I've a notebook with a RTL8188CE Half PCI-e Wireless card. It can connect succesfully in all my wifi (different routers), but it can't sustain tx with one of them, a TP-Link WR1043ND rev 1.x, wich has Atheros AR9103 wifi. With Windows 8 my wireless card works just fine. I've been reading that there's a issue with this driver (rtl8192se) with kernel higher than 3.9. One friend of mine has another notebook with a RTL8192se (rtl8192se driver also) wich behaves in the same way. I can provide any other info that's needed, just point me in the right direction.
There are new versions of all the Realtek rtlwifi drivers that I am getting ready for submission to the kernel. To test these, git clone http://github.com/lwfinger/rtlwifi_new.git cd rtlwifi_new make sudo make install Next reboot. For the make to work, you will need the kernel header package from your distro.
Ok, just for the record, it's actually rtl8192ce, not rtl8192se driver I stated. On the other hand, rtlwifi drivers from that git repo hang my notebook, with a kernel panic the moment I enabled the hardware switch. The compilation did throw some warnings but it get completed. I'll retry all the process after my notebook shutdown: the batery is internal, no power key response, no sysreq key, no nothing :( Also, which kernel versions is this driver compatible with? Fedora (my actual distro) it's currently using 3.15
This code is supposed to be for kernels 3.0 and newer. It also should not generate any warnings. Please post them.
Created attachment 145631 [details] Compilation output (make) This is the output from make.
Created attachment 145641 [details] Compilation output (make install) This is the output from 'make install'
Created attachment 145651 [details] gcc version GCC version, just in case
Created attachment 145661 [details] Kernel Panic And the kernel panic after restart and hardware switch on
I found the problem. Please pull, make, and install again. Please put text attachments as blabla.txt. If you do that, the browser can display them directly. As blabla.bin, they have to be dwnloaded and opened with a text editor.
(In reply to Larry Finger from comment #8) > Please put text attachments as blabla.txt. If you do that, the browser can > display them directly. As blabla.bin, they have to be dwnloaded and opened > with a text editor. Ok, sorry about that, I'll keep that in mind next time. No more kernel panic, so, I'll be testing this next two or three days. But it has already sent 300 icmp pings without stopping :) Thanks for your fast response, and work!
I've been testing with my problematic wifi connection, and it seems to be working ok. At 2nd and 3rd try, the tx would behave in the same way I had described in the bug report but only after I've had switch off-on the hardware key, till a restart. The dmesg output show something related to the hardware key in the rtl module, but I can't reproduce it anymore. Now, restarting, suspending, hardware switch on-off-on, rfkill block-unblock, and the connection remains pretty stable. Only in one try I've got 30 ping packages lost in a 8000 row (which I can't attribute only to the wifi driver). In the next day or so, I'll try in the other notebook I've mentioned earlier (RTL8192SE) and will report back. That being said, this bug could be closed unless there's another reason to maintain it open.