Bugzilla Bug 59578

Attached is a new ebuild for ndiswrapper-0.9.  This fixes bug [58883]. 
Additionally, it fixes problems with /etc/modules.d/ndiswrapper incorrectly
calling loadndiswrapper; this fixes bug [58244].  Finally, it removes an
unneeded test, and adjust the comments to fit the new ebuild.

Reproducible: Always
Steps to Reproduce:
1.
2.
3.

Created an attachment (id=36858) [edit]
ebuild for ndiswrapper-0.9

that's bug 58244 and bug 58883 that are fixed by this ebuild.

Created an attachment (id=36890) [edit]
File to go in net-wireless/ndiswrapper/files.

This file is moved into /etc/modules.d/.  It is needed to alias the wifi
interfface to the ndiswrapper module.  This version removes the explicit call
to loadndisdriver, as it is no longer needed (and can causes problems).

*** Bug 58244 has been marked as a duplicate of this bug. ***

*** Bug 59553 has been marked as a duplicate of this bug. ***

*** Bug 58883 has been marked as a duplicate of this bug. ***

this ebuild is now in Portage... however I'd like it to get some testing before
it's unleashed on the masses and I would like to come up with a good
standardization for new users where they can put their Windows drivers and just
make the ebuild a bit more new user friendly rather then... Ok... it emerged...
now what!? So... the ebuild is in package.mask. Expect some changes without a
revision bump. Howver, it should be stable enough for prior users to use it.

If you want to test it out... edit /etc/portage/package.unmask and add...
=net-wireless/ndiswrapper-0.9

if it's successfully working for you, post here. Any changes/ideas/comments are
welcome here. Any changes to the ebuild please submit a diff against the one
posted here rather then a full ebuild.

Last but not least... don't bug latexer about this... bug me.

It works for me using A Dell Latitude D600 Laptop with the Truemobile 1350 card
(802.11g).  I've variously used drivers from HP, Broadcom, eMachines, and Dell;
usually with similar results and performance (according to the ndiswrapper
developers, it's the same basic driver).

Would it make more sense to move this to ~x86 for more testing, intest of
having to explicitly unmask it?  Otherwise, only people specifically interested
in it (e.g. those motivated enough to search bugs.gentoo.org) are going to
actually test it.

For user configuration, a few comments/suggestions:

* Add a note directing the users to
http://ndiswrapper.sourceforge.net/supported_chipsets.html to obtain
appropriate drivers.  Running "lspci -n | grep 'Class 0280'" should help them
find it.  Unless someone writes a full-blown wifi chip detector/driver
downloader, this is likely the best solution.  The installation instructions
for ndiwrapper are pretty simple as well.

* Once the drivers are "installed" by ndiswrapper you don't really need to keep
the original .exe/.zip files around; the required files are all copied to
/etc/ndiswrapper/*.

* If we do want a place to store the windows drivers permenantly, how about
/usr/lib/win32?  There's already a bunch of .dll and other Windows files tucked
away in there; just make an "ndiswrapper" directory, and put the files in that.

Made enough changes to the ebuild to make me happy enough to unmask it.

Closing this bug out. If you have any other problems with ndiswrapper... please
file them as new bugs.