diff options
| author | peter <peter@FreeBSD.org> | 2003-07-21 20:59:27 +0000 |
|---|---|---|
| committer | peter <peter@FreeBSD.org> | 2003-07-21 20:59:27 +0000 |
| commit | 57defe6f29377d5df5ee2d95638673b88a928681 (patch) | |
| tree | 26895b6e6a6ea1ab74e45c03e03d1531b64980dc /sys/i386/acpica | |
| parent | b03a1147332ddbe41f75a85dd618e5b77d7554ef (diff) | |
| download | FreeBSD-src-57defe6f29377d5df5ee2d95638673b88a928681.zip FreeBSD-src-57defe6f29377d5df5ee2d95638673b88a928681.tar.gz | |
Commit Ian Dowse's workaround for acpi resume resetting after the
LAZY_SWITCH changes. He pointed out the acpi code sets up an identity
mapping in the current vmspace and that got messed up by the %cr3 being
out of sync with the current page directory. As a workaround, restore
%cr3 across the sleep/resume. A more complete fix would be to undo the
lazy state and clear the pm_active bit from the borrowed pmap, but this
works and people are currently hurting. I'll clean this up.
This is mostly Ian's patch, plus a PAE tweak from me.
Diffstat (limited to 'sys/i386/acpica')
| -rw-r--r-- | sys/i386/acpica/acpi_wakeup.c | 8 |
1 files changed, 8 insertions, 0 deletions
diff --git a/sys/i386/acpica/acpi_wakeup.c b/sys/i386/acpica/acpi_wakeup.c index 81a396a..472438f 100644 --- a/sys/i386/acpica/acpi_wakeup.c +++ b/sys/i386/acpica/acpi_wakeup.c @@ -193,6 +193,7 @@ acpi_sleep_machdep(struct acpi_softc *sc, int state) static vm_page_t opage = NULL; int ret = 0; int pteobj_allocated = 0; + u_int32_t cr3; u_long ef; struct proc *p; @@ -209,6 +210,12 @@ acpi_sleep_machdep(struct acpi_softc *sc, int state) if ((p = curproc) == NULL) p = &proc0; pm = vmspace_pmap(p->p_vmspace); + cr3 = rcr3(); +#ifdef PAE + load_cr3(vtophys(pm->pm_pdpt)); +#else + load_cr3(vtophys(pm->pm_pdir)); +#endif if (pm->pm_pteobj == NULL) { pm->pm_pteobj = vm_object_allocate(OBJT_DEFAULT, PTDPTDI + 1); pteobj_allocated = 1; @@ -295,6 +302,7 @@ out: vm_object_deallocate(pm->pm_pteobj); pm->pm_pteobj = NULL; } + load_cr3(cr3); write_eflags(ef); |
