#asahi-dev · 2024-01-14 11:10
jannegI was wondering why u-boot's truetype console was displaying cyan tinted text until I remembered that's how xrgb8888 white looks when rendered as xrgb2101010#asahi-dev · 2024-01-13 17:53
jannegfor some reason though only in the efi console and not in u-boot's console#asahi-dev · 2024-01-13 17:52
jannegand now with cursor support: https://github.com/jannau/u-boot-1/tree/asahi-vidconsole-fixes#asahi-dev · 2024-01-11 20:58
jannegno need to test anything. there is no cursor#asahi-dev · 2024-01-11 20:58
jannegoh, I forgot the link to my github branch: https://github.com/jannau/u-boot-1/tree/asahi-vidconsole-fixes#asahi-dev · 2024-01-11 20:53
jannegno, unrelated to HiDPI. Simon enabled the cursor only for expo#asahi-dev · 2024-01-11 20:49
jannegHiDPI font support is still a hack and not upstreamable, only missing feature is cursor support#asahi-dev · 2024-01-11 20:48
jannegmarcan: u-boot asahi-releng rebased on asahi / v2024.01 with most EFI_SIMPLE_TEXT_OUTPUT_PROTOCOL issues fixed#asahi-dev · 2024-01-11 20:16
jannegkettenis: is there a reason for "stdout=serial,vidconsole"? efi_console will only take the size from vidconsole if stdout starts with "vidconsole". changing the order fixes make the efi text console display filled instead of 80x25#asahi-dev · 2024-01-06 17:06
jannegmaz: yes, reload#asahi-dev · 2024-01-06 16:36
jannegmaz: https://gist.github.com/jannau/079ce2d5a35fedcdaad7bd9e71bac9c9#asahi-dev · 2024-01-06 12:09
janneginteresting, macos 13.6.3 is installable on M3#asahi-dev · 2024-01-06 11:54
jannegwhy apple? "parent_addr = 0x0000000210000000" from t8122 (m3) main MMIO range#asahi-dev · 2024-01-06 11:53
jannegkettenis: fixup commits in a PR are the best way to ensure it's not forgotten#asahi-dev · 2024-01-06 10:24
jannegkettenis: pull request would work but still depends on marcan to merge it. I might end up including it in my work tree to avoid conflicts#asahi-dev · 2024-01-04 14:30
jannegleio: FCVTN/FCVTN2#asahi-dev · 2024-01-04 14:21
jannegm1 has half-preciosion float (5-bit exponnent)#asahi-dev · 2023-12-30 16:10
janneglevel triggered and disabling the interrupt after it has triggered and reenable it when dcp notices the link is gone#asahi-dev · 2023-12-30 16:01
jannegI'm not sure if we need the detect pin at all. as it is now a level triggered irq would be enough#asahi-dev · 2023-12-30 15:50
jannegI think I started with just the gpio to check pin and didn't thought much about it#asahi-dev · 2023-12-30 15:47
jannegI don't remember#asahi-dev · 2023-12-30 15:44
jannegI interpret block as either a bunch of continuously data or possibly functional blocks and not as blocking#asahi-dev · 2023-12-30 15:40
jannegbecause it has an unknown and annoying interface: uint32_t set_block_dcp(arg1=uint64_t, arg2=uint, arg3=uint, arg4=Array(8, ulong), arg5=uint, data=SizedBytes(0x1000, "length"), length=ulong, unknArry=Array(4, uint))#asahi-dev · 2023-12-30 15:27
jannegthere are also several other gamma related calls in the DCP/iomfb endpoint. search for gamma in proxyclient/hv/trace_dcp.py#asahi-dev · 2023-12-30 15:26
jannegyes, macOS supports gamma LUTs and at least one of the nightshift macOS apps still works and uses that. that results in annoying set_block_dcp calls on the DCP endpoint#asahi-dev · 2023-12-30 15:11
jannegbut the better option would be to investigate how to set gamma LUTs with DCP#asahi-dev · 2023-12-30 15:09
jannegyes, won't work with that protocol unless you look at the gamma lut and check if it can be represented by a CTM which is not necessarily the case#asahi-dev · 2023-12-30 15:02
jannegwhy not? does it just has an interface to set the gamma lut?#asahi-dev · 2023-12-30 14:58
jannegyou have to start before the gamma lut. ctm would need to the same input as the code which computes the gamma lut#asahi-dev · 2023-12-30 14:45
jannegcan you check with drm_info if the CTM property is still set after exiting? if yes, you have to remove it, if no it's likely a driver bug#asahi-dev · 2023-12-30 14:43
janneg1<<32, 0, 0, 0, 1<<32, 0, 0, 0, 1<<32#asahi-dev · 2023-12-30 14:42
jannegyes, (sign << 63) | (integer << 32) | fraction#asahi-dev · 2023-12-30 14:40
jannegS31.32 is a sign bit, 31-bit integer and 32-bit fractional#asahi-dev · 2023-12-30 14:39
jannegj`ey: I think the drm client is supposed to reset properties on exit#asahi-dev · 2023-12-28 21:38
jannegthere are 4 cd3217 instances but I think the 4th one is magsafe since that should still use usb pd#asahi-dev · 2023-12-28 21:33
jannegI don't think so#asahi-dev · 2023-12-28 21:20
jannegI think atcphy does already the same. that's used to get usb2, usb3 and dp phys from the single atcphy, see atcphy_xlate()#asahi-dev · 2023-12-28 21:07
jannegbut since I have the phy_set_mode call anyway to turn dptx-phy on that was the most abvious hack to make configure atcphy#asahi-dev · 2023-12-28 21:06
jannegnot sure what the best solution for the dp-only phy is. I was also thinking about using the the typec_switch/mux directly to configure atc-phy to dp-only in normal orientation#asahi-dev · 2023-12-28 20:56
jannegI still need to find a better solution for that#asahi-dev · 2023-12-28 20:56
jannegignore the dcp->index that is required for dptx-phy since that has an integrated mux to select the dcp source#asahi-dev · 2023-12-28 20:54
jannegthere is a `phy_set_mode_ext(dptx->atcphy, PHY_MODE_DP, dcp->index);` in dptxep.c#asahi-dev · 2023-12-28 20:53
jannegcan't we do the mode switching I did from dcp via phy_set_mode() conditionally #asahi-dev · 2023-12-28 20:50
jannegsven: yes, both mac studios use atcphy3 for usb-c (on the first die)#asahi-dev · 2023-12-27 11:08
jannegstarting now#asahi-dev · 2023-12-27 10:09
jannegstream will be here https://streaming.media.ccc.de/37c3/eins#asahi-dev · 2023-12-27 10:07
jannegtalk on how apple silicon usb-c access applies to the iphone 15 in 55 minutes at 37C3: https://events.ccc.de/congress/2023/hub/en/event/apple_s_iphone_15_under_the_c/#asahi-dev · 2023-12-26 23:14
jannegthanks#asahi-dev · 2023-12-26 23:07
jannegvadorovsky: hej, if the patch applies to upstream u-boot please send it to the u-boot ml, if not please create a github pr#asahi-dev · 2023-12-26 19:00
jannegkettenis: see https://github.com/AsahiLinux/linux/commit/00f0043a5d78919cf14aee3999592b66792115d4