IRC channel logs


back to list of logs

<damo22>Sending IPIs to APIC ID 1... <crash>
<damo22> * On x86 it's required to boot all logical CPUs at least once so
<damo22> * that the init code can get a chance to set CR4.MCE on each
<damo22> * CPU. Otherwise, a broadcasted MCE observing CR4.MCE=0b on any
<damo22> * core will shutdown the machine.
<damo22>thats all i could find
<damo22>Started cpu 1 (lapic id 0001)
<damo22>coreboot/src$ vim cpu/x86/lapic/lapic_cpu_init.c
<damo22>it was the IPI sequence
<damo22>Sending IPIs to APIC ID 1...
<damo22>Waiting for IPI...+done
<damo22>Waiting for IPI...+done
<damo22>After apic write
<damo22>Waiting for IPI...+done
<damo22>APIC delivery error 0x4 <----
<damo22>Started cpu 1 (lapic id 0001) HANG
<damo22>hmm i didnt fix it after all but i got closer
<zamfofex>Yes, I feel like Serenity OS has a lot of visual charm. Whereas the Hurd’s charm is mostly technical.
<damo22>bit 2 = SendAcceptError: a message sent by this apic was not accepted by any other apic
<damo22>maybe it only needs one round of SIPI
<Gooberpatrol66>technically charming, the best kind of charming
<damo22>according to the BKDG for my cpu, INIT IPI is not valid for deassert on a level triggered mode
<damo22>argh, it still fails to accept SIPI
<damo22>youpi: i think it accepted the SIPI this time, but ran the reset vector bootblock and tried to reboot coreboot on the AP
<damo22>or maybe it just crashed
<Pellescours>damo22: I tried your branch and each time, the assertion failed. The memory for AP is greater than 1MB, I tried to boot multiple times and each time I get the assertion
<youpi>talking about glibc 2.39, it's already in the git repo of debian
<youpi>but we are still waiting for packages to fix their bugs against 2.38
<flavioc>youpi: is anybody working on the GDB x86_64 port? I have a mostly working patch, just needs some polishing before sending for review
<youpi>iirc somebody was, but nothing was published
<flavioc>ok, I will try to publish it later today
<solid_black>hi luckyluke!