RFC: Make testbsp useful on x86_64

[reet-]

Hi,

I'm currently trying to hack the testbsp example to work on the x86_64 target. The goal is to have a good coverage of the expected BSP functionality for an upcoming x86_64 variant (Muen, see [1]).

If I run the binary via QEMU/KVM, I get a #GP exception, see below:

$ /opt/gnatpro-23.1-x86_64-elf-linux64-bin/bin/../libexec/gnatemulator/x86_64-elf/23.1/qemu-system-x86_64 -d int -gdb tcp::1234 -serial mon:stdio -nographic -no-reboot -L /opt/gnatpro-23.1-x86_64-elf-linux64-bin/bin/../share/gnatemulator/x86_64-elf/pc-bios/ -cpu SandyBridge,-check -bios - -device loader,file=bb-runtimes/examples/test-bsp/testbsp
Start TestBSP
  * check check against unexpected restart, OK
  * check initialized board, OK
  * check initialized interrupts, OK
  * check data initialization, OK
=== Clock tick duration: 1 ns
  * check clock is running, OK
  * check clock counted up at least 1 msec, OK
  * check clock counted up less than 2 msec, OK
  * check clock reading takes less than 20 usec, OK
  * check clock reading takes at least 5 ns, OK
=== Clock_Read duration: 67 ns
  * check all timer tests, OK
  . Initializing context
  * check Context_Switch to self, OK
  . Testthread (123)
  * check Context_Switch to test thread increments shared var, OK
  . Testthread got control back  after context switch
  * check repeated context switch, OK
=== Context_Switch duration: 351 ns
  * check context switches properly commit shared variables, OK
  * check all context switch tests, OK
  . Alarm_Interrupt_ID = 255
  . Priority_Of_Interrupt (Alarm_Interrupt_ID) = 255
  . Installing interrupt handler
check_exception old: 0xffffffff new 0xd
     0: v=0d e=0000 i=0 cpl=0 IP=0008:0000000000107e40 pc=0000000000107e40 SP=0010:00000000003ce898 env->regs[R_EAX]=7463656a626f2064
RAX=7463656a626f2064 RBX=0000000000123358 RCX=0000000000123698 RDX=0000000000114300
RSI=00000000000000ff RDI=00000000001af300 RBP=00000000003ce8d0 RSP=00000000003ce898
R8 =0000000000123698 R9 =0000000000000000 R10=00000000001b1c82 R11=00000000003cecc0
R12=0000000000102b37 R13=00000000003cec48 R14=0000000000123108 R15=00000000003cec48
RIP=0000000000107e40 RFL=00000002 [-------] CPL=0 II=0 A20=1 SMM=0 HLT=0
ES =0010 0000000000000000 ffffffff 00cf9300 DPL=0 DS   [-WA]
CS =0008 0000000000000000 ffffffff 00af9a00 DPL=0 CS64 [-R-]
SS =0010 0000000000000000 ffffffff 00cf9300 DPL=0 DS   [-WA]
DS =0010 0000000000000000 ffffffff 00cf9300 DPL=0 DS   [-WA]
FS =0010 0000000000000000 ffffffff 00cf9300 DPL=0 DS   [-WA]
GS =0000 00000000001b03a0 0000ffff 00009300 DPL=0 DS   [-WA]
LDT=0000 0000000000000000 0000ffff 00008200 DPL=0 LDT
TR =0018 00000000001b03a0 0000007b 00008900 DPL=0 TSS64-avl
GDT=     00000000001260e0 00000027
IDT=     00000000001af3a0 00000fff
CR0=80010011 CR2=0000000000000000 CR3=0000000000128000 CR4=00040220
DR0=0000000000000000 DR1=0000000000000000 DR2=0000000000000000 DR3=0000000000000000 
DR6=00000000ffff0ff0 DR7=0000000000000400
CCS=00000000000000ff CCD=7463656a626f2064 CCO=LOGICQ
EFER=0000000000000500
FATAL PROCESSOR EXCEPTION RAISED: General Protection ExceptionError Code:  0
Rebooting...

Before I invest more time, is this supposed to work, or is my testbsp port to x86 missing something?

0x107e40 is inside the system__bb__threads__queues__change_priority function, I guess the #GP is a result of the borked value in RAX, used as pointer.

Thanks
reto

[1] - https://github.com/codelabs-ch/bb-runtimes

[CLAassistant]

All committers have signed the CLA.