Re: [Qemu-devel] [RFC] vfio: migration to trace points

[Alexander Graf]

On 03.09.14 10:45, Eric Auger wrote:
> This patch removes all DPRINTF and replace them by trace points.
> A few DPRINTF used in error cases were transformed into error_report.
> 
> Signed-off-by: Eric Auger <address@hidden>
> 
> ---
> 
> - __func__ is removed since trace point name does the same job
> - HWADDR_PRIx were replaced by PRIx64
> 
> Besides those changes format strings were kept the same. in few
> cases however I was forced to change them due to parsing errors
> (always related to parenthesis handling). This is indicated in
> trace-events. Cases than are not correctly handled are given below:
> - "(%04x:%02x:%02x.%x)" need to be replaced by " (%04x:%02x:%02x.%x)"
> - "%s read(%04x:%02x:%02x.%x:BAR%d+0x%"PRIx64", %d) = 0x%"PRIx64 ->
>   "%s read(%04x:%02x:%02x.%x:BAR%d+0x%"PRIx64", %d = 0x%"PRIx64 ->
> - "%s write(%04x:%02x:%02x.%x:BAR%d+0x%"PRIx64", 0x%"PRIx64", %d)"
>   "%s write(%04x:%02x:%02x.%x:BAR%d+0x%"PRIx64", 0x%"PRIx64", %d"
> This is a temporary fix.
> 
> - This leads to a too large amount of trace points which may not be
> eligible as trace points - I don't know?-
> - this transformation just is tested compiled on PCI. Tested on platform
>   qemu configured with --enable-trace-backends=stderr
> - in future, format strings and calls may be simplified by using a single
>   name argument instead of domain, bus, slot, function.

I think it's a nice step into the right direction.


Alex