OS hangs without going to KDL when performing large/long compiles

[electrodeyt]

A semi-reproducible way to cause this (for me) is to run haikuporter on gcc (I ran haikuporter -S -j8 --no-source-packages --get-dependencies gcc). KDL does not start and the serial debugger does not show anything either (although I have included the very short log of what happens before it hangs)

Tested on hrev58043 and hrev58067. Hardware used is a Fujitsu mini PC with a Intel i5 6500T; hang happens with and without SMP (although the non-SMP hang was not triggered when compiling gcc, given that it was running a multi-threaded python script before it hanged, I'm pretty sure its the same issue)

Serial log (both the same for both hrevs i tried):

CPU: no microcode provided
CPU 0: type 0 family 6 extended_family 0 model 14 extended_model 5 stepping 3, string 'GenuineIntel'
CPU 0: vendor 'Intel' model name 'Intel(R) Core(TM) i5-6500T CPU @ 2.50GHz'
CPU 0: apic id 0, package 0, core 0, smt 0
CPU 0: cache sharing: L1 id 0, L2 id 0, L3 id 0
CPU 0: features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clfsh ds acpi mmx fxsr sse sse2 ss htt tm pbe sse3 pclmulqdq dtes64 monitor dscpl vmx smx est tm2 ssse3 fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tscdeadline aes xsave avx f16c rdrnd syscall nx pdpe1gb long dts itb arat pln ecmd ptm hwp hwp_notify hwp_actwin hwp_epp hdc aperfmperf epb tsc_adjust sgx bmi1 hle avx2 smep bmi2 erms invpcid rtm mpx rdseed adx smap cflushopt intel_pt xsaveopt xsavec xgetbv1 xsaves
CPU 0: patch_level 0xba
CPU: using TSC frequency from CPUID
reserve_io_interrupt_vectors: reserved 16 vectors starting from 0
mark_page_range_in_use(0x0, 0x100): start page is before free list
mark_page_range_in_use(0x0, 0xa0): start page is before free list
... started gcc compile and a while later (time seems to be random, have had it happen both very quickly and after quite a bit of time) it hangs here ...

[waddlesplash]

That serial log looks very incomplete; it's missing most of the boot and then run?

[electrodeyt]

Serial log

[electrodeyt]

Woops, forgot to actually enable serial debug in the bootloader (just saw output on the serial port and thought it was working), enabled it and attached it. It's mostly complete, but does miss some of the early boot because apparently screen didn't log it all; sorry about all that!

[waddlesplash]

This is still missing a lot, but at least it gets to where I presume the hang actually happens. Not much of interest in it, though, but that's not surprising.

I guess you can't enter KDL via the keyboard shortcut (Alt+SysRq+D?) If not, then this may be another candidate for a change I have in mind to the ICI system to detect other cores hanging.

[electrodeyt]

Haven't tried that yet, will try in a bit but I have to actually find another keyboard first because it will not detect my USB keyboard after it hangs (At the moment I am switching my keyboard between the Haiku box and my main PC; which is where the USB disconnect logs come from)