20 Nov
2015
20 Nov
'15
3:28 a.m.
Hi,
On Nov 18 2015 23:17, Takashi Iwai wrote:
On Sun, 15 Nov 2015 10:26:00 +0100, Takashi Sakamoto wrote:
- err = spkr_volume_command(oxfw, &oxfw->volume_min,
- if (strcmp(oxfw->card->driver, "FireWave") == 0) {
spkr->mixer_channels = 6;
spkr->mute_fb_id = 0x01;
spkr->volume_fb_id = 0x02;
- }
- if (strcmp(oxfw->card->driver, "FWSpeakers") == 0) {
spkr->mixer_channels = 1;
spkr->mute_fb_id = 0x01;
spkr->volume_fb_id = 0x01;
- }
What's the merit of such explicit individual conditional over the constant table in the current implementation? The latter is more error-prone and simpler in general.
I'm also concerned about it. Yes, the usage of 'struct ieee1394_device_id.driver_data' is nicer than thse condition statements, in this point.
My intension of a part of this patch series is to enclose model-dependent parameters inner model-dependent files, instead of adding module-public structure. This idea, itself, is not so bad, I think.
There may be better ways to detect models and assign to structure but I don't still find it.
Thanks
Takashi Sakamoto