On Thu, Oct 15, 2020 at 11:18:18PM +0200, Greg Kurz wrote:
> While reading the code _again_ I spotted a memory leak and I realized
> that the realize/unrealize paths are uselessly complex and not really
> symmetrical.
>
> This series fixes the leak and re-shuffles the code to make it cleaner.
>
> Tested with 'make check', travis-ci and manual hotplug/unplug of CPU
> cores. Also tested error paths by simulating failures when creating
> interrupt presenters or when setting the vCPU id.
>
> v2: - enforce symmetry between realize and unrealize
> - unrealize vCPUs with qdev_unrealize()
> - one loop to create/realize and to unrealize/delete vCPUs
Applied to ppc-for-5.2.
>
> ---
>
> Greg Kurz (5):
> spapr: Fix leak of CPU machine specific data
> spapr: Unrealize vCPUs with qdev_unrealize()
> spapr: Drop spapr_delete_vcpu() unused argument
> spapr: Make spapr_cpu_core_unrealize() idempotent
> spapr: Simplify spapr_cpu_core_realize() and spapr_cpu_core_unrealize()
>
>
> accel/tcg/user-exec-stub.c | 4 ++
> hw/ppc/spapr_cpu_core.c | 69 ++++++++++++++++++---------------------
> target/ppc/translate_init.c.inc | 2 +
> 3 files changed, 37 insertions(+), 38 deletions(-)
>
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson