<tasslehoff>
So, the usb benchmark thing I asked about the other day. We're thinking of creating a product where the physical connections will be usb and usb-otg. Then we may have to connect to all sorts of things, like usb-ethernet adapters, mhl, mass storage devices and android phones. I've been tasked with figuring out the actual bandwith we can expect from usb on the imx6.
<tasslehoff>
anyone got a) benchmarks or info on the actual speed I can expect. b) suggestions for tests I should run to get a fairly certain image of the performance on the usb ports?
paulk-collins has quit [Quit: Quitte]
FelixH has joined #imx6-dev
cnxsoft1 has joined #imx6-dev
cnxsoft has quit [Ping timeout: 240 seconds]
FelixH has quit [Quit: Leaving]
<rz2k_>
tasslehoff: performance is not your first priority, first priority would be getting info about how hard memory leaks/kernel crashes with your usb devices
<rz2k_>
example: company where I was working before on imx6 product now has problems in production env: NAND degrades significantly over time
<rz2k_>
like 2 weeks of worktime and fs dies, chip itself is ok (tested many times with all available tools)
aholler_ is now known as aholler
<tasslehoff>
rz2k_: I didn't quite understand that first sentence.
<rz2k_>
when planning usage of imx6 in such project as yours, you need to build up a test rig and test for crashes first, because you might end finding that one of usb devices you want to connect crashes the kernel
<rz2k_>
then think about how well it will perform
paulk-collins has joined #imx6-dev
<tasslehoff>
rz2k_: I see, and kinda agree. But if the performance is less than we need I can spare the testing of the devices, cause than we will need to find another solution.
<rz2k_>
for example, I had problems with one of the USB soundcards connected to imx6 for some time, turned out to be a bug in the kernel
<rz2k_>
spent many hours bisecting mainline until found a fix :)
<tasslehoff>
rz2k_: :)
<tasslehoff>
rz2k_: I don't really know which usb devices will be connected yet.
<rz2k_>
from my experience - imx6 as usb-hid device connected to a windows pc was slow, copying 350mb of data took 40 minutes.
<rz2k_>
(dont ask why someone needed to use usb-hid for such transfer :)
<rz2k_>
try to ask at meta-freescale, if I remember correctly, there was a package called imx-test, that had unit tests for every IP in imx6, so it might have something for usb
<rz2k_>
meta-freescale mailing list
<tasslehoff>
rz2k_: ah. good tip. thanks.
FelixH has joined #imx6-dev
FelixH has quit [Ping timeout: 255 seconds]
<aholler>
just use the fs-gadget and transfer a file to easily get an idea of speed
<aholler>
it's enough to play hd-movies ;)
<tasslehoff>
:)
FelixH has joined #imx6-dev
steeve has quit [Remote host closed the connection]
cnxsoft1 has quit [Quit: cnxsoft1]
cnxsoft has joined #imx6-dev
dv^laptop has joined #imx6-dev
cnxsoft has quit [Quit: cnxsoft]
FelixH has quit [Ping timeout: 244 seconds]
dv^laptop has quit [Client Quit]
dv^laptop has joined #imx6-dev
tasslehoff has quit [Quit: WeeChat 0.4.2]
paulk-collins has quit [Quit: Quitte]
steeve has joined #imx6-dev
FelixH has joined #imx6-dev
pdgendt has quit [Ping timeout: 265 seconds]
pdgendt has joined #imx6-dev
FelixH has quit [Quit: Leaving]
neooo has joined #imx6-dev
dv^laptop has quit [Remote host closed the connection]
TuTizz has quit [Quit: Quitte]
neooo has quit [Quit: Leaving]
steeve has quit [Remote host closed the connection]
steeve has joined #imx6-dev
paulk-collins has joined #imx6-dev
frsc has quit [Ping timeout: 252 seconds]
zaxari has quit [Quit: WeeChat 1.1.1]
zaxari has joined #imx6-dev
hste_ has quit [Remote host closed the connection]
hste_ has joined #imx6-dev
bfederau has quit [Remote host closed the connection]