| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 00:10 | |
| josch | vagrantc, minute: this patch might fix the linux build on 7.1.5: https://source.mnt.re/reform/reform-debian-packages/-/merge_requests/192 | 00:25 |
|---|---|---|
| josch | minute: do you want to rebase MR 188 on top of that and/or cherry pick the commit from there so that your branch builds the most recent kernel? | 00:27 |
| - murphnj (QUIT: Quit: murphnj) (~murphnj@user/murphnj) | 00:40 | |
| - Kooda (QUIT: Quit: WeeChat 4.4.4) (~kooda@natsu.upyum.com) | 00:54 | |
| + Kooda (~kooda@natsu.upyum.com) | 00:56 | |
| + robin__ (~robin@user/terpri) | 01:18 | |
| - robin_ (QUIT: Ping timeout: 260 seconds) (~robin@user/terpri) | 01:23 | |
| + oliverD (~Thunderbi@user/oliverd) | 01:45 | |
| - oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd) | 01:48 | |
| vagrantc | josch: thanks, will check! | 02:07 |
| vagrantc | josch: well, it applies, presuming the kernel doesn't fail to build ... looks good. :) | 02:25 |
| - AnimaInvicta (PART: !!unknown attribute: msg!!) (~AnimaInvi@88-169-191-86.subs.proxad.net) | 02:52 | |
| * robin__ -> robin | 03:05 | |
| + oliverD (~Thunderbi@user/oliverd) | 03:09 | |
| - paperManu (QUIT: Ping timeout: 265 seconds) (~paperManu@204.48.0.160) | 03:15 | |
| - oliverD (QUIT: Ping timeout: 257 seconds) (~Thunderbi@user/oliverd) | 03:18 | |
| + colinsane (~colinunin@97-113-130-182.tukw.qwest.net) | 03:19 | |
| + oliverD (~Thunderbi@user/oliverd) | 03:31 | |
| - oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd) | 03:33 | |
| + oliverD (~Thunderbi@user/oliverd) | 03:42 | |
| - oliverD (QUIT: Read error: Connection reset by peer) (~Thunderbi@user/oliverd) | 03:43 | |
| - kop316 (QUIT: Remote host closed the connection) (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:04 | |
| + kop316 (m-6f6zq6@static.138.159.90.157.clients.your-server.de) | 04:05 | |
| - leonardo_alberto (QUIT: Quit: Client closed) (~leonardo_@2a0c:5a86:130c:9a00:d9fa:7e77:9b50:72c0) | 04:18 | |
| - paperManu_ (QUIT: Ping timeout: 265 seconds) (~paperManu@204.48.0.160) | 04:29 | |
| - Gooberpatrol66 (QUIT: Remote host closed the connection) (~Gooberpat@user/gooberpatrol66) | 04:50 | |
| stephano | josch: sorry, I meant to message about 7.1.5, but today's work day was nuts. I have a fix here: https://github.com/cetola/mnt-build/blob/main/xtra-patches/mnt-overrides/7.1/qcom-mnt-reform2/4004-qcom-pci-add-d3cold.patch | 04:58 |
| stephano | The fix is in mainline, 7.2-rc, 2cc0e7454c78 | 05:03 |
| + robin_ (~robin@user/terpri) | 05:18 | |
| - robin (QUIT: Ping timeout: 260 seconds) (~robin@user/terpri) | 05:21 | |
| stephano | Oh, and by "I have a fix", I mean "I copied what they did in 7.2". So my fix is mostly ctrl-c followed by a cunning use of ctrl-v. | 05:35 |
| - colinsane (QUIT: Ping timeout: 259 seconds) (~colinunin@97-113-130-182.tukw.qwest.net) | 05:39 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50) | 05:53 | |
| + mjw (~mjw@S0106b0f5306f19f3.vc.shawcable.net) | 06:00 | |
| - n_to (QUIT: Quit: quitidiquit) (~n_to@2a03:4000:6:3662:24b1:57ff:fec6:76c1) | 06:50 | |
| + n_to (~n_to@2a03:4000:6:3662:24b1:57ff:fec6:76c1) | 06:50 | |
| + qbit_ (~qbit@user/qbit) | 07:10 | |
| - qbit (QUIT: Ping timeout: 256 seconds) (~qbit@user/qbit) | 07:12 | |
| * qbit_ -> qbit | 07:12 | |
| + oliverD (~Thunderbi@user/oliverd) | 07:14 | |
| josch | stephano: thank you, your solution is better. I used my own power of the copypasta and added you as the git author in MR 192 if that's okay? | 07:14 |
| - oliverD (QUIT: Ping timeout: 253 seconds) (~Thunderbi@user/oliverd) | 07:18 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 08:13 | |
