| Summary: | kmod-wifi fails dependency tests | ||
|---|---|---|---|
| Product: | Fedora | Reporter: | Steve Hayter <steve.hayter> |
| Component: | madwifi-kmod | Assignee: | Nicolas Chauvet <kwizart> |
| Status: | RESOLVED INVALID | ||
| Severity: | normal | CC: | fedora |
| Priority: | P1 | ||
| Version: | 10 | ||
| Hardware: | All | ||
| OS: | GNU/Linux | ||
| namespace: | |||
|
Description
Steve Hayter
2009-04-02 23:07:08 CEST
That kernel is available, it was pushed yesterday: https://www.redhat.com/archives/fedora-package-announce/2009-April/msg00065.html Your yum just chose a not-up2date mirror. The whole problem is explained in https://www.redhat.com/archives/fedora-devel-list/2008-August/msg00041.html Just wait and it'll solve itself BTW: please reminds that ath5k (within kernel) will obsoletes madwifi starting with 2.6.29 kernel. You will need to remove it, and use ath5k/ath9h instead (which currently support most chipsets). (In reply to comment #2) > BTW: please reminds that ath5k (within kernel) will obsoletes madwifi starting > with 2.6.29 kernel. You will need to remove it, and use ath5k/ath9h instead > (which currently support most chipsets). Note that I don't agree with dropping madwifi in a stable release (e.g. F9 or F10) if there is no full replacement ("most chipsets" indicates ath5k/ath9k is none yet; seems the AP/MasterMode support also is not that stable yet). (In reply to comment #3) > (In reply to comment #2) > > BTW: please reminds that ath5k (within kernel) will obsoletes madwifi starting > > with 2.6.29 kernel. You will need to remove it, and use ath5k/ath9h instead > > (which currently support most chipsets). > > Note that I don't agree with dropping madwifi in a stable release (e.g. F9 or > F10) if there is no full replacement ("most chipsets" indicates ath5k/ath9k is > none yet; seems the AP/MasterMode support also is not that stable yet). > There is nothing to discuss in this area. Inkernel 2.6.29 ath5k "IS" a full replacement and better driver than madwifi (monitor mode doesn't need patching , and master mode is fully supported), Futhermore inkernel 2.6.29 ath9k is now the reference implementation against madwifi. If dropping madwifi by F-10 could have been a little too earlier wrt newer chipset support, I 've always said that it was supposed to be droped sooner or later. Actually -- the "bug" was entirely my own doing. I did a yum clean all; yum update and it did in fact "fix itself". My bad. Thanks for the suggestions though. And, FYI, I actually had to roll back to the madwifi driver because the kernel driver, frankly, sucks. Using the embedded kernel driver, my throughput drops to 4Mbps, the connection itself continually drops and reconnects, and overall system performace suffers -- when the connection drops, the entire system freezes until the connection is reestablished. Blacklisting the kernel module and using madwifi resolves all of the issues and gives me solid throughput of 48Mbps. I've posted several comments to this effect on several forums over the past few months, and, as is normally the case with something "new and better", no one has ever responded to it. The one bug report I opened at Fedora was subsequently closed with no comments. PulseAudio is just as crappy, yet, apparently, I can't be having issues with it either because it's "newer and better" too... Steve (In reply to comment #5) > Actually -- the "bug" was entirely my own doing. I did a yum clean all; yum > update and it did in fact "fix itself". My bad. > > Thanks for the suggestions though. > > And, FYI, I actually had to roll back to the madwifi driver because the kernel > driver, frankly, sucks. please test with kernel 2.6.29. There is no interested in feedback with kernel lower than this. (report problem on Fedora only, but you can put the link here also). Once kernel 2.6.29 is here I will disable the dependency on the kmod from the madwifi package. |