You could always ask whomever sent you that file to create another one without using Mac's proprietary BS.

Yeah. . . I think it's an anomally from a 7zip file created on a Mac.

IT's part of a "Capture the Flag" exercise, but I think it's a red herring.

But that leaves me stuck with: termbin.com/hk8j

CrtxReavr: Appears to be gibberish. But, I could be wrong!

so I attempted to try out iwlwifi instead of iwm, without success, the driver loads, wlan0 attaches to it, but it fails to actually connect

I see these messages that seem relevant: termbin.com/wmkj

I have issues with the emmc of my new board. "mmcsd0: Error indicated: 1 Timeout", followed by an sdhci_pci1_slot0 Controller timeout and register dump. Last message just after the dump is "Tuning failed, using fixed sampling clock".

Apparently, people used "hw.sdhci.enable_msi=0" to 'fix' this, but since the posts mentioning this are about a decade old, I am not sure if that is still helpful.

disabling MSI/MSI-X can always be useful if the hardware is broken or not wired up correctly, which is not that uncommon

Okay, so I will give that one a try. Is there anything else I can try? I mean, apart from getting rid of the board, lol

ivy: and for MSI-X is there an additional tunable, or is that implied/included with enable_msi=0?

since there's no enable_msix for this driver, MSI-X probably isn't supported at all

Remilia: Yup

Ok I have a major issue which I keep ignoring but I keep getting hard poweroffs because its 30C+ in London and cooling capacity is limited in this weather... there is no way of controlling the fan speed right now, and libreboot has a bug where 70C+ the system powers off. I have tried to manually downclock the cpu before but it seems that it is ignored, which I assume might be due to powerd?

Anyways, is there any tips on how I can prevent the cpu rising about 70C OS side?

I dont think it would be hard to write a basic daemon which checks the coretemp every second to see if its risen above say 60C and drop the clock speed

but I assume that would be reinventing the wheel, so any suggestions?

polarian: have you tried to tweak EDC/TDC/PPT values?

just undervolting/underclocking won't always do it if the mobo simply pushes too much power

70C is too low of a temp to be shutting down, libreboot should fix that tbh

dstolfa: yeah it needs reverse engineering

and the dev in charge of the device is busy...

also no I havent touched the EDC/TDC/PPT values, nor would I know how to tweak them

well its an old laptop, and the dual core cpu eats 35w under max load

the usual issue is firefox, rending a bloated website with too much js, cpu hits 100%, 2-3 mins later coretemp exceeds 70C, poweroff

or if I want to compile something, I point a desk fan at the bottom of the laptop, and leave it upside down for better airflow, keeps the cpu nice and cool and allows it to compile at full clock speed :P

sounds like your best bet is to revert to regular firmware until libreboot fixes the issue or just suffer the bug

wait a second, I just had a stupid idea... hypothetically if I cut the pwm wire, the fan should then spin full speed 24/7...

(its quiet enough anyways, just no silent)

or... the fan might not spin at all with no voltage to the pwm wire

but a quick google search indicates its the former, not the latter...

dstolfa: nahhhhhhhhh... no fun!

I didnt mention that the acpi code is not done, so I cant see battery charge, or suspend when lid shuts, or wake from keyboard interaction

wanna go to sleep? doas acpiconf -s3

want to wake up, power button only

lets be real if I wanted a solution which worked out of the box I wouldn't be running freebsd on a laptop

wait so PWM fans spin depending on the percentage of time the PWM pin is high, so if it is high 24/7 say I spliced the cable and connected it to 5V, in theory... it should always spin at 100%

you could but if it's a high powered fan, you might kill it quite quickly

ughhh

I have a Rpi pico... uses like 0.2w :)

could shove one in the case somewhere, have it bring the pwm pin high and low xD

hmm in fact, pico can take data via usb port along with power iirc... sacrifice a usb port, write a basic program to feed the pico the fan speed desired, have the program hook into coretemp and then boom, EC bypassed!

curious anyone use zfs from ports? specifically zfs 2.3.3 for raidz expansion?

polarian: Maybe consider getting new fans, keeping your room cool, and remove any dust accumulated inside your computers and servers.

Especially servers, since those are powered on 24/7, they tend to collect dust really quick.

remiliascarlet: the servers dont have that issue :)

although I havent cleaned mine for 1.5 years

it must be caked in dust

I havevent looked at the hardware in that long too :P

my laptop has been cleaned out, there is nothing more I can do :/

its a libreboot issue

70C+

sure? repaste the cooling system

I connected a 20TB usb disk to my 11.3 system (it's FreeNAS), created a zpool on it, then decided to do things differently and removed the zpool. I tried to "gpart destroy da0" to start from scratch, but gpart complains "device busy"

how can find out why it is busy?

sphex_, rwp: Do you folks have bugs open about this? Seems like that'd be a critically useful step to getting resolution.

treefrob: was it a raw zpool? because if it was there shouldn't be anything to destroy

I think it will give a device busy error if there are partitions present but do not have a handy empty drive to test

Remilia, you're right. It refuses if partitions exist, unless "-F" is given. Did that, and it destroyed the partition table

I've created a new GPT table with a single partition and used mkfs.ext4 to create a filesystem on da0p1, but I can't mount it with "mount [-t ext2fs] /dev/da0p1 /mnt/backup". The complaint is "Invalid argument" :(

treefrob:

you tried: mount -t ext4 /dev/da0p1 /mnt/backup

need to specify the actual filesystem

sig`: ext4 is ext2 in a trenchcoat and FreeBSD's mount option for ext4 is ext2fs

see `man ext2fs`

first tried: mkdir -p /mnt/backup ???

powerd never seems to downclock or do fuck all ngl

I have played with it and ran it on powerd_flags="-a hiadaptive -b minimum -i 40 -r 75"

but it always ends up clothing at max 24/7

in fact I manually dropped the clock speeds down to the minimum level and the performance is identical, and the heat issue has gone away... I know if I was compiling or something it would be a massive difference (its half the max clock), but I dont see why powerd isn't dropping the clock speed down to min when the cpu is mostly idling

mason, I keep hearing everyone else talk about the problem as if it is already well known. But I don't have a bug number for it. I might presume that it is not a problem in 15. Eventually I will get to 15 and then see if it is still a problem or not.

rwp: Probably worth opening a bug anyway, just in case.