[for-5.0,0/3] ide: fix potential memory leaks (plus one via-ide bugfix)

Message

Mark Cave-Ayland March 24, 2020, 9:05 p.m. UTC

This was supposed to be a simple patchset to switch via-ide and cmd646-ide
over to use qdev gpio in the same way as Peter's patch did for sil3112, but
at the same time I spotted a silly mistake in my last set of via-ide
patches which is included as patch 1.

I'm not sure exactly why Coverity CID 1421984 isn't triggered by the
via-ide and cmd646-ide devices, however given the simplicity of the fix it
seems worth doing just to keep everything the same and ensure it won't
suddenly appear in future.

The via-ide changes were tested using the instructions provided by Zoltan
for MIPS fulong2e and PPC pegasos2, whilst the cmd646 change was tested
using one of my SPARC64 Linux test images.

Signed-off-by: Mark Cave-Ayland <[e-mail]>


Mark Cave-Ayland (3):
  via-ide: don't use PCI level for legacy IRQs
  via-ide: use qdev gpio rather than qemu_allocate_irqs()
  cmd646-ide: use qdev gpio rather than qemu_allocate_irqs()

 hw/ide/cmd646.c | 9 ++++-----
 hw/ide/via.c    | 7 ++++---
 2 files changed, 8 insertions(+), 8 deletions(-)