Howdy I am chasing a bug with FreeBSD on RELEASE >13.0
When booting off the removable eMMC module it now fails at bootup. On 13.0 It was fine.
I have a little clue with kernel backtrace error now.
/usr/src/sys/dev/mmc/host/dwmmc.c:368
So hunting that down I don't see any error. Do you?
The offending line is this:
if ((cmd->opcode == MMC_WRITE_MULTIPLE_BLOCK ||
I have the nagging suspicion that something is turning off the emmc power on bootup.
mmc0 EXT_CLD not found
Is the last thing on screen before crash. I really believe it is a dtb problem.
Does anybody have any knowledge of what emmc.dtbo does?
I have seen this overlay on some systems but not this one.
It is not a gnu one but FreeBSD overlay I believe like the rpi-spigen.dtbo
extcld seems to have something to do with the emmc5.0 spec.
I am out of my league trying to scratch two sticks together.
What I do know on Hummingboard eMMC need some loader quirks at one point in time.
When booting off the removable eMMC module it now fails at bootup. On 13.0 It was fine.
I have a little clue with kernel backtrace error now.
/usr/src/sys/dev/mmc/host/dwmmc.c:368
So hunting that down I don't see any error. Do you?
Code:
dwmmc_tasklet(struct dwmmc_softc *sc)
{
struct mmc_command *cmd;
cmd = sc->curcmd;
if (cmd == NULL)
return;
if (!sc->cmd_done)
return;
if (cmd->error != MMC_ERR_NONE || !cmd->data) {
dwmmc_next_operation(sc);
} else if (cmd->data && sc->dto_rcvd) {
if ((cmd->opcode == MMC_WRITE_MULTIPLE_BLOCK ||
cmd->opcode == MMC_READ_MULTIPLE_BLOCK) &&
sc->use_auto_stop) {
if (sc->acd_rcvd)
dwmmc_next_operation(sc);
} else {
dwmmc_next_operation(sc);
}
}
}
The offending line is this:
if ((cmd->opcode == MMC_WRITE_MULTIPLE_BLOCK ||
I have the nagging suspicion that something is turning off the emmc power on bootup.
mmc0 EXT_CLD not found
Is the last thing on screen before crash. I really believe it is a dtb problem.
Does anybody have any knowledge of what emmc.dtbo does?
I have seen this overlay on some systems but not this one.
It is not a gnu one but FreeBSD overlay I believe like the rpi-spigen.dtbo
extcld seems to have something to do with the emmc5.0 spec.
I am out of my league trying to scratch two sticks together.
What I do know on Hummingboard eMMC need some loader quirks at one point in time.