something that have been fixed long ago. Besides there's many other case that a user might need a softdep, there's no reason for arch include one for only this case.

Additional info:
* package version(s)
* config and/or log files etc.


Steps to reproduce:
This task depends upon

Closed by  Doug Newgard (Scimmia)
Wednesday, 13 May 2015, 21:45 GMT
Reason for closing:  None
Additional comments about closing:  Requested by reporter.
Comment by Dave Reisner (falconindy) - Monday, 04 May 2015, 14:49 GMT
You probably shouldn't recommend removal of things when you don't understand why they were added. In this case, the softdep exists because ohci_hcd and uhci_hcd would bind to devices which were capable of running at faster speeds, but wouldn't, because ehci_hcd wasn't yet available. If you'd like to propose that this file be removed, please actually test that this is no longer the case.

1) Benchmark USB transfer speeds on USB2.0 and USB 3.0 capable devices.
2) Remove the rule from your machine (and initramfs), reboot.
3) Repeat step 1.
4) Report your findings.
Comment by Tom Yan (tom.ty89) - Monday, 04 May 2015, 16:02 GMT
   usb_modules (4.4 KiB)
Comment by Dave Reisner (falconindy) - Monday, 04 May 2015, 16:18 GMT
Start here: https://lkml.org/lkml/2013/9/10/371

If you have further questions/concerns, please take it up with linux-modules@vger.kernel.org or the kernel folk.
Comment by Tom Yan (tom.ty89) - Monday, 04 May 2015, 17:53 GMT
Comment by Dave Reisner (falconindy) - Monday, 04 May 2015, 18:02 GMT
If you're concerned with the conf file, then you're concerned with the softdeps. They're intimately related. Suggesting otherwise is lunacy. To quote your own original report:

> Probably a workaround for something that have been fixed long ago.
You must actually prove that such a problem no longer exists before I'll be convinced that we should get rid of it. I've given you the problem that it solves. You haven't shown me any evidence that it's been fixed somewhere else.
Comment by Tom Yan (tom.ty89) - Monday, 04 May 2015, 19:37 GMT

Loading...