system/freebsd-src
system/freebsd-src
1
0
Fork 0
mirror of https://github.com/freebsd/freebsd-src synced 2024-10-15 04:43:53 +00:00
Code Issues Packages Projects Releases Wiki Activity

riscv: set kernel_pmap hart mask more precisely

Browse source 
In pmap_bootstrap(), we fill kernel_pmap->pm_active since it is
invariably active on all harts. However, this marks it as active even
for harts that don't exist in the system, which can cause issue when the
mask is passed to the SBI firmware via sbi_remote_sfence_vma().
Specifically, the SBI spec allows SBI_ERR_INVALID_PARAM to be returned
when an invalid hart is set in the mask.

The latest version of OpenSBI does not have this issue, but v0.6 does,
and this is triggering a recently added KASSERT in CI. Switch to only
setting bits in pm_active for harts that enter the system.

Reported by:	Jenkins
Reviewed by:	markj
Differential Revision:	https://reviews.freebsd.org/D27080
This commit is contained in:
Mitchell Horne 2020-11-05 00:52:52 +00:00
parent 01db2f5461
commit caaddb88e8
Notes: svn2git 2020-12-20 02:59:44 +00:00 
svn path=/head/; revision=367356
2 changed files with 10 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
sys/riscv/riscv/mp_machdep.c
View file

@ -46,6 +46,7 @@ __FBSDID("$FreeBSD$");
#include <sys/systm.h>
#include <sys/bus.h>
#include <sys/cpu.h>
#include <sys/cpuset.h>
#include <sys/kernel.h>
#include <sys/ktr.h>
#include <sys/malloc.h>
@ -266,6 +267,9 @@ init_secondary(uint64_t hart)
/* Enable external (PLIC) interrupts */
csr_set(sie, SIE_SEIE);
/* Activate this hart in the kernel pmap. */
CPU_SET_ATOMIC(hart, &kernel_pmap->pm_active);
/* Activate process 0's pmap. */
pmap_activate_boot(vmspace_pmap(proc0.p_vmspace));

7
sys/riscv/riscv/pmap.c
View file

@ -573,7 +573,12 @@ pmap_bootstrap(vm_offset_t l1pt, vm_paddr_t kernstart, vm_size_t kernlen)
rw_init(&pvh_global_lock, "pmap pv global");
CPU_FILL(&kernel_pmap->pm_active);
/*
* Set the current CPU as active in the kernel pmap. Secondary cores
* will add themselves later in init_secondary(). The SBI firmware
* may rely on this mask being precise, so CPU_FILL() is not used.
*/
CPU_SET(PCPU_GET(hart), &kernel_pmap->pm_active);
/* Assume the address we were loaded to is a valid physical address. */
min_pa = max_pa = kernstart;