| + RandyK (~RandyK@user/randyk) | 08:14 | |
| - mjw (QUIT: Ping timeout: 265 seconds) (~mjw@S0106b0f5306f19f3.vc.shawcable.net) | 08:14 | |
| - marty (QUIT: Ping timeout: 245 seconds) (~marty@static-23-234-100-217.cust.tzulo.com) | 08:16 | |
| + marty (~marty@static-23-234-101-23.cust.tzulo.com) | 08:18 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 08:19 | |
| + RandyK (~RandyK@user/randyk) | 08:20 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 08:27 | |
| + RandyK (~RandyK@user/randyk) | 08:27 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 08:28 | |
| + RandyK (~RandyK@user/randyk) | 08:29 | |
| - se6astian (QUIT: ) (~quassel@2a03:b0c0:3:f0:0:1:9e0a:7000) | 08:37 | |
| + BAndiT1983 (~quassel@2a03:b0c0:3:f0:0:1:9e0a:7000) | 08:37 | |
| + se6astian (~quassel@2a03:b0c0:3:f0:0:1:9e0a:7000) | 08:37 | |
| + anuejn (~quassel@2a03:b0c0:3:f0:0:1:9e0a:7000) | 08:37 | |
| + oliverD (~Thunderbi@user/oliverd) | 08:47 | |
| + robin__ (~robin@user/terpri) | 08:49 | |
| - robin_ (QUIT: Ping timeout: 260 seconds) (~robin@user/terpri) | 08:52 | |
| josch | minute: I saw that you removed comments from ./linux/config. Currently, linux/build.sh just does `cat config >>linux/debian/config/arm64/config` which we could easily replace by `grep -v '#' config >> ...` and then linux/config can have comments. Okay? | 09:29 |
| - oliverD (QUIT: Ping timeout: 251 seconds) (~Thunderbi@user/oliverd) | 10:08 | |
| minute | josch: ah, sounds good, thanks! | 10:10 |
| minute | stephano: thanks for finding the pcie patch! | 10:11 |
| josch | minute: I rebased your rk3588-pocket-dp-altmode-and-sc branch on top of MR 192 with stephano's patch in it. The built just finished in josch/reform-debian-packages and now a rk3588 pocket reform system image with that kernel is building in https://source.mnt.re/josch/reform-system-image/-/pipelines/6485 | 10:16 |
| josch | Once that is done I'll boot that on my pocket and see what happens. Then, I'll fash the new firmware from https://source.mnt.re/reform/pocket-reform/-/merge_requests/86 The latest commit is from one month ago. That's the latest state? | 10:20 |
| minute | josch: hmm no the latest commit is from july 18 | 10:36 |
| josch | ah, good i asked i was looking at the wrong branch, thank you! | 10:38 |
| + oliverD (~Thunderbi@user/oliverd) | 10:48 | |
| josch | I/O error, dev mmcblk0, sector 1072 op 0x1:(WRITE) flags 0x4800 phys_seg 63 prio class 2 | 11:13 |
| josch | Buffer I/O error on dev mmcblk0, logical block 74, lost async page write | 11:13 |
| josch | mmc0: error -84 whilst initialising SD card | 11:13 |
| josch | meson-gx-mmc ffe05000.mmc: exceeding card's volts | 11:13 |
| josch | sd-cards falling like flies here... | 11:13 |
| josch | there are sd-cards i can buy that go for 15 cents per GB. Unfortunately that doesn't mean that i can buy a bag of 8 GB cards for cheap... :( | 11:17 |
| minute | ohh | 11:29 |
| josch | hrm.... :) https://mister-muffin.de/p/Ui5q.jpg | 11:33 |
| josch | but when gdm started the display was fine again | 11:34 |
| - lrvick (QUIT: Quit: WeeChat 4.6.3) (~lrvick@2600:1700:5430:a94f::1e80) | 11:36 | |
| josch | sysctl and keyboard are still on the old firmware. With the new kernel, battery shows 0% and /sys/bus/spi/drivers/reform2_lpc misses the spi1.0 directory | 11:39 |
| josch | (reform2_lpc is loaded) | 11:39 |
| josch | in dmesg i have multiple "mnt_sc spi1.0: mntsc: timeout 1000" messages (expected on the old firmware i guess) | 11:39 |
| josch | and then power_supply BAT0: driver failed to report `status' property: -16 which explains the 0% battery | 11:40 |
| josch | oh this gets spammed a lot | 11:40 |
| josch | minute: my dmesg has over 2k timeout messages with mnt_sc on the old firmware | 11:41 |
