And that's with cpu capped to 1,4 instead of 1,7
Openssl benchmark is also impressive, triple the times of official fw:)
Maybe there is a quickfire command that I can issue to turn USB on? Seems like USB hub is detected from dmesg
And that's with cpu capped to 1,4 instead of 1,7
Openssl benchmark is also impressive, triple the times of official fw:)
Maybe there is a quickfire command that I can issue to turn USB on? Seems like USB hub is detected from dmesg
Joined: 03 Jan 2010 Posts: 7568 Location: YWG, Canada
Posted: Fri May 13, 2016 21:43 Post subject:
does tx rate reporting work on this ath10k fw? & after a few hours is the 5ghz interface still fine to associate with or does it become very slow & difficult to connect? _________________ LATEST FIRMWARE(S)
BrainSlayer wrote:
we just do it since we do not like any restrictions enforced by stupid cocaine snorting managers
Seems it's working ok so far, no problems yet
Btw, considering compiling options, I noticed that almost every firmware is compiled with Os optimization and from my experience using O3 gives about 10-15% performance boost (even with entware packages).
Isn't it the same case with firmware? Why no one compiles it with O3 flag? I guess there's enough room on flash for that. Maybe there is a reason?
And there's just a cosmetic thing - on status page it shows ipq8064
Edit: seems the problem with WiFi persists, after 9 hours of uptime couldn't reconnect to WiFi with devices that left the WiFi area for a while. Those that were connected all the time keep working fine
Seems it's working ok so far, no problems yet
Btw, considering compiling options, I noticed that almost every firmware is compiled with Os optimization and from my experience using O3 gives about 10-15% performance boost (even with entware packages).
Isn't it the same case with firmware? Why no one compiles it with O3 flag? I guess there's enough room on flash for that. Maybe there is a reason?
One thing that I'm aware of is that debugging highly optimized code is difficult. For example, if you sent a stack dump to the developer, it would be much easier use an analyzer and see what's going on if the code is not highly optimized. Usually developers don't go to the highest level of optimization, strip symbols out of object modules, etc., until the final, release builds.
Noticed that kernel gets tainted with code 32 just for your info. Happens only when WiFi is enabled.
Guess it's nothing to worry about?
Btw is there any workaround for 5ghz WiFi issue?
does tx rate reporting work on this ath10k fw? & after a few hours is the 5ghz interface still fine to associate with or does it become very slow & difficult to connect?
Joined: 03 Jan 2010 Posts: 7568 Location: YWG, Canada
Posted: Mon May 16, 2016 21:20 Post subject:
<Kong> wrote:
tatsuya46 wrote:
does tx rate reporting work on this ath10k fw? & after a few hours is the 5ghz interface still fine to associate with or does it become very slow & difficult to connect?
Next build come with updated backports, maybe it helps.
Will upload new builds after I ran a few tests.
from what i see its a issue with ath10k fw 10.2.4 & 10.4.1/10.4.3, maybe qualcomm finally fixed it with 10.4-3.2 _________________ LATEST FIRMWARE(S)
BrainSlayer wrote:
we just do it since we do not like any restrictions enforced by stupid cocaine snorting managers