[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH 2/2] bulk: Prefer fast cpu_env() over slower CPU QOM cast macro



On 26/1/24 18:09, Thomas Huth wrote:
On 25/01/2024 17.56, Philippe Mathieu-Daudé wrote:
Mechanical patch produced running the command documented
in scripts/coccinelle/cpu_env.cocci_template header.

Signed-off-by: Philippe Mathieu-Daudé <philmd@xxxxxxxxxx>
---

  114 files changed, 273 insertions(+), 548 deletions(-)

A huge patch ... I wonder whether it would make sense to split it up by target architecture to ease the review?

...
diff --git a/hw/i386/vmmouse.c b/hw/i386/vmmouse.c
index a8d014d09a..eb0613bfbe 100644
--- a/hw/i386/vmmouse.c
+++ b/hw/i386/vmmouse.c
@@ -74,8 +74,7 @@ struct VMMouseState {
  static void vmmouse_get_data(uint32_t *data)
  {
-    X86CPU *cpu = X86_CPU(current_cpu);
-    CPUX86State *env = &cpu->env;
+    CPUX86State *env = cpu_env(CPU(current_cpu));

No need for the CPU() cast here, current_cpu is already
of type "CPUState *".

Yes, Paolo noticed and I fixed for v2.

I'll stop here, please respin with the cpu_env(CPU(current_cpu)) fixed to
cpu_env(current_cpu), and please split the patch by target CPU types.

Well I don't know, this is an reproducible mechanical patch..
But indeed as Paolo you found an optimization so worth not making
human review a pain.

I was about to post v2 but I'll see how to split.

Thanks for the review!

Phil.



 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.