| josch | right, so /sys/class/power_suply/BAT0/ exists but accessing its content gives EBUSY | 11:47 |
| josch | whoops, i am missing usb device 2e8a:000a | 11:51 |
| josch | yay i have the new sysctl firmware and i can see it in lsusb again :) | 12:12 |
| josch | minute: my battery gauge works again \o/ | 12:14 |
| josch | hrm... except that it immediately jumps to 100% when i plug it in agai (but reform-power-daemon recognizes the plug-in/out events) and i can now read /sys/class/power_suply/BAT0/ again | 12:14 |
| josch | lets discharge it for a while and see if the battery gauge reports start making sense | 12:15 |
| minute | josch: the rotation stuff is because of the phantom dp display i think | 12:18 |
| minute | josch: all these errors are expected | 12:18 |
| josch | nice | 12:19 |
| josch | meh... battery percentage is stuck at 94% it seems :( | 12:19 |
| minute | josch: except the 100% thing | 12:19 |
| josch | minute: it was the same problem on the old firmware -- maybe some hardware is borked for me? | 12:19 |
| minute | josch: mhm? that sounds like a gauge thing. could reset the gauge by unplugging the batteries | 12:19 |
| minute | josch: and plugging them back in. | 12:19 |
| minute | josch: although you have nonstandard cells right? so we could walk through implementing the setting for the higher capacity | 12:20 |
| josch | minute: would toggling the standby switch not have a similar effect? | 12:20 |
| josch | minute: yes, i have non-standard cells but this used to work in the past | 12:20 |
| josch | wait... by unplugging the batteries, do you mean unplugging the cells or unplugging the battery board? | 12:21 |
| + murphnj (~murphnj@user/murphnj) | 12:31 | |
| + paperManu (~paperManu@204.48.0.160) | 12:54 | |
| minute | josch: no, toggling the standby switch does not have that effect. | 12:56 |
| minute | josch: unplugging the cells. | 12:56 |
| minute | josch: because the cells power the gauge chip | 12:56 |
| josch | Then running the cells down to zero should have a similar effect? | 13:05 |
| josch | I unplugged both cells and then replugged them. | 13:05 |
| josch | Now battery is stuck at 88% when AC power is not connected. | 13:06 |
| josch | Well, it's probably just a me-problem because of the custom cells. No big deal. | 13:06 |
| - cli (QUIT: Ping timeout: 263 seconds) (~m-vsauiy@user/cli) | 13:10 | |
| + cli (~m-vsauiy@user/cli) | 13:13 | |
| minute | i've now rebased my alt-mode branch on linux7.1.5 branch | 13:14 |
| minute | josch: running the cells to zero might not have the same effect. only if the cells themselves both shut off their voltage | 13:15 |
| minute | josch: we need to inform the SC about the larger cell capacity | 13:15 |
| minute | josch: could you try booting with this DTB? i've enabled the display reset gpio, driven by mnt-sc, in this one http://minute.social/secret-room/rk3588-mnt-pocket-reform.dtb | 13:16 |
| josch | sure, sec | 13:18 |
| minute | (also testing that one now, but i'm on mb2) | 13:20 |
| minute | panel still works for me at least | 13:20 |
| minute | on reboot, i have the weird stripey mode in barebox, but it gets fixed when linux starts up | 13:22 |
| josch | minute: i'm using u-boot (as it's still the default) here | 13:22 |
| josch | with your dtb, display still works | 13:22 |
| minute | ah very good | 13:22 |
| minute | internal display (v2) still fades out when toggling it off in gnome, but comes on again cleanly | 13:24 |
| minute | hmm i'm not seeing a dev_info() output with the gpio name in dmesg though | 13:26 |
| minute | oh damn, there's a piece of code missing to actually register the gpio controller m)) | 13:30 |
