rarias/nixos-riscv
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Add problems with the PLIC to the journal

Browse Source
This commit is contained in:
Rodrigo Arias 2024-09-04 12:43:46 +02:00
parent 272fd211b2
commit b075d6fa2a
1 changed files with 121 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

121
JOURNAL.md
View File

@ -3914,3 +3914,124 @@ frequency is:
Assuming that it actually works well. Let's try that and see if we can have a
more realistic reading for the CLINT interrupt frequency.
Something is going on with the plictool:
~ # plictool
plic=0x40800000 nsources=1024 ncontexts=15872
src=1 pend=1 prio=1 ctx=1 thre=0 <--- zero
src=2 pend=0 prio=1
src=3 pend=0 prio=1
src=4 pend=1 prio=1
src=33 pend=0 prio=0 ctx=1 thre=2
~ # plictool
plic=0x40800000 nsources=1024 ncontexts=15872
src=1 pend=1 prio=1 ctx=1 thre=2 <--- now changed to 2???
src=2 pend=0 prio=1
src=3 pend=0 prio=1
src=4 pend=1 prio=1
src=33 pend=0 prio=0 ctx=1 thre=2
Either the register is changing its value on its own, or the tool is causing it.
Let's revert the threshold to 0, but writing to the threshold address, should be
at 0x40a01000.
~ # devmem 0x40a01000
0x00000002
~ # devmem 0x40a01000
0x00000002
~ # devmem 0x40a01000
0x00000002
~ # devmem 0x40a01000 8 0
~ # devmem 0x40a01000
0x00000000
~ # devmem 0x40a01000
0x00000000
~ # devmem 0x40a01000
0x00000000
~ # plictool -c 2 -s 5 <-- limiting the sources and contexts
plic=0x40800000 nsources=5 ncontexts=2
src=1 pend=1 prio=1 ctx=1 thre=0 <-- good
src=2 pend=0 prio=1
src=3 pend=0 prio=1
src=4 pend=1 prio=1
~ # plictool -c 2 -s 5
plic=0x40800000 nsources=5 ncontexts=2
src=1 pend=1 prio=1 ctx=1 thre=0
src=2 pend=0 prio=1
src=3 pend=0 prio=1
src=4 pend=1 prio=1
~ # plictool -c 2 -s 5
plic=0x40800000 nsources=5 ncontexts=2
src=1 pend=1 prio=1 ctx=1 thre=0 <--- stable
src=2 pend=0 prio=1
src=3 pend=0 prio=1
src=4 pend=1 prio=1
~ # cat /proc/interrupts
CPU0
1: 1 SiFive PLIC 1 Edge ttyS1
5: 195413 RISC-V INTC 5 Edge riscv-timer
~ # cat /proc/int^Crupts
~ # echo a > /dev/ttyS1
~ # cat /proc/interrupts
CPU0
1: 1 SiFive PLIC 1 Edge ttyS1
5: 197102 RISC-V INTC 5 Edge riscv-timer
~ # plictool -c 2 -s 5
plic=0x40800000 nsources=5 ncontexts=2
src=1 pend=1 <--- wtf prio=1 ctx=1 thre=0
src=2 pend=0 prio=1
src=3 pend=0 prio=1
src=4 pend=1 prio=1
~ # devmem 0x40a01004
0x00000000 <--- the claim register continues to give 0 instead of 1
~ # devmem 0x40a01004
0x00000000
~ # devmem 0x40a01004
0x00000000
~ # devmem 0x40a01004
0x00000000
~ # devmem 0x40a01004
0x00000000
~ # devmem 0x40a01004 8 5 <-- try writing 5
~ # devmem 0x40a01004
0x00000000 <--- no change
~ # devmem 0x40a01004
0x00000000
~ # devmem 0x40a01004
0x00000000
~ # devmem 0x40a01004
0x00000000
~ # devmem 0x40a01004 8 1 <-- try with 1
~ # devmem 0x40a01004
0x00000000
~ # devmem 0x40a01004
0x00000000
~ # devmem 0x40a01004
0x00000000
~ # plictool -c 2 -s 5
plic=0x40800000 nsources=5 ncontexts=2
src=1 pend=0 <--- now it is gone prio=1 ctx=1 thre=0
src=2 pend=0 prio=1
src=3 pend=0 prio=1
src=4 pend=1 prio=1
~ # cat /proc/interrupts
CPU0
1: 3 <-- and we have >1 SiFive PLIC 1 Edge ttyS1
5: 213775 RISC-V INTC 5 Edge riscv-timer
~ # echo a > /dev/ttyS1
~ # cat /proc/interrupts
CPU0
1: 3 <--- stuck again SiFive PLIC 1 Edge ttyS1
5: 214551 RISC-V INTC 5 Edge riscv-timer
~ # plictool -c 2 -s 5
plic=0x40800000 nsources=5 ncontexts=2
src=1 pend=1 prio=1 ctx=1 thre=1 <--- again set to 1
src=2 pend=0 prio=1
src=3 pend=0 prio=1
src=4 pend=1 prio=1
Clearly something is not working well in the PLIC.