From 4c315c27661502a0813b129e41c0bf640c34a8d6 Mon Sep 17 00:00:00 2001 From: Markus Armbruster Date: Thu, 1 Oct 2015 10:59:58 +0200 Subject: qdev: Protect device-list-properties against broken devices MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Several devices don't survive object_unref(object_new(T)): they crash or hang during cleanup, or they leave dangling pointers behind. This breaks at least device-list-properties, because qmp_device_list_properties() needs to create a device to find its properties. Broken in commit f4eb32b "qmp: show QOM properties in device-list-properties", v2.1. Example reproducer: $ qemu-system-aarch64 -