| minute | josch: FYI i've now overwritten that .dtb with a debug version for myself, don't download again | 13:38 |
| josch | thanks for the heads-up! | 13:43 |
| josch | i see that you pushed more things to the rk3588-pocket-dp-altmode-and-sc branch -- i'll upgrade to a kernel built from that branch once you tell me that it'd make sense to do so | 13:47 |
| minute | *bugfixing intensifies* | 13:47 |
| minute | josch: yes, mntsc driver's gpio stuff is still borked, fixing atm | 13:47 |
| - oliverD (QUIT: Ping timeout: 261 seconds) (~Thunderbi@user/oliverd) | 13:47 | |
| josch | Okay, i'm now off to child-caring duties -- back in 6 hours or so! | 13:48 |
| josch | minute: you could apply this patch and then put back the comments in linux/config: https://paste.debian.net/hidden/2691052b | 13:50 |
| minute | josch: thx! see ya! | 13:58 |
| - paperManu (QUIT: Ping timeout: 260 seconds) (~paperManu@204.48.0.160) | 14:30 | |
| + paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 14:38 | |
| minute | annoying, gpio_ext_enable(), when called via SPI interface, can crash the sysctl fw | 15:14 |
| minute | hm, I2C_TIMEOUT is 500*1000 us, that's 500 ms?! | 15:16 |
| minute | i guess that's not ok to call in a 5ms timer interrupt | 15:17 |
| minute | yeah that seems to be related... uff uff | 15:19 |
| minute | ah no we're nowadays on an edge-triggered interrupt | 15:23 |
| - gidzit (QUIT: Ping timeout: 264 seconds) (~gidzit@82-64-82-48.subs.proxad.net) | 15:28 | |
| minute | aha, triggering on rising edge of SS0 works better than on falling edge | 15:56 |
| stephano | josch: Thanks! The attribution in MR 192 works. minute: Glad the PCIe patch was useful. As I update Arch, I'll post those upstream finds here next time. | 15:58 |
| - aperezdc (QUIT: Remote host closed the connection) (~aperezdc@2a03:6000:6e61:633::43) | 16:20 | |
| + aperezdc (~aperezdc@46.23.89.43) | 16:21 | |
| - RandyK (QUIT: Remote host closed the connection) (~RandyK@user/randyk) | 17:30 | |
| + RandyK (~RandyK@user/randyk) | 17:31 | |
| - marty (QUIT: Ping timeout: 248 seconds) (~marty@static-23-234-101-23.cust.tzulo.com) | 17:36 | |
| + marty (~marty@146.70.171.155) | 17:38 | |
| + AnimaInvicta (~AnimaInvi@88-169-191-86.subs.proxad.net) | 18:47 | |
| - paperManu (QUIT: Ping timeout: 253 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 19:25 | |
| - S0rin (QUIT: Quit: WeeChat 4.6.3) (~S0rin@user/s0rin) | 19:26 | |
| + S0rin (~S0rin@user/s0rin) | 19:38 | |
| + mjw (~mjw@S0106b0f5306f19f3.vc.shawcable.net) | 19:42 | |
| minute | ok, this build's kernel/mntsc driver toggles the panel's reset gpio successfully via mntsc https://source.mnt.re/reform/reform-debian-packages/-/jobs/23439 | 19:45 |
| minute | but it makes a WARN at drivers/gpio/gpiolib.c:4010 for some reason | 19:45 |
| minute | ohh | 19:45 |
| minute | /* Should be using gpiod_set_value_cansleep() */ | 19:45 |
| minute | WARN_ON(desc->gdev->can_sleep); | 19:46 |
| minute | kinda weird though, this means that this has to be patched in the panel driver | 19:47 |
| - mjw (QUIT: Ping timeout: 252 seconds) (~mjw@S0106b0f5306f19f3.vc.shawcable.net) | 19:59 | |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50) | 20:08 | |
| + paperManu (~paperManu@bras-base-mtrlpq0776w-grc-27-65-94-77-245.dsl.bell.ca) | 20:17 | |
| stephano | minute: https://community.mnt.re/t/pocket-screen-off-after-idle-and-cant-get-on-again/4304/2 | 20:40 |
| stephano | that "cansleep" warning has been around for a while | 20:40 |
| minute | stephano: ahh interesting. | 20:42 |
| minute | stephano: i've now changed the cansleep to false in the mntsc driver, but i can also revert that and change it in the panel driver instead... the whole thing feels a bit weird though | 20:43 |
