Home | Docs | Forums | Lists | Bugs | Planet | Store | GMN | Get Gentoo!
View Bug Activity | Format For Printing | XML | Clone This Bug
This is an ebuild for at76c503, which has support for atmel USB wireless devices. The ebuild has only been tested up to downloading the package. Reproducible: Always Steps to Reproduce: 1. 2. 3.
Created an attachment (id=48684) [edit] contains ebuild, manifest, and digest
According to the website the 2.6 kernel requires a version greater than 0.11 (Currenly only CVS as 0.11 is the most recent release). The cvs version has worked for me in the past, maybe a cvs snapshot would be better than 0.11. It might be good to know what has been holding the next release back.
Can you please attach the ebuild as a plain text file? We dont need digests and manifests .. they can be easily created.
Created an attachment (id=48850) [edit] The ebuild attached as a text file
Yes, you are right. I forgot that the source I downloaded was from CVS. I need to modify the ebuild, I will do the testing this upcoming weekend to sort this out.
If you plan to update the ebuild, it would be nice if you could rewrite it and take another ebuild that uses "inherit linux-mod" as base, such as cloop or ipw2100 or madwifi-driver or orinoco or bcm4400 (~x86 ebuilds).
Hey Stefan, Thanks for pointing that out, I will use that as a base so it knows that it is installing modules. I will be writing the cvs version as well for 2.6 kernels. Walter
Created an attachment (id=49000) [edit] updated ebuild containing cvs snapshot download ability
I get this error when trying to merge the ebuild: /usr/lib/portage/bin/ebuild.sh: line 1560: /home/walterw/portage/net-wireless/at76c503/at76c503-20040119.ebuild: Permission denied Please let me know what needs to be fixed to make this ebuild work as it is my first and I hope to do some more for webapps and such. Walter
Created an attachment (id=49125) [edit] at76c503a-0.12_beta19.ebuild please test and comment
Maybe we need adep on atmel-firmware?
Created an attachment (id=49158) [edit] cvs module works, cannot make/make install just yet
Hi Stefan, Thanks for your quick reply. So the ebuild actually works better on my machine than my server. I can get as far as getting the source from cvs, but make gives me errors, and it is copying files to the wrong places. It should have a kernel version in between modules and build. I suspect this should be a simple fix and once we figure it out, it should compile and install okay. >>> Source unpacked. mkdir -p .tmp_versions cp /lib/modules/./build/.tmp_versions/*.mod /var/tmp/portage/at76c503a-cvs-20050121/work/at76c503a/.tmp_versions cp: cannot stat `/lib/modules/./build/.tmp_versions/*.mod': No such file or directory make: [modules] Error 1 (ignored) make -C /lib/modules/./build SUBDIRS=/var/tmp/portage/at76c503a-cvs-20050121/work/at76c503a MODVERDIR=/var/tmp/portage/at76c503a-cvs-20050121/work/at76c503a/.tmp_versions \ EXTRA_CFLAGS=" -DCOMPILE_FIRMWARE_INTO_DRIVER" modules make: *** /lib/modules/./build: No such file or directory. Stop. make: *** [modules] Error 2 !!! ERROR: net-wireless/at76c503a-cvs-20050121 failed. !!! Function src_compile, Line 59, Exitcode 2 !!! emake failed !!! If you need support, post the topmost build error, NOT this status message.
Created an attachment (id=49160) [edit] first, success build Using linux_mod functions, the module does now compile and install correctly. May need to do some testing with this though.
