---------- Forwarded message --------- From: Hank Barta <hba...@gmail.com> Date: Tue, Sep 13, 2022 at 12:54 PM Subject: Re: Bug#1019700: mmc0: Timeout waiting for hardware cmd interrupt. To: Bjørn Mork <bj...@mork.no>
Hi Bjørn, Many thanks for the prompt reply. In the mean time I have done the following: * Reimaged my SD card with `20220808_raspi_4_bookworm.img.xz` from Debian Tested images. (5.18.14-1 kernel) * Booted and noted *no* SD card timeouts. Rebooted and power cycled 3 times each with the same result. * Performed `apt update && apt upgrade -y` and rebooted. (5.19.6-1 kernel) * First boot - repeated SD timeouts and unable to log in. Power cycled to force reboot * Second reboot - no SD card timeouts. Added `dtparam=sd_poll_once=on` to `/boot/firmware/config.txt` * Third boot - repeated SD card timeouts. Evetually I was able to log in to the console. Network is not fully up. The repeated SD timeouts seem to be slowing normal boot. Actually I may not have been logged in but in the console that presents when there is a problem booting. I exited and now I see a login prompt. And Ethernet finally came up. 737 seconds post boot according to console messages. (It was some time later before I could ssh in.) The SD timeout messages stopped. I have a login prompt at the console but it takes about 30s to login. The system is now responsive, but WiFi modules did not load. I count 52 timeout messages in dmesg output. There is no response to <ctrl><alt><del> at the console. Tried to shutdown using `shutdown -r now` and the system hangs. The system is most certainly not operating normally. Does Debian use the device tree? This is a Debian system, not R-Pi OS. If I reboot enough times I will get a clean boot followed by normal operation. I have tried different SD cards, USB SSDs and Pi 4Bs all with the same result so I do not believe this is a H/W problem. I do recall the previous SD timeout issue and I worked around that by inserting an SD card post boot but that no longer works. This seems to be a new problem. best, hank On Tue, Sep 13, 2022 at 11:32 AM Bjørn Mork <bj...@mork.no> wrote: > Hank Barta <hba...@gmail.com> writes: > > > ** Kernel log: > > [ 723.735217] mmc0: sdhci: Timeout: 0x00000000 | Int stat: 0x00018000 > > [ 723.741743] mmc0: sdhci: Int enab: 0x00ff1003 | Sig enab: 0x00ff1003 > > [ 723.748270] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000001 > > [ 723.754797] mmc0: sdhci: Caps: 0x45ee6432 | Caps_1: 0x0000a525 > > [ 723.761324] mmc0: sdhci: Cmd: 0x00000502 | Max curr: 0x00080008 > > [ 723.767851] mmc0: sdhci: Resp[0]: 0x000001aa | Resp[1]: 0x00000000 > > [ 723.774379] mmc0: sdhci: Resp[2]: 0x00000000 | Resp[3]: 0x00000000 > > [ 723.780905] mmc0: sdhci: Host ctl2: 0x00000000 > > [ 723.785404] mmc0: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0x00000000 > > [ 723.791930] mmc0: sdhci: ============================================ > > [ 733.923993] mmc0: Timeout waiting for hardware cmd interrupt. > > These repeated messages are normal on the RPi4 if you boot it without an > SD card. E.g. from USB or network. If that's what you intend to do, > then you can avoid the repeated messages by adding > > dtparam=sd_poll_once=on > > to the config.txt file in your firmware partition. Often mounted as > /boot/firmware/. > > The effect depends on which device-tree you are using. I believe it > will only work with the ones coming with the Raspberry Pi firmware. See > > https://github.com/raspberrypi/firmware/blob/master/boot/overlays/README > > for docs. > > > Bjørn > -- Beautiful Sunny Winfield -- Beautiful Sunny Winfield
