Bug 276523 - serial console not functional orange pi 5 plus
Summary: serial console not functional orange pi 5 plus
Status: In Progress
Alias: None
Product: Base System
Classification: Unclassified
Component: arm (show other bugs)
Version: 14.0-STABLE
Hardware: arm64 Any
: --- Affects Only Me
Assignee: freebsd-arm (Nobody)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2024-01-22 11:13 UTC by rdunkle
Modified: 2024-01-24 08:28 UTC (History)
1 user (show)

See Also:

Attachments
var/log/messages from boot -v (34.56 KB, text/plain)
2024-01-22 11:13 UTC, rdunkle 		no flags Details
output --> sysctl -ad (466.95 KB, text/plain)
2024-01-22 16:05 UTC, rdunkle 		no flags Details
acpidump -b (5.53 KB, application/x-zip-compressed)
2024-01-22 18:06 UTC, rdunkle 		no flags Details
ns8250: Try to get the clock-frequency property (1.38 KB, patch)
2024-01-23 10:08 UTC, Emmanuel Vadot 		no flags Details | Diff
acpidump after patch (5.53 KB, application/x-zip-compressed)
2024-01-23 14:16 UTC, rdunkle 		no flags Details
boot -v terminal output and var-log-messages (11.24 KB, application/x-zip-compressed)
2024-01-24 08:28 UTC, rdunkle 		no flags Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description rdunkle 2024-01-22 11:13:34 UTC 
Created attachment 247841 [details]
var/log/messages from boot -v

FreeBSD14-stable  orange pi 5 plus
I see the same problem in FreeBSD 13.2
serial console works during the efi boot steps.
The console looks like this:
Setting currdev to disk0p2:
Loading /boot/defaults/loader.conf
Loading /boot/defaults/loader.conf
Loading /boot/device.hints
Loading /boot/loader.conf
The processors are identified and then serial output stops.
Final output looks like this:
output to serial always stops here--> I/TC: Secondary CPU 7 switching to normal world boot.
---
# dmesg | grep uart
uart0: <16750 or compatible> iomem 0xfeb50000-0xfeb50fff irq 21 on acpi0
uart0: fast interrupt
uart0: PPS capture mode: DCD
-- never see that typical line-->  uart0: console (115200,n,8,1)

I filed a bug report with the github efi project.
https://github.com/edk2-porting/edk2-rk3588/issues/118
---
I made the change as suggested to: 
sys/dev/uart/uart_dev_ns8250.c
line 452
{"HISI0031", &uart_ns8250_class, 0, 2, 0, 24000000, UART_F_BUSY_DETECT, "HiSilicon / Synopsys Designware UART"},
-----
I see the same problem with serial port.
Need guidance on how to fix this problem
Comment 1 Emmanuel Vadot freebsd_committer freebsd_triage 2024-01-22 13:08:21 UTC 
Is there an ACPI method or property to get the base clock ?
Patching like you did will break what ever other hardware this compat was added for.
Comment 2 rdunkle 2024-01-22 16:05:42 UTC 
Created attachment 247847 [details]
output --> sysctl -ad

attached a sysctl list from a non-patched system
Comment 3 Emmanuel Vadot freebsd_committer freebsd_triage 2024-01-22 16:50:42 UTC 
sysctls won't help, do :
pkg install acpica-tools
/usr/local/bin/acpidump -b

and post the .dat file here.
Comment 4 rdunkle 2024-01-22 18:06:20 UTC 
Created attachment 247848 [details]
acpidump -b
Comment 5 Emmanuel Vadot freebsd_committer freebsd_triage 2024-01-23 10:08:20 UTC 
Created attachment 247879 [details]
ns8250: Try to get the clock-frequency property

Please try with this patch.
Comment 6 rdunkle 2024-01-23 14:16:53 UTC 
Created attachment 247889 [details]
acpidump after patch
Comment 7 rdunkle 2024-01-23 14:18:05 UTC 
the same results, no terminal.  attached new acpidump
Comment 8 Emmanuel Vadot freebsd_committer freebsd_triage 2024-01-23 18:34:09 UTC 
No need to upload acpidump each time, this won't modify the table.
Also it looks like that my patch is useless because we already try to get the clock-frequency property (https://cgit.freebsd.org/src/tree/sys/dev/uart/uart_bus_acpi.c#n99) so I don't really know why it doesn't work here.
Can you share a boot -v, it might contain more info
Comment 9 rdunkle 2024-01-24 08:28:59 UTC 
Created attachment 247917 [details]
boot -v terminal output and var-log-messages