≡
[15:11:40] <dddddd> I guess that setting up audio conflicts with serial-jack use. I'm wondering how that situation is handled. Will see...
[15:17:59] <jo0nas> dddddd: flipping serial-jack is the "gpio set PL9" mentioned at https://en.opensuse.org/HCL:Teres-I
[15:26:09] <dddddd> I imagine this: serial-jack working until the kernel setups audio, then serial is lost.
[15:27:00] <dddddd> So, if serial is the only way to login, you can't.
[15:33:02] <dddddd> Worst situation: u-boot starts using the kernel DT and serial is unavailable quite early. One always can prepare custom boot, if it's insterested in serial-jack, but works by default now ;)
[16:01:28] <jo0nas> dddddd: what are you talking about? Did you actually read the page I shared?
[16:02:51] <dddddd> yes """The USB keyboard is slow to come up, and the audio jack is connected to the serial console by default. Add usb stop; usb start; and gpio set PL9 somewhere in the flow of automatically executed scripts."""
[16:03:48] <dddddd> in the context of 'no automated image creation'
[16:09:07] <dddddd> serial by default, but if it's used later by the kernel for audio (via pl12 if I'm reading the patch correctly) the signals would be mixed, what am I missing?
[16:12:54] <dddddd> Also, the wording in the suse page is confusing. If serial is on by default then, gpio set does enable for real the serial, or they are trying to teach how to avoid the serial?
[16:24:13] <jo0nas> dddddd: here is the detailed explanation: https://patchwork.kernel.org/patch/10792589/
[16:32:17] <dddddd> the "best solution" allows to grab logs, but not login (if audio driver is loaded, which you can't unload because you can't login).
[16:50:28] <jo0nas> dddddd: if you have good ideas then I suggest you share them at the mailinglist where it is discussed: http://lists.infradead.org/pipermail/linux-arm-kernel/2019-February/629541.html
[16:52:29] <dddddd> Thanks for the pointers; I'm just informally commenting, sorry for the noise
[16:56:22] <diego71> dddddd: you can blacklist the audio driver in that case
[16:57:30] <dddddd> sure, or not compiling it in the first place, or change dts...
[17:00:29] <diego71> the blacklist is more quick, and can be reverted easily
[17:00:58] <dddddd> Obviously, for the final user it's better to default to audio... which is a change for everybody (specially for those of us tinkering with the thing)
IRC #olimex 2019-05-28
[15:02:24] <dddddd> xcko, if just got into -next, it's for the 5.3 merge window, AFAIU. Thanks for the link![15:11:40] <dddddd> I guess that setting up audio conflicts with serial-jack use. I'm wondering how that situation is handled. Will see...
[15:17:59] <jo0nas> dddddd: flipping serial-jack is the "gpio set PL9" mentioned at https://en.opensuse.org/HCL:Teres-I
[15:26:09] <dddddd> I imagine this: serial-jack working until the kernel setups audio, then serial is lost.
[15:27:00] <dddddd> So, if serial is the only way to login, you can't.
[15:33:02] <dddddd> Worst situation: u-boot starts using the kernel DT and serial is unavailable quite early. One always can prepare custom boot, if it's insterested in serial-jack, but works by default now ;)
[16:01:28] <jo0nas> dddddd: what are you talking about? Did you actually read the page I shared?
[16:02:51] <dddddd> yes """The USB keyboard is slow to come up, and the audio jack is connected to the serial console by default. Add usb stop; usb start; and gpio set PL9 somewhere in the flow of automatically executed scripts."""
[16:03:48] <dddddd> in the context of 'no automated image creation'
[16:09:07] <dddddd> serial by default, but if it's used later by the kernel for audio (via pl12 if I'm reading the patch correctly) the signals would be mixed, what am I missing?
[16:12:54] <dddddd> Also, the wording in the suse page is confusing. If serial is on by default then, gpio set does enable for real the serial, or they are trying to teach how to avoid the serial?
[16:24:13] <jo0nas> dddddd: here is the detailed explanation: https://patchwork.kernel.org/patch/10792589/
[16:32:17] <dddddd> the "best solution" allows to grab logs, but not login (if audio driver is loaded, which you can't unload because you can't login).
[16:50:28] <jo0nas> dddddd: if you have good ideas then I suggest you share them at the mailinglist where it is discussed: http://lists.infradead.org/pipermail/linux-arm-kernel/2019-February/629541.html
[16:52:29] <dddddd> Thanks for the pointers; I'm just informally commenting, sorry for the noise
[16:56:22] <diego71> dddddd: you can blacklist the audio driver in that case
[16:57:30] <dddddd> sure, or not compiling it in the first place, or change dts...
[17:00:29] <diego71> the blacklist is more quick, and can be reverted easily
[17:00:58] <dddddd> Obviously, for the final user it's better to default to audio... which is a change for everybody (specially for those of us tinkering with the thing)
