[PATCH v2 1/3] ahci: Fix leak of IRQState

[Fabiano Rosas <farosas@suse.de>]

ASAN spotted an indirect leak in ahci-test:

Indirect leak of 576 byte(s) in 6 object(s) allocated from:
    #0 0x556112a08ca0 in malloc
    #1 0x7f21591e595d in g_malloc
    #2 0x7f21591ffe38 in g_slice_alloc
    #3 0x7f21591c988d in g_hash_table_new_full
    #4 0x556114d756f1 in object_initialize_with_type ../qom/object.c:506:23
    #5 0x556114d77590 in object_new_with_type ../qom/object.c:706:5
    #6 0x556114d777e8 in object_new ../qom/object.c:722:12
    #7 0x556114d6d2be in qemu_allocate_irq ../hw/core/irq.c:94:25
    #8 0x556114d6d1bc in qemu_extend_irqs ../hw/core/irq.c:82:16
    #9 0x556114d6d329 in qemu_allocate_irqs ../hw/core/irq.c:89:12
    #10 0x5561135da9bd in ahci_realize ../hw/ide/ahci.c:1644:12
    #11 0x55611360156a in pci_ich9_ahci_realize ../hw/ide/ich.c:132:5

Although ahci_realize() frees the return of qemu_allocate_irqs() right
away, the individual IRQStates are still left around.

Issue qemu_free_irq() at ahci_uninit().

Reviewed-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
---
 hw/ide/ahci.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/hw/ide/ahci.c b/hw/ide/ahci.c
index 6b04762c4a..86f6297dc9 100644
--- a/hw/ide/ahci.c
+++ b/hw/ide/ahci.c
@@ -1688,6 +1688,7 @@ void ahci_uninit(AHCIState *s)
             }
             ide_exit(ide_state);
         }
+        qemu_free_irq(ad->port.irq);
         object_unparent(OBJECT(&ad->port));
     }
 
-- 
2.53.0