You are not logged in.
I've been struggling for the past few hours to get Wi-Fi to work on my Compaq 615. After thoroughly following the Broadcom wireless guide the device still failed to show up on lspci. The PCI-ID / Chip ID is 14e4:4315 / BCM4312.
Then suddenly after restoring the BIOS defaults it showed up. This was only temporary however as a reboot made it disappear once again.
The lsmod diff showed a couple of differences in modules -- reloading BIOS added the modules bluetooth, btusb, ecb and lib80211_crypt_tkip. I tried adding them to my module-load.d/wl.conf without any noticable changes after rebooting.
rfkill list results:
Before BIOS reset:
0: hp-wifi: Wireless LAN
Soft blocked: yes
Hard blocked: no
...After BIOS reset:
0: phy0: Wireless LAN
Soft blocked: no
Hard blocked: no
1: hp-wifi: Wireless LAN
Soft blocked: yes
Hard blocked: no
...Now for the really odd part; if I run 'sudo rfkill unblock wifi' in the faulty state (without the device showing up on lspci), arch will start to output code in the terminal with terms such as add_worker, manage_workers, process, eventually freeze, then crash. After forcing a shutdown by holding down the power button, the next boot will provide working Wi-Fi drivers. Even more mysteriously, the rfkill unblock crash results in working drivers for one(?) additional reboot then disappears again. Forcing a shutdown by pressing the power button without crashing the system through rfkill unblock does not solve the device issue.
This really boggles my mind as a new Arch Linux (and Linux for that matter) user. I really can't wrap my head around what is going on. Although I've gotten my Wi-Fi to work, there's got to be a more efficient solution to my problem.
Thank you.
Last edited by SebN (2014-07-13 18:43:39)
Offline
Perhaps also compare the contents of the output of the 'dmesg' command?
What surprises me is the different modules being loaded after a bios reset.
I suggest you provide some output of dmesg after
- a bios reset
- after a faulty state
- after a forced 'good' state
We can use some regex magic and diff to compare them.
fs/super.c : "Self-destruct in 5 seconds. Have a nice day...\n",
Offline
I posted the dmesg outputs to my github: https://github.com/seb-nyberg/dmesg
Hopefully the regex magic can detect some abnormality!
By the way, I've noticed that the Wi-Fi working after a reboot is a result of unblocking it, not really related to crashing the computer. So, right now I could keep Wi-Fi working by making sure its unblocked at shutdown.
Offline
I did some quick scrolling:
dmesg_bios_reset + dmesg_forced_state uniquely contains
pci 0000:06:00.0: [14e4:4315] type 00 class 0x028000
pci 0000:06:00.0: reg 0x10: [mem 0x92000000-0x92003fff 64bit]
pci 0000:06:00.0: supports D1 D2
INFO @wl_cfg80211_attach : Registered CFG80211 phy
lib80211_crypt: registered algorithm 'TKIP'
eth0: Broadcom BCM4315 802.11 Hybrid Wireless Controller 6.30.223.141 (r415941)
wl 0000:06:00.0: no hotplug settings from platform (a lot...)
dmesg_bios_reset uniquely contains
pci_bus 0000:01: hash matches
dmesg_faulty uniquely contains
hp_wmi: query 0x5 returned error 0xea74You could consider blacklisting the HP WMI module. And maybel try to disable some virtualization options in your BIOS. I suggest you try these two options one by one. And, of course, try the other option *after* reverting the first one back to it's initial state.
fs/super.c : "Self-destruct in 5 seconds. Have a nice day...\n",
Offline
Blacklisting the hp_wmi module did the trick. rfkill now only shows the phy device but that doesn't really matter.
In case someone googles this issue later on -- I am using the broadcom-wl-dkms AUR package which automatically adds a blacklist conf to /etc/modprobe.d/. All I had to do was to add blacklist hp_wmi and things worked out great!
Thanks Rexilion ![]()
Offline
Why are you using an out of tree kernel module?
I googled your model, I think the native b43 driver should work
I also saw some hits about overheating problems. You might want to keep an eye on any temp sensors. However, this could not be related to Linux, but perhaps to bad/aging hardware??
fs/super.c : "Self-destruct in 5 seconds. Have a nice day...\n",
Offline