#
✕ nick:jannau
#asahi-dev · 2022-04-07 05:31
jannauto inestigate that further you have to look at dcp traces (hv/trace_dcp.py) from running macos under m1n1's hypervisor
#asahi-dev · 2022-04-07 05:30
jannaunametable[m]: there is no documention. I'm not sure if anyone has checked but the expectation is that backlight is controlled through DCP
#asahi-dev · 2022-04-07 05:28
jannaurqou_: problem seems to be "nvme-apple 393cc0000.nvme: ANS did not boot". I've seen nvme (and usb) failures with broken DCP but never looked closely at them since the problem was clearly DCP. Although I'm still wondering how DCP can break nvme
#asahi-dev · 2022-04-06 21:01
jannaurqou_: t600x is fixed, m1n1 didn't set the dart type
#asahi-dev · 2022-04-06 16:30
jannauGlanzmann: main ubuntu arm64 mirror is http://ports.ubuntu.com/
#asahi-dev · 2022-04-06 06:54
jannaurqou_: https://www.jannau.net/uboot_dcp_rsv_mem.diff would fix the u-boot problem, I'll look after *dcp-nb on t600x tonight
#asahi-dev · 2022-04-06 05:53
jannaukevans91: you could use the vmlinuz-asahi from https://cdn.asahilinux.org/aarch64/asahi/linux-asahi-5.17.rc7.20220310.5-2-aarch64.pkg.tar.xz it will not boot fully without modules/initrd but I guess you're just interested in whether it still boots
#asahi-dev · 2022-04-06 05:04
jannauI haven't tested it with u-boot myself yet, also only tested on t8103
#asahi-dev · 2022-04-06 05:01
jannaurqou_: not a problem with https://github.com/jannau/m1n1/tree/dcp-nb and https://github.com/jannau/linux/tree/asahi-dcp-nb it changes the way how the premapped memory regions are expressed in the fdt
#asahi-dev · 2022-04-05 21:48
jannaufixed
#asahi-dev · 2022-04-05 21:17
jannauthat's for bare metal, HV support is in m1n1/hv.py line 156ff
#asahi-dev · 2022-04-05 21:13
jannauhypervisor not well tested on t6002 means support for starting CPUs on the second die is missing
#asahi-dev · 2022-04-05 20:45
jannaualexsv: see https://github.com/jannau/m1n1/tree/dcp-nb and https://github.com/jannau/linux/tree/asahi-dcp-nb it changes the way how the premapped memory regions are expressed in the fdt, u-boot broke with the previous one
#asahi-dev · 2022-04-05 16:12
jannauarnd: instructions for running linux/u-boot as VM guest are here https://github.com/AsahiLinux/docs/wiki/m1n1%3AUser-Guide#running-a-linux-kernel-as-a-vm-guest-with-virtual-uart
#asahi-dev · 2022-04-05 16:10
jannauthen tools/run_guest.py in proxyclient can be used to execute m1n1 + payload from the usb connected computer
#asahi-dev · 2022-04-05 16:08
jannauah, yes. m1n1 needs to be standalone on the device without payload
#asahi-dev · 2022-04-05 16:08
jannauarnd: there should be two ttyACM devices when m1n1 is loaded
#asahi-dev · 2022-04-05 16:06
jannauarnd: debug port is the one next to the ethernet port but that should only matter for the hw uart
#asahi-dev · 2022-04-05 06:29
jannaumain cpu also maintains the iommu for display processor, dcp and disp0 both have their own dart
#asahi-dev · 2022-04-05 05:42
jannauI don't think we have documentation besides the code and the tracing implementation
#asahi-dev · 2022-04-05 05:40
jannauthere is also a dp-crossbar which probably needs to to be configured
#asahi-dev · 2022-04-05 05:40
jannau3. the usb-c ports have to be configured for displayport alt mode. that probaly needs to to build on on sven's USB3 work: https://github.com/AsahiLinux/linux/tree/atc-WIP with this m1n1 pull request https://github.com/AsahiLinux/m1n1/pull/170
#asahi-dev · 2022-04-05 05:36
jannauthe nodes for the external display(s) are called dcpextN and dispextN
#asahi-dev · 2022-04-05 05:35
jannauapple calls the nodes for the main display dcp and disp0 in their device-tree
