#
✕ nick:sven
#asahi · 2021-05-25 09:13
svenI don't understand that question
#asahi · 2021-05-25 09:02
svenor, well, you could give one port to Mac OS and one to m1n1
#asahi · 2021-05-25 09:02
svenno
#asahi · 2021-05-24 18:59
sveni think we'll eventually only need a phy driver
#asahi · 2021-05-24 18:50
svendo we really need the glue node?
#asahi · 2021-05-24 18:34
svenfair enough
#asahi · 2021-05-24 18:33
sveneven for host mode?
#asahi · 2021-05-24 18:33
svenhuh, really?
#asahi · 2021-05-24 18:30
svenkettenis: does u-boot take care of the usb phy init for host mode? then i can probably get away with just adding the dwc3 device nodes in the dart patchset so that it actually becomes useful for something
#asahi · 2021-05-24 18:29
svennow back to the DART
#asahi · 2021-05-24 18:28
svenalright, sent
#asahi · 2021-05-24 17:51
svenkettenis: https://f.svpe.de/f8ea95271af6bec833d5448efe87d2addca671993c9be1c92e8dcd464368bed7_0000-cover-letter.patch is that enough stressing that we really want to express the clock topology in the device tree or can you think of more points why this is a good idea that i should add?
#asahi · 2021-05-24 17:13
svenbut if you force it to use the full 64bit address space and set the two bypass bits in the dart TCR it'll work fine
#asahi · 2021-05-24 17:13
sventhe linux driver ignores dma-ranges and only uses 32bit addresses by default
#asahi · 2021-05-24 17:11
svendwc3 also works in bypass mode fwiw
#asahi · 2021-05-24 17:07
svenmarcan did some experiments where he just disconnected the D+/D- lines and that worked just fine as well
#asahi · 2021-05-24 17:07
svenyeah, that sounds just like what i saw
#asahi · 2021-05-24 17:05
svennice. i assume bsd itself needs almost no changes if you just do the phy init in u-boot?
#asahi · 2021-05-24 14:02
sven(without manually resetting dwc3)
#asahi · 2021-05-24 14:02
svenjust wondering if reusing a port works because i couldn't get that to work in either device or host mode
#asahi · 2021-05-24 14:01
svenhttps://github.com/AsahiLinux/m1n1/blob/main/proxyclient/i2c.py#L112
#asahi · 2021-05-24 14:01
svenyeah, i know :-)
#asahi · 2021-05-24 13:57
sven(i.e. disconnect a device and connect it again)
#asahi · 2021-05-24 13:57
svencan you use a port twice there?
#asahi · 2021-05-23 11:56
sveni can show the draft here before i submit it :)
#asahi · 2021-05-23 11:56
svenyes
#asahi · 2021-05-23 11:55
svenalright, let me rebase that and submit it then
#asahi · 2021-05-23 11:55
sveni can mention it in the first mail as an alternative though
#asahi · 2021-05-23 11:54
sveni had a prototype but uhh... no one else really does it and it doesn't look like it's obviously-correct
#asahi · 2021-05-23 11:54
svenmarcan: well, i could just submit the thing that maps a page for each clock to start a discussion
#asahi · 2021-05-16 13:23
svenjannau: https://f.svpe.de/60643d23cac86f10491e5fcc4c8c1291ff82b54c760bfee731a7c04651ed1608_np_chksum2.py that one should be ~10x faster and maybe still produce the same checksums
#asahi · 2021-05-16 13:03
svenjannau: https://f.svpe.de/3a908a03b5060fd7911166967486d80e895b4384cadb298ce80739091829e03e_np_chksum.py that might work. not much faster though i think
#asahi · 2021-05-16 12:56
sventurns out that my stupid numpy code is also slower than the pure python code :D
#asahi · 2021-05-16 12:56
svenmeh.
#asahi · 2021-05-16 12:56
svenoh, it doesn't?
#asahi · 2021-05-16 12:55
svenanyone with access to m1n1 and a mac around? i think i can reduce the checksum to a few numpy ops but i'm on a train and can't test anything right now
#asahi · 2021-05-15 18:19
svenit's still weird that some parts of the dwc controller have to use the first dart and others the second dart though
#asahi · 2021-05-15 18:18
svenbut the maintainers seemed to prefer the separate node thing *shrug*
#asahi · 2021-05-15 18:18
sveni still don't have a strong opinion on that
#asahi · 2021-05-15 18:15
svennow i just need to fix that 16k/4k pages issue and then i can submit v3
#asahi · 2021-05-15 18:15
svenregarding the DART stuff I wanted to make sure this DAPF or the SMMU instances don't need special treatment first. i think they don't.
#asahi · 2021-05-15 18:14
sveni could just submit the clock stuff which maps a single page for each clock
#asahi · 2021-05-15 18:13
svenhides
#asahi · 2021-05-14 12:55
svenyeah, that sounds the least confusing
#asahi · 2021-05-14 12:54
sven(erm. no. i'm just confused)
#asahi · 2021-05-14 12:54
sven(which is useless because VHE, but still)
#asahi · 2021-05-14 12:53
svenoh dear, there's also a mask for el2 i think
#asahi · 2021-05-14 12:51
svenugh.. this is going to be seriously confusing
#asahi · 2021-05-14 12:50
svenyes :/
#asahi · 2021-05-14 12:49
svenso... SPRR_EL0_PERM_MASK_EL1 and SPRR_EL1_PERM_MASK_EL1? :D

plus ›