Hi Larry,

Thanks for a detailed post with your findings. We understand your
frustration here. But I have a strange issue in my L440 which has R8192EE
Wifi Adapter and 3.17.0-031700rc6_3.17.0-031700rc6.201409211935 kernel
which is the latest available kernel. I installed this kernel few days back
and tried with new driver from github repo. I don't find any errors in the
new kernel make log (as you were mentioning in the above post - which I saw
in default 3.13 kernel). But after installing this kernel and new driver, I
am still not able to connect to Wifi networks - It tries several times and
fails to connect. I can see following messages repeatedly on dmesg while it
tries connecting.

[   41.078474] wlan0: authenticate with b0:a8:6e:40:38:02
[   41.111422] wlan0: direct probe to b0:a8:6e:40:38:02 (try 1/3)
[   41.314257] wlan0: direct probe to b0:a8:6e:40:38:02 (try 2/3)
[   41.518483] wlan0: direct probe to b0:a8:6e:40:38:02 (try 3/3)
[   41.722704] wlan0: authentication with b0:a8:6e:40:38:02 timed out
[   41.875292] wlan0: authenticate with ac:4b:c8:67:da:82
[   41.891899] wlan0: send auth to ac:4b:c8:67:da:82 (try 1/3)
[   41.995006] wlan0: send auth to ac:4b:c8:67:da:82 (try 2/3)
[   42.099117] wlan0: send auth to ac:4b:c8:67:da:82 (try 3/3)
[   42.203231] wlan0: authentication with ac:4b:c8:67:da:82 timed out

I can see the following device driver details using lshw -C "Network"
command:

 *-network
       description: Wireless interface
       product: RTL8192EE PCIe Wireless Network Adapter
       vendor: Realtek Semiconductor Co., Ltd.
       physical id: 0
       bus info: pci@0000:02:00.0
       logical name: wlan0
       version: 00
       serial: 34:23:87:24:72:87
       width: 64 bits
       clock: 33MHz
       capabilities: bus_master cap_list ethernet physical wireless
       configuration: broadcast=yes driver=r8192ee
driverversion=3.17.0-031700rc6-generic firmware=N/A latency=0 link=no
multicast=yes wireless=IEEE 802.11bgn
       resources: irq:30 ioport:5000(size=256) memory:f2400000-f2403fff

Do you or anybody facing similar issue? If yes, can you please help me out
in figuring cause of this issue. I am clueless and missing out something
here. I already feeling like I chose a wrong OS Ubuntu.

For curiosity I have one more question - Does this new hardware RTL8192
works on older version of Ubuntu or any other Linux distro (probably
desktop/lite version) for that matter? Which one you safely bet? I am ready
to migrate to any other version of Linux which can fix this issue for ever
!!

Thanks in advance,
Venkatesha T R


On Thu, Sep 25, 2014 at 9:45 PM, Larry Finger <email address hidden>
wrote:

> I am beginning to understand this problem, but I do not have a general
> fix.
>
> In kernel 3.14, the authors of mac80211 changed the API (applications
> interface) for routine ieee80211_is_robust_mgmt_frame() to be able to
> intercept a certain kind of error. For those drivers that could not fit
> the new model, which includes the Realtek codes, they created a new
> entry point that used the old method. It was named
> _ieee80211_is_robust_mgmt_frame(), i.e. it has an underscore. For out-
> of-kernel drivers such as those in the rtlwifi_new repo, there must be
> code that determines at compile time which API needs to be used. Such
> code is in these drivers. If the kernel version is 3.14 or newer, use
> the underscored version. If older use the one without.
>
> Why does it fail here? The reason is that Ubuntu applied the API change
> to their 3.13 kernels! Now the decision gets harder. If the kernel is
> 3.14, or if this is Ubuntu with version 3.13.11 + some unknown patch
> level, use the new API. That gets quite difficult; however, if I take
> the simple step of having all 3.13 versions use the new API, then I
> break every user of 3.13 that does not run Ubuntu. I refuse to do that
> for the following reason. Ubuntu should not have changed the API in mid
> 3.13!! In doing so, they make this kind of problem for everyone trying
> to provide out-of-kernel drivers. For me to fix the problem for Ubuntu
> users by breaking every other distro's users would be effectively to
> reward Ubuntu for what I consider to be their bad behavior.
>
> How do I know this is the problem. First of all, the make log that was
> posted shows "warning: passing argument 1 of
> ‘ieee80211_is_robust_mgmt_frame’ from incompatible pointer type [enabled
> by default]".  It is true that warnings can usually be ignored; however,
> when arguments have an incompatible pointer type, the end result will
> usually be a kernel crash. When that warning occurs for a routine that
> has changed API, it really sticks out. Secondly, I see this change in
> Ubuntu's 3.11.0-36 patch file.
>
> What can Ubuntu users do? To my thinking, switching to a 3.14 or newer
> kernel would be best. If you really want to stay with Ubuntu's 3.13.0-36
> kernel, then run the following command in the rtlwifi_new directory:
>
> sed -i
> 's/ieee80211_is_robust_mgmt_frame/_ieee80211_is_robust_mgmt_frame/g'
> rtl8*/trx.c
>
> That will change the code to the new API for every kernel. Of course,
> that will break every kernel older than 3.13, and the command should be
> run *ONLY* if you see warnings of the kind noted above.  To revert that
> change, you will need to run 'git checkout -f'.
>
> Sorry that it has taken so long to sort this out, but I really never
> expected the API to change in the middle of kernel 3.13 for any distro.
> If the tone seems to be  too bitchy, remember that I'm really angry
> about what has happened, and all the time I have wasted on sorting it
> out. My comments here are a lot milder than the way I really want to say
> them.
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1239578
>
> Title:
>   No wireless support for Realtek RTL8192EE [10ec:818b]
>
> Status in HWE Next Project:
>   Confirmed
> Status in HWE Next trusty series:
>   Confirmed
> Status in “linux” package in Ubuntu:
>   Confirmed
>
> Bug description:
>   Open this bug to trace device driver RTL8192EE.
>
>   We have seen this wireless card on some laptops, id is [10ec:818b],
>   driver is RTL8192EE.
>
>   Status: waiting for this driver being moved out of staging directory.
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/hwe-next/+bug/1239578/+subscriptions
>



-- 

*Thanks and Regards*
*--------------------------------------------------------------------------------------------------------------------*
*Venkatesh T R*
*9945040858*
A Man with Many Dimensions...