Re: apple m1 status update (20230610)

[sc dying <sc.dying%gmail.com@localhost>]

On Sat, May 10, 2025 at 2:03 AM Jun Ebihara <jun%soum.co.jp@localhost> wrote:
>
> many thanx!
>
> 2025年5月8日(木) 15:38 sc dying <sc.dying%gmail.com@localhost>:
> > Do you see any errors from uaudio on console?
>
> no errors.

Good news for me.
Did you see any errors before patch?

>
> > Your kernel works well on my rpi4 and nanopi neo3.
> > Can you try attached patch?
>
> I patched kernel put to
> https://ftp.netbsd.org/pub/NetBSD/misc/jun/aarch64/apple-m1/netbsd.2025-05-10
>
> > Please patch to original xhci.c, rebuild kernel with
> > XHCI_DEBUG, boot new kernel with your D10s, and play audio.
> > You can see hw.xhci.debug MIB value as 0, please let it 0.
> > `vmstat -u usbhist' shows descriptor values, please send me it.
>
> NetBSD m1.soum.co.jp 10.99.14 NetBSD 10.99.14
> (GENERIC64.202505100000UTC) #3: Sat May 10 10:39:24 JST 2025
> jun%amir63d.soum.co.jp@localhost:/usr/src/sys/arch/evbarm/compile/GENERIC64
> evbarm
> m1$ sysctl -a |grep hw.xh
> hw.xhci.debug = 0
> m1$ vmstat -u usbhist
> 50000 entries, next is 115
 (snip)

Than you.
I could not find any problem.

These two are isochronous pipes.

> 1746841934.687156 xhci_open#24@5: called: addr 1 depth 1 port 1 speed 3
> 1746841934.687156 xhci_open#24@5:  slot 1 dci 2 epaddr 0x01 attr 0x05
> 1746841934.687156 xhci_open#24@5:  mps 0x188 maxb 0 bival 1 pipe
> 0xffff00097fd0de00
> 1746841934.687215 xhci_setup_ctx#0@5: slot 1 dci 2: ival 0 mps 0x188
> 1746841934.687215 xhci_setup_ctx#0@5:  mult 0 maxburst 0 mep 0x188 atl 0x188

0x188 (bytes) / 2^(1 - 1) (125us) = 3136000 (bytes/s)
Looks 392khz/32bit, or so?

> 1746841934.687383 xhci_open#25@5: called: addr 1 depth 1 port 1 speed 3
> 1746841934.687383 xhci_open#25@5:  slot 1 dci 3 epaddr 0x81 attr 0x11
> 1746841934.687383 xhci_open#25@5:  mps 0x4 maxb 0 bival 4 pipe
> 0xffff0009a698e500
> 1746841934.687415 xhci_setup_ctx#0@5: slot 1 dci 3: ival 3 mps 0x4
> 1746841934.687415 xhci_setup_ctx#0@5:  mult 0 maxburst 0 mep 0x4 atl 0x4

Do you know what this isoch-IN endpoint is?

> 1746841934.687725 xhci_device_isoc_enter#0@5: bival 0 ival 1 isoc_next 0x178
> 1746841934.687725 xhci_device_isoc_enter#0@5: mfindex 0x13f isoc_next 0x178

> 1746841998.153689 xhci_open#26@6: called: addr 1 depth 1 port 1 speed 3
> 1746841998.153690 xhci_open#26@6:  slot 1 dci 2 epaddr 0x01 attr 0x05
> 1746841998.153690 xhci_open#26@6:  mps 0x188 maxb 0 bival 1 pipe
> 0xffff0009a5f9fb00
> 1746841998.153692 xhci_setup_ctx#0@6: slot 1 dci 2: ival 0 mps 0x188
> 1746841998.153693 xhci_setup_ctx#0@6:  mult 0 maxburst 0 mep 0x188 atl 0x188
> 1746841998.153875 xhci_open#27@6: called: addr 1 depth 1 port 1 speed 3
> 1746841998.153875 xhci_open#27@6:  slot 1 dci 3 epaddr 0x81 attr 0x11
> 1746841998.153875 xhci_open#27@6:  mps 0x4 maxb 0 bival 4 pipe
> 0xffff000999e91280
> 1746841998.153878 xhci_setup_ctx#0@6: slot 1 dci 3: ival 3 mps 0x4
> 1746841998.153878 xhci_setup_ctx#0@6:  mult 0 maxburst 0 mep 0x4 atl 0x4
> 1746841998.154038 xhci_device_isoc_enter#0@6: bival 0 ival 1 isoc_next 0xc5
> 1746841998.154039 xhci_device_isoc_enter#0@6: mfindex 0x8c isoc_next 0xc5
>
>
> > The patch differs from previosu
> > - replace aprint_{debug,verbose}_dev with aprint_normal_dev
> >   to show if HCC supports CFC.
> > - add debug logs to know descriptor values to be configfured.
> >
> > Cheers,
> > --
> >
> > On Wed, May 7, 2025 at 9:27 PM Jun Ebihara <jun%soum.co.jp@localhost> wrote:
> > >
> > > 2025年5月2日(金) 17:55 <sc.dying%gmail.com@localhost>:
> > > > I'm revisiting this problem.
> > > > https://mail-index.netbsd.org/port-arm/2023/06/18/msg008275.html
> > > > If interested, could you please test new patch?
> > >
> > > Thanx! I apply your patch and make /netbsd.But output sound is still wicked with
> > > uaudio0: Topping (0x152a) D10s (0x8750), rev 2.00/1.04, addr 4
> > >
> > > the kernel put it on
> > > https://cdn.netbsd.org/pub/NetBSD/misc/jun/aarch64/apple-m1/netbsd.2025-05-02
> > >
> > > dmesg:
> > > https://github.com/ebijun/NetBSD/blob/2dfb47313e3db1534a8b07f06f28841ddf514fc4/dmesg/aarch64/Apple_Mac_mini_M1_2020
> > >
> > > --
> > > Jun Ebihara
>
>
>
> --
> Jun Ebihara