* Re: Question about memory trace with execlog plugin
[not found] <CyPTzkaf9gPwB4yBJ2L2v7OEa1ifWfc6Ys8CFowNoUee6ap4yEu3YThfluX0AObgYU4UcWQFHNbfWrw0kZ-wNQF3gER8mhXZg2Vc74SOstw=@protonmail.com>
@ 2025-04-05 14:25 ` Philippe Mathieu-Daudé
0 siblings, 0 replies; only message in thread
From: Philippe Mathieu-Daudé @ 2025-04-05 14:25 UTC (permalink / raw)
To: Yannis Bolliger, qemu-discuss@nongnu.org
Cc: QEMU Developers, Pierrick Bouvier, Alex Bennée
Cc'ing plugin experts
On 5/4/25 13:25, Yannis Bolliger via wrote:
> Hi everyone,
>
> I am using the execlog plugin to generate a trace of memory accesses, in
> particular I am interested in the physical RAM addresses.
> Now I'd like to identify which of these memory accesses are related to a
> kernel buffer -> user buffer copy operation.
> To that end I am logging the physical addresses as obtained by
> page_to_phys (+ the offset of the RAM in the memory map) in the kernel
> before the copy is done.
> I'd expect to find memory accesses in my trace that go to the buffer
> addresses I logged in the kernel, but I cannot find any.
>
> Is it possible to correlate these addresses? What do I need to look out
> for?
> I may have a wrong understanding of how things work.
>
> Any help is appreciated,
> Yannis
>
> Some background info:
>
> * qemu-system-aarch64 on an x86_64 host.
> * android kernel (cuttlefish)
>
^ permalink raw reply [flat|nested] only message in thread