Add support for SteelSeries Arctis 7. #16
Conversation
src/devices/steelseries_arctis7.c
Outdated
| // the range of the Arctis 7 seems to be from 0 to 0x12 (18) | ||
| num = map(num, 0, 128, 0x00, 0x12); | ||
|
|
||
| unsigned char *buf = calloc(31, 1); |
There was a problem hiding this comment.
Good catch.
Regarding memcpy vs. memmove: I try to avoid memcpy since its potential benefit (increased speed potentially through prefetching) outweigh the risk (not safe with overlapping regions). Especially in this case it won't make any difference, neither runtime-related nor behavioral.
|
|
||
| if (num) | ||
| { | ||
| memmove(buf, data_on, sizeof(data_on)); |
There was a problem hiding this comment.
memmove is usually used when there is overlapping memory, normally I would use memcpy, but it doesn't really matter much
|
Also I tested wether it still works with my Headset, there doesn't seem to be any problem and I also guess that there won't be with other ones. Otherwise the interface id's of the other headsets must be adapted. |
Ionic
force-pushed
the
feature/support-steelseries-arctis7
branch
from
b834421 to
64e2f85
Compare
|
Updated the branch. Some more cleanup, less code duplication etc. |
Touches on Sapd#8, but doesn't fix it. Should be backwards-compatible if all device structures are nulled (trivially true for static variables like here) and enumerating the HID devices returns the default one that would be opened via hid_open().
Currently only supports setting (or disabling) the sidetone feature.
Ionic
force-pushed
the
feature/support-steelseries-arctis7
branch
from
64e2f85 to
b6f6f4e
Compare
|
Err... I don't get why git sometimes changes more commits than would be needed, so threw in a quick branch rebase. |
|
Thank you, I merged it into master |
…tis7 Add support for SteelSeries Arctis 7.
This currently only supports setting (or disabling) the sidetone feature.
Sadly, this control station falls out of line by using Output Requests instead of Feature Requests, so the code is a tad different from all the other implementations.
Further, the request needs to be sent to a non-default interface, which touches on #8. Thus, I've implemented initial interface opening, which SHOULD be backwards compatible due to initializing the device structure statically (and thus nulling the memory block) and assuming that
hid_enumeratereturns the default interface as the first one - which would be whathid_open(vid, pid)would normally open.I was not able to actually test this, though, so please go ahead and make sure it doesn't break your devices.