On 03.01.19 18:26, Ralf Ramsauer wrote:
Hi Jan,
On 1/3/19 5:45 PM, Jan Kiszka wrote:
On 03.01.19 00:12, Ralf Ramsauer wrote:
This is the second step to get rid of arch_handle_exit(). There's no
need to
call vmreturn() from arch_handle_exit(). Let's move this to assembly.
Signed-off-by: Ralf Ramsauer <[email protected]>
---
hypervisor/arch/arm64/entry.S | 4 +++-
hypervisor/arch/arm64/traps.c | 2 +-
2 files changed, 4 insertions(+), 2 deletions(-)
diff --git a/hypervisor/arch/arm64/entry.S
b/hypervisor/arch/arm64/entry.S
index 97033e91..464386d9 100644
--- a/hypervisor/arch/arm64/entry.S
+++ b/hypervisor/arch/arm64/entry.S
@@ -315,7 +315,8 @@ bootstrap_vectors:
mov x30, xzr
mov x0, sp
bl arch_handle_exit
- b .
+ /* take the fast exit path, sp is already in place */
+ b __vmreturn
.endm
.align 11
@@ -413,6 +414,7 @@ shutdown_el2:
vmreturn:
/* x0: union registers* */
mov sp, x0
+__vmreturn:
You missed my comment:
The only remaining caller of vmreturn is arch_cpu_activate_vmm() which
is not really a hot-path.
Ack, understood.
I would suggest to make the common case jump
free and let vmreturn jump here instead.
I don't get this comment, could you clarify it a bit? What do you mean
with "common case jump free"?
If you mean the hyp_vectors with common case (fast path), then I can't
get them jump free, as the interrupt vectors don't have enough space for
the code.
Ah, the jump starting point is a macro! Ok, then my comment makes no sense
anymore.
Jan
--
Siemens AG, Corporate Technology, CT RDA IOT SES-DE
Corporate Competence Center Embedded Linux
--
You received this message because you are subscribed to the Google Groups
"Jailhouse" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
