* [PATCH 1/3] lguest: fix sense if IF flag on interrupt injection
@ 2007-07-20 12:11 Rusty Russell
2007-07-20 12:12 ` [PATCH 2/3] lguest: trivial: We now have asm/processor-flags.h, so use it Rusty Russell
0 siblings, 1 reply; 3+ messages in thread
From: Rusty Russell @ 2007-07-20 12:11 UTC (permalink / raw)
To: Linus Torvalds; +Cc: lkml - Kernel Mailing List, virtualization
The sense of the IF bit is backwards in the host interrupt handling.
This means we always save "IF=1" on the stack when injecting an
interrupt. It turns out this is almost always correct (unless the
guest is taking a page fault in an interrupt due to an unpopulated
vmalloc mapping), so went unnoticed.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
---
drivers/lguest/interrupts_and_traps.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff -r 209f5cd5cda5 drivers/lguest/interrupts_and_traps.c
--- a/drivers/lguest/interrupts_and_traps.c Fri Jul 20 14:53:40 2007 +1000
+++ b/drivers/lguest/interrupts_and_traps.c Fri Jul 20 21:34:06 2007 +1000
@@ -38,12 +38,12 @@ static void set_guest_interrupt(struct l
ss = lg->regs->ss;
}
- /* We use IF bit in eflags to indicate whether irqs were disabled
- (it's always 0, since irqs are enabled when guest is running). */
+ /* We use IF bit in eflags to indicate whether irqs were enabled
+ (it's always 1, since irqs are enabled when guest is running). */
eflags = lg->regs->eflags;
- if (get_user(irq_enable, &lg->lguest_data->irq_enabled))
- irq_enable = 0;
- eflags |= (irq_enable & X86_EFLAGS_IF);
+ if (get_user(irq_enable, &lg->lguest_data->irq_enabled) == 0
+ && !(irq_enable & X86_EFLAGS_IF))
+ eflags &= ~X86_EFLAGS_IF;
push_guest_stack(lg, &gstack, eflags);
push_guest_stack(lg, &gstack, lg->regs->cs);
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH 2/3] lguest: trivial: We now have asm/processor-flags.h, so use it.
2007-07-20 12:11 [PATCH 1/3] lguest: fix sense if IF flag on interrupt injection Rusty Russell
@ 2007-07-20 12:12 ` Rusty Russell
2007-07-20 12:15 ` [PATCH 3/3] lguest: override sched_clock Rusty Russell
0 siblings, 1 reply; 3+ messages in thread
From: Rusty Russell @ 2007-07-20 12:12 UTC (permalink / raw)
To: Linus Torvalds; +Cc: lkml - Kernel Mailing List, virtualization
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
diff -r f41cd1b8d7ef drivers/lguest/lguest_asm.S
--- a/drivers/lguest/lguest_asm.S Fri Jul 20 13:46:40 2007 +1000
+++ b/drivers/lguest/lguest_asm.S Fri Jul 20 14:17:04 2007 +1000
@@ -2,9 +2,7 @@
#include <linux/lguest.h>
#include <asm/asm-offsets.h>
#include <asm/thread_info.h>
-
-/* FIXME: Once asm/processor-flags.h goes in, include that */
-#define X86_EFLAGS_IF 0x00000200
+#include <asm/processor-flags.h>
/*
* This is where we begin: we have a magic signature which the launcher looks
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH 3/3] lguest: override sched_clock
2007-07-20 12:12 ` [PATCH 2/3] lguest: trivial: We now have asm/processor-flags.h, so use it Rusty Russell
@ 2007-07-20 12:15 ` Rusty Russell
0 siblings, 0 replies; 3+ messages in thread
From: Rusty Russell @ 2007-07-20 12:15 UTC (permalink / raw)
To: Linus Torvalds; +Cc: lkml - Kernel Mailing List, virtualization
Guests currently use the default scheduler clock: this means they
always use jiffies even if TSC is actually available. It doesn't make
any noticeable difference here, but it's a better thing to do.
Also remove commented-out asm/sched-clock.h from -mm tree.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
---
drivers/lguest/lguest.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff -r d9973bf15010 drivers/lguest/lguest.c
--- a/drivers/lguest/lguest.c Fri Jul 20 21:40:49 2007 +1000
+++ b/drivers/lguest/lguest.c Fri Jul 20 22:08:10 2007 +1000
@@ -39,7 +39,6 @@
#include <asm/e820.h>
#include <asm/mce.h>
#include <asm/io.h>
-//#include <asm/sched-clock.h>
/* Declarations for definitions in lguest_guest.S */
extern char lguest_noirq_start[], lguest_noirq_end[];
@@ -57,6 +56,7 @@ struct lguest_data lguest_data = {
.blocked_interrupts = { 1 }, /* Block timer interrupts */
};
struct lguest_device_desc *lguest_devices;
+static cycle_t clock_base;
static enum paravirt_lazy_mode lazy_mode;
static void lguest_lazy_mode(enum paravirt_lazy_mode mode)
@@ -363,6 +363,11 @@ static struct clocksource lguest_clock =
.read = lguest_clock_read,
};
+static unsigned long long lguest_sched_clock(void)
+{
+ return cyc2ns(&lguest_clock, lguest_clock_read() - clock_base);
+}
+
/* We also need a "struct clock_event_device": Linux asks us to set it to go
* off some time in the future. Actually, James Morris figured all this out, I
* just applied the patch. */
@@ -439,6 +444,7 @@ static void lguest_time_init(void)
lguest_clock.mult = (((u64)NSEC_PER_SEC<<8)/ACTHZ) << 8;
lguest_clock.mask = CLOCKSOURCE_MASK(32);
}
+ clock_base = lguest_clock_read();
clocksource_register(&lguest_clock);
/* We can't set cpumask in the initializer: damn C limitations! */
@@ -584,6 +590,7 @@ __init void lguest_init(void *boot)
paravirt_ops.time_init = lguest_time_init;
paravirt_ops.set_lazy_mode = lguest_lazy_mode;
paravirt_ops.wbinvd = lguest_wbinvd;
+ paravirt_ops.sched_clock = lguest_sched_clock;
hcall(LHCALL_LGUEST_INIT, __pa(&lguest_data), 0, 0);
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2007-07-20 12:15 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-07-20 12:11 [PATCH 1/3] lguest: fix sense if IF flag on interrupt injection Rusty Russell
2007-07-20 12:12 ` [PATCH 2/3] lguest: trivial: We now have asm/processor-flags.h, so use it Rusty Russell
2007-07-20 12:15 ` [PATCH 3/3] lguest: override sched_clock Rusty Russell
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox