#
…

2023-01-03

00:25jnn has joined #asahi-dev
00:25jn has quit [Read error: Connection reset by peer]
00:32jnn has quit [Remote host closed the connection]
00:34jn has joined #asahi-dev
01:04hertz_ has joined #asahi-dev
01:09hertz has quit [Ping timeout: 480 seconds]
04:40eiln has joined #asahi-dev
05:00c10l has quit [Quit: Bye o/]
05:00c10l has joined #asahi-dev
05:18jluthra has quit [Remote host closed the connection]
05:18jluthra has joined #asahi-dev
05:30whomst has joined #asahi-dev
05:36eiln has quit [Remote host closed the connection]
05:46eiln has joined #asahi-dev
06:33chadmed_ has joined #asahi-dev
06:35zrides has joined #asahi-dev
06:36eilnwhat a self-own, there's a semi-enforced iova upper limit .. "vm-size"
06:37eilnconsidering only dma to L2 fails when in the upper range, i think it's a distance issue ("hint" seems right)
06:37eilnoke back to work :)
06:37zrides has quit [Remote host closed the connection]
06:52SSJ_GZ has joined #asahi-dev
07:06pthariensflame has joined #asahi-dev
07:10pthariensflame has quit []
07:15eiln has quit [Remote host closed the connection]
07:21bps has joined #asahi-dev
07:38chadmed_ has quit [Ping timeout: 480 seconds]
07:38bps has quit [Ping timeout: 480 seconds]
07:53chadmed_ has joined #asahi-dev
08:20SSJ_GZ has quit [Remote host closed the connection]
08:29chadmed has quit [Ping timeout: 480 seconds]
08:33cylm_ has joined #asahi-dev
08:46amarioguy has quit [Remote host closed the connection]
08:46llrcombs has quit []
08:46amarioguy has joined #asahi-dev
08:47rcombs has joined #asahi-dev
08:59bps has joined #asahi-dev
09:08chadmed_ has quit [Ping timeout: 480 seconds]
09:27whomst has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
09:28maria has quit [Ping timeout: 480 seconds]
09:50bps has quit [Ping timeout: 480 seconds]
09:59chadmed has joined #asahi-dev
10:00maria has joined #asahi-dev
10:38kujeger has joined #asahi-dev
10:41jannauthe dart_hw_reset() in runtime resume seems break PCIe. no issues so far after replacing the reset with a restore, i.e. instead of resetting tcr and ttbr it restores the old values and calls invalidate_tlb() only after the values are restored
10:45jannaustill looks like the PCI's subsystem understanding of runtime pm is incompatible with powering the iommu down on suspend
11:03marcanthe PCIe subsystem doesn't have an understanding of runtime PM
11:03marcanit's not surprising that is broken
11:05MajorBiscuit has joined #asahi-dev
11:11jannauit looked like it uses generic runtime pm callbacks if pci drivers don't have their own callbacks. I stopped looking after that
11:25kettenisif DMA is quiesced before the iommu is powered down, it should work
11:27kettenisbut that does mean there is a dependency that needs to be enforced
11:27bps has joined #asahi-dev
11:33jannauhmm, I mistook the quirks for a generic suspend/resume handling so there is no runtime pm the PCIe drivers
11:34jannauI would have expected that the device link / runtime pm code keeps the supplier on in that case
11:36svenbah... and the acio driver is back to resetting the soc *sigh*
12:00chengsun_ has joined #asahi-dev
12:00chengsun has quit [Read error: Connection reset by peer]
12:09realwakka has joined #asahi-dev
12:17c10l has quit [Ping timeout: 480 seconds]
12:49jannauthe issue might be that pm_runtime_enable() returns true for the pcie device in dart_probe_device(). sysfs claims that runtime pm is not used for them, but dart_probe_device() is too early to check the device runtime pm status anyway
12:49jannauugly workaround for now: add DL_FLAG_RPM_ACTIVE if dev_is_pci(dev) is true
12:51jannauI wonder why this is not a problem on other systems. runtime pm in other iommu drivers looks close enough
13:01c10l has joined #asahi-dev
13:02chadmed has quit [Remote host closed the connection]
13:10kettenisjannau: which other iommu drivers?