Created an attachment (id=49161) [edit] updated driver, fixed path information When installing the modules, this warning is reported: >>> Install at76c503a-cvs-20050121 into /var/tmp/portage/at76c503a-cvs-20050121/image/ category net-wireless * Installing at76_usbdfu module * Preparing file for modules.d... [ ok ] * Installing at76c503-i3861 module /usr/lib/portage/bin/ebuild.sh: line 1: ${#MODULESD_AT76C503-I3861_ALIASES[*]}:bad substitution /usr/lib/portage/bin/ebuild.sh: line 143: [: ==: unary operator expected * Preparing file for modules.d... [ ok ] * Installing at76c503-i3863 module /usr/lib/portage/bin/ebuild.sh: line 1: ${#MODULESD_AT76C503-I3863_ALIASES[*]}:bad substitution /usr/lib/portage/bin/ebuild.sh: line 143: [: ==: unary operator expected * Preparing file for modules.d... [ ok ] * Installing at76c503-rfmd-acc module /usr/lib/portage/bin/ebuild.sh: line 1: ${#MODULESD_AT76C503-RFMD-ACC_ALIASES[*]}: bad substitution /usr/lib/portage/bin/ebuild.sh: line 143: [: ==: unary operator expected * Preparing file for modules.d... [ ok ] * Installing at76c503-rfmd module /usr/lib/portage/bin/ebuild.sh: line 1: ${#MODULESD_AT76C503-RFMD_ALIASES[*]}: bad substitution /usr/lib/portage/bin/ebuild.sh: line 143: [: ==: unary operator expected * Preparing file for modules.d... [ ok ] * Installing at76c503 module * Preparing file for modules.d... [ ok ] * Installing at76c505-rfmd module /usr/lib/portage/bin/ebuild.sh: line 1: ${#MODULESD_AT76C505-RFMD_ALIASES[*]}: bad substitution /usr/lib/portage/bin/ebuild.sh: line 143: [: ==: unary operator expected * Preparing file for modules.d... [ ok ] * Installing at76c505-rfmd2958 module /usr/lib/portage/bin/ebuild.sh: line 1: ${#MODULESD_AT76C505-RFMD2958_ALIASES[*]}: bad substitution /usr/lib/portage/bin/ebuild.sh: line 143: [: ==: unary operator expected * Preparing file for modules.d... [ ok ] * Installing at76c505a-rfmd2958 module /usr/lib/portage/bin/ebuild.sh: line 1: ${#MODULESD_AT76C505A-RFMD2958_ALIASES[*]}: bad substitution /usr/lib/portage/bin/ebuild.sh: line 143: [: ==: unary operator expected * Preparing file for modules.d... [ ok ]
sorry, I wont add it as cvs-module-ebuild because that will create a lot of load for the cvs-server, does the ebuild I attached, which uses the debain sources, work for you?
Hi Stefan, Is it possible that you can check out a version, say 20050124, and upload that to gentoo mirrors to reduce load and that would be used instead of cvs? It would spread the load over all the gentoo mirrors rather than one server. That way we are more certain the ebuild written will work, as if any internal changes are made to the code (at76c503a), we should change the ebuild and release a new ebuild for that cvs version. I didn't see the attached ebuild. Walter
http://bugs.gentoo.org/attachment.cgi?id=49125 Can you please test this ebuild?
Hi Stefan, I am at school and will have to wait till this weekend to test the ebuild. Do you put this stuff into portage, or what does the process involve? Thanks, Walter
as I do not own the hardware, I will put it in, when you have tested it.
Hi Stefan, I just ran the installer, and it works correctly for both emerging and unmerging the package. I did not get any errors. The only thing is for the driver location (although it doesn't matter), I put them in /lib/modules/${KV}/kernel/drivers/net/wireless/at76c503a/ instead of /lib/modules/${KV}/net. It works either way, but I guess there are reasons for putting it there as well. Walter
I added the ebuild to portage now, it is /net for the reason that it does not get overwritten if you recompile your kernel with a different .config
Hi Stefan, I did the testing on AMD64, so could you add it to ~amd64? Thanks, Walter
please open a new bug for that and assign it to amd64@gentoo.org unfortunately I am not allowed to do anything here, as I dont own an amd64 system. You are free to donate me one, if you want me to test that and add the keyword :)