| stephano | minute: oh, I assumed it was set to cansleep on purpose on the mntsc. :) | 20:44 |
| stephano | minute: You can certainly set the RP2040 to "cansleep = false" assuming you never use that feature. Probably smart to flag that somewhere though in case some day someone wants to set SLEEP or DORMANT modes. | 20:50 |
| minute | stephano: those are unrelated. the "sleep" is a kernel specific context thing, i.e. when in an IRQ handler call tree, i think you can't use msleep()... etc functions | 20:52 |
| minute | https://docs.kernel.org/timers/delay_sleep_functions.html#delay-and-sleep-mechanisms | 20:54 |
| minute | also called "atomic" context | 20:54 |
| minute | https://lwn.net/Articles/274695/ | 20:55 |
| minute | > When the kernel is running in process context, it is allowed to go to sleep if necessary. But when the kernel is running in atomic context, things like sleeping are not allowed. Code which handles hardware and software interrupts is one obvious example of atomic context | 20:55 |
| - paperManu (QUIT: Ping timeout: 251 seconds) (~paperManu@bras-base-mtrlpq0776w-grc-27-65-94-77-245.dsl.bell.ca) | 20:56 | |
| stephano | minute: ooooohhhhh okay. wow I was confused. | 20:57 |
| minute | funnily, a very similar problem bit me _in_ the rp2040 code | 20:57 |
| minute | a longer time ago, when i discovered that msleep() should not be used in irq handlers in pico-sdk, as they'll make it crash | 20:57 |
| minute | that's when i converted some stuff to udelay(), in the keyboard fw and sysctl fw code | 20:58 |
| minute | ah not udelay, it's called busy_wait_us() there | 20:58 |
| stephano | I see: src/sysctl.c:43:// called from timer interrupt, no sleep allowed here!... src/spi_com.c:29: - no sleep_ms() calls | 20:59 |
| minute | yes! | 21:00 |
| minute | and the linux kernel has similar constraints | 21:00 |
| + paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 21:08 | |
| minute | another kernel package to test... https://source.mnt.re/reform/reform-debian-packages/-/jobs/23445 | 21:08 |
| josch | minute: anything special i should look out for? | 21:08 |
| josch | and would you like me to test that kernel package with the new firmware or with the old? | 21:09 |
| minute | josch: oh one sec, let me test it first :D | 21:09 |
| minute | interesting, the panel doesn't seem to come up anymore in linux | 21:14 |
| minute | ah i'm sillyh | 21:16 |
| minute | deleted too much when patching the patch, kconfig+makefile change missing now | 21:17 |
| minute | hmmm weird, with the panel driver loaded manually as a module now, i get dw-mipi-dsi2 fde30000.dsi: command interface is busy | 21:27 |
| minute | hm, but "rmmod panel_mnt_pocket_reform" works without crashing | 21:29 |
| minute | it doesn't clean up the sysfs_create_group though! dirty | 21:30 |
| ^alex | who's Colonel Module | 21:38 |
| minute | uff everything behaves totally differently when the panel is loaded as a module on rk3588 (vs compiled in)....... | 21:56 |
| + Gooberpatrol66 (~Gooberpat@user/gooberpatrol66) | 22:20 | |
| - talos (QUIT: Ping timeout: 260 seconds) (~talos@2600:6c5d:f0:1fd0:fd40:6f77:566d:1388) | 22:25 | |
| minute | *extreme mnt-pocket-reform-panel driver hacking in progress* | 22:43 |
| - exark (QUIT: Ping timeout: 265 seconds) (~exark@user/exark) | 23:18 | |
| + talos (~talos@2600:6c5d:f0:1fd0:e9d6:603:e5d:a18f) | 23:22 | |
| - murphnj (QUIT: Remote host closed the connection) (~murphnj@user/murphnj) | 23:22 | |
| - paperManu (QUIT: Ping timeout: 260 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca) | 23:50 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
