request ms9132: quite successful already with just adding missing MODULE_ALIAS and USB_DEVICE_AND_INTERFACE_INFO

[jakobi]

Request to add ms9132.

Both request and a heads up for a simple USB3-A to HDMI dongle (based
on Macrosilicon ms9132, usb 345f:9132) showing only some rectangular artefacts (speed permitting; possibly the artifacts are kind of compression restarts?).

Procedure

1. make it compile for the kernel in use: For 6.1 e.g. select a late 2022 commit, suitably switch between dma_buf_map and iosys_map, add missing include drm/drmedid.h) (this is for a Kali 2023.1 stick, stuck at kernel 6.1, as I cannot update the kernel),_
2. modprobe drm_shmem_helper,
3. add v345Fp9132 in MODULE_ALIAS and id_table
4. run insmod.sh to recompile and test in X11

Relevant changes

diff 'dd--color=auto' -u -r ms912x/ms912x.mod.c ms912xMOD/ms912x.mod.c
--- ms912x/ms912x.mod.c 2023-10-18 20:21:16.700507428 +0000
+++ ms912xMOD/ms912x.mod.c 2023-10-19 01:19:16.422745923 +0000
@@ -116,3 +116,4 @@

 MODULE_ALIAS("usb:v534Dp6021d*dc*dsc*dp*icFFisc00ip00in*");
 MODULE_ALIAS("usb:v534Dp0821d*dc*dsc*dp*icFFisc00ip00in*");
+MODULE_ALIAS("usb:v345Fp9132d*dc*dsc*dp*icFFisc00ip00in*");
diff '--color=auto' -u -r ms912x/ms912x_drv.c ms912xMOD/ms912x_drv.c
--- ms912x/ms912x_drv.c 2023-10-18 20:04:13.723435640 +0000
+++ ms912xMOD/ms912x_drv.c 2023-10-19 00:29:41.244899105 +0000
@@ -273,6 +273,7 @@
 static const struct usb_device_id id_table[] = {
   { USB_DEVICE_AND_INTERFACE_INFO(0x534d, 0x6021, 0xff, 0x00, 0x00) },
   { USB_DEVICE_AND_INTERFACE_INFO(0x534d, 0x0821, 0xff, 0x00, 0x00) },
+  { USB_DEVICE_AND_INTERFACE_INFO(0x345f, 0x9132, 0xff, 0x00, 0x00) },
   {},
 };
 MODULE_DEVICE_TABLE(usb, id_table);

Tests in X11

Lenovo T61: Due to usb2 speed, on T61 I see a ton of rectangular artefacts, and a stable 1920x1080 image is only shown when idle. Usable for manpages and vi, not for browsing less than fully static pages. Youtube is good for jokes only...

Yoga 720 usb3: I've only rarely artefacts. I seem to be able to trigger them somewhat when moving windows, terminal use is near perfect, and even youtube in the browser is sluggisch, delayed(1+s), but with near always sane display. Audio-over-hdmi seems to be missing.

[rhgndf]

Kernel 6.1

Can you try the branch kernel-6.1?

USB 2

I have no idea what's causing it. In my tests, I figure out sometimes minor artefacts are due to the desktop environment, but no major artefacts are seen. Some people have reported more severe artefacts which could be similar to yours. More debugging is needed, so if you are able to help debug, that would be welcome.

USB 3

I don't have usb3 device to test, so thanks for testing!

Source modifications

Nice work on the usb3 side! For the sources files, only the modifications to ms912x_drv.c is needed. Could you do a PR and attach the lsusb -v output also?

[rhgndf]

It might be fixed with ac7e2a5d38c765de21e8df646c969a03c6175422.

[rhgndf]

Latency issue might be fixed with b763760e9c83bcc299ebd08972d79e7bebbc3c2d

Could you test the new commits to see if it has been fixed with USB3? Also, audio over HDMI works on my side, but I only have a USB2 dongle. Once its good I will add the VID PID in.

[buddyjojo]

I tested my ms9132 after adding the vid/pid, audio worked fine and didn't notice too much latency or graphical glitches. Though I did get some screen tearing but I did all my testing on Wayland so could be an issue with that. Need any more info/testing from this device?

[rhgndf]

Since I don't have a USB 3 device and it looks like this works as is without modification, the vid/pid is added in 738aef1b659927b21be62a71fbe6f296bea86ae0.