#asahi-dev · 2022-04-05 05:35
jannauon the display processor side there seem to be no or nor huge difference between the main display adn external displays
#asahi-dev · 2022-04-05 05:33
jannauit will currently not work if you boot via u-boot (to be fixed very soon, probably tonight CEST)
#asahi-dev · 2022-04-05 05:32
jannauthe second part is dcp, there is working linux driver for the main display in https://github.com/jannau/linux/tree/asahi-dcp, it requires this m1n1 branch https://github.com/jannau/m1n1/tree/dcp
#asahi-dev · 2022-04-05 05:30
jannau1. macos under m1n1 hypervisor (https://github.com/AsahiLinux/docs/wiki/SW:Hypervisor), for the main display part we have hv/trace_dcp.py as tracing script
#asahi-dev · 2022-04-05 05:28
jannaualexsv: welcome. have you already started looking into it? For USB-C displayport alt mode there are 3 things you'll have to look at
#asahi-dev · 2022-04-04 18:43
jannaudcp working Thierry's latest iommu map reserved memory bindings proposal
#asahi-dev · 2022-04-04 17:46
jannauin m1n1 or linux? maybe iboot/m1n1 doesn't clear the irq after shutdown
#asahi-dev · 2022-04-03 13:10
jannaunot sure if it is wise to leave buffer partially initialized on errors. we check bfr->size in apple_rtkit_free_buffer()
#asahi-dev · 2022-04-03 13:05
jannausven: buffer->iova is not set in the dma_alloc_coherent() case in apple_rtkit_common_rx_get_buffer
#asahi-dev · 2022-04-03 13:00
jannauno, we have no way to decide that on the devicetree level. if anything m1n1 should remove the unneeded power domain
#asahi-dev · 2022-04-03 12:08
jannausigh, the new reserved memory/iommu mapping devicetree bindings proposed by Thierry are a pain to handle in m1n1
#asahi-dev · 2022-04-03 10:44
jannaudoubling it it means the encoder as to check twice as many frames whether which block is good predictor for current block
#asahi-dev · 2022-04-03 10:43
jannauI haven't looked at the interface apple offers but increasing the number of reference frames obviously makes the encoding more complex
#asahi-dev · 2022-04-03 10:40
jannauh264/h265 encoding complexity also depends massively on the encoding options
#asahi-dev · 2022-04-02 16:07
jannausven: or not although the error looks different: "nvme nvme0: Device not ready; aborting reset, CSTC=0x3"
#asahi-dev · 2022-04-02 15:56
jannausven: with the patch not reproducible so far, I'll test how fast I can reproduce it without those writel
#asahi-dev · 2022-04-02 15:02
jannausven: yes, with m1 max. I heven't really tried to reproduce it on the mac mini
#asahi-dev · 2022-04-02 14:00
jannauthere's a new comment, shmem_setup() has the iova argument for dcp
#asahi-dev · 2022-04-02 13:27
jannauit will not work for review comments sent to you alone. for everything else you could filter for messages with your email address in to: or cc:
#asahi-dev · 2022-04-02 13:18
jannauif you want to have mails completly separate you could use lei to retrieve them from the mailing list archives: https://people.kernel.org/monsieuricon/lore-lei-part-1-getting-started
#asahi-dev · 2022-04-02 09:47
jannauI've pushed updated m1n1 t6002-bringup, additional changes: skip initialization of interrupts on die 1 and cosmetic fixes
#asahi-dev · 2022-04-02 09:45
jannaumarcan: failed display init on the ultra was my fault, DVA was wrong for the ultra as it has 1 additional page mapped compared to the mac mini
#asahi-dev · 2022-04-01 18:24
jannauno problem
#asahi-dev · 2022-04-01 18:21
jannauand fdt_appendprop_{u32,u64,cell} for your convenience
#asahi-dev · 2022-04-01 18:20
jannauj`ey: fdt_appendprop()?
#asahi-dev · 2022-04-01 18:00
jannauyes, but you have care about endianness for multi-byte data

plus ›