13:11kettenisI had a quick look the other day, and the only iommu drivers that seem to do runtime pm are the SoC-specific ones
13:11roxfan2 is now known as roxfan
13:12kettenisthose are basically only used for the GPUs
13:13realwakka has quit [Remote host closed the connection]
13:13jannauyes, the soc specific ones and arm-smmu (not arm-smmu-v3)
13:20djorz has joined #asahi-dev
13:22marcanjannau: the problem is probably the same thing I ran into with DCP, that the genpd framework shuts down the PD as soon as *one* device attaches that claims to be RPM aware
13:22marcanwhich meant that the DART attaching to the DCP power domain before the FB or DCP itself was killing it
13:23marcanmy hack with DCP was to set the domain to always on and unset that in the DCP driver
13:23marcankilling the PCIe power domain breaks the init that m1n1 does
13:27jannaupcie works initially, ethernet seems to randomly break when trasmitting data. I think the issue is the reset in the dart resume function
13:28jannauI suspect it resets the dart while the device is in use
13:36systwi has quit [Ping timeout: 480 seconds]
13:36jannaups_apcie_gp_sys is always-on on t600x so it ought to be the dart reset
13:40systwi has joined #asahi-dev
13:40marcancould be, yeah
13:43jannauwell, the issue is that the dart is runtime suspended / resumed while the device is in use
13:43jannaubut responsible for the breakage is the reset as long as the power domain stays on
13:49hightower3 has joined #asahi-dev
13:50___nick___ has joined #asahi-dev
13:51___nick___ has quit []
13:53___nick___ has joined #asahi-dev
13:54Mudb0y has joined #asahi-dev
13:54Mudb0y has quit [Read error: Connection reset by peer]
13:54Mudb0y has joined #asahi-dev
13:56hightower4 has quit [Ping timeout: 480 seconds]
13:57Mudb0y has quit []
14:16Mudb0y has joined #asahi-dev
14:21Mudb0y has quit []
14:29bcrumb has joined #asahi-dev
14:31c10l has quit [Quit: Bye o/]
14:34bcrumb has quit []
14:35bcrumb has joined #asahi-dev
14:36c10l has joined #asahi-dev
14:39bcrumb has quit []
15:05c10l has quit [Quit: Bye o/]
15:15hightower3 has quit [Ping timeout: 480 seconds]
15:21MajorBiscuit has quit [Quit: WeeChat 3.6]
15:22MajorBiscuit has joined #asahi-dev
15:24c10l has joined #asahi-dev
15:24MajorBiscuit has quit []
15:34c10l has quit [Quit: Bye o/]
15:54c10l has joined #asahi-dev
15:55balrog has quit [Quit: Bye]
16:02bcrumb has joined #asahi-dev
16:09bcrumb has quit [Quit: WeeChat 3.7.1]
16:11balrog has joined #asahi-dev
16:25bcrumb has joined #asahi-dev
16:38hightower2 has joined #asahi-dev
16:50cylm_ has quit [Ping timeout: 480 seconds]
16:54bps has quit [Ping timeout: 480 seconds]
17:01nopeslide1 has joined #asahi-dev
17:07nopeslide has quit [Ping timeout: 480 seconds]
17:08bcrumb has quit [Quit: WeeChat 3.7.1]
17:25hightower2 has quit [Ping timeout: 480 seconds]
17:26SSJ_GZ has joined #asahi-dev
17:51hightower2 has joined #asahi-dev
18:01bps has joined #asahi-dev
18:01bcrumb has joined #asahi-dev
18:14bcrumb has quit [Quit: WeeChat 3.7.1]
18:25bcrumb has joined #asahi-dev
18:26bcrumb has quit []
18:28pthariensflame has joined #asahi-dev
18:29pthariensflame has quit []
18:41bcrumb has joined #asahi-dev
18:47bcrumb has quit [Quit: WeeChat 3.7.1]
18:57hertz_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
19:44whomst has joined #asahi-dev
20:01hertz has joined #asahi-dev
20:26lewurm has quit [Ping timeout: 480 seconds]
20:34c10l has quit [Quit: Bye o/]
20:35c10l has joined #asahi-dev
20:35lewurm has joined #asahi-dev
20:48lewurm has quit [Ping timeout: 480 seconds]
20:55rmkmarcan: good bye
20:55rmk has left #asahi-dev [#asahi-dev]
20:58psykose_ has quit [Remote host closed the connection]
20:58psykose has joined #asahi-dev
21:00lewurm has joined #asahi-dev
21:08___nick___ has quit [Ping timeout: 480 seconds]
21:15lewurm has quit [Ping timeout: 480 seconds]
21:25hertz has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
21:31blazra has joined #asahi-dev
21:51lewurm has joined #asahi-dev
22:14chadmed has joined #asahi-dev
22:26SSJ_GZ has quit [Ping timeout: 480 seconds]
23:32blazra has quit [Remote host closed the connection]
23:34bps has quit [Ping timeout: 480 seconds]
23:34blazra has joined #asahi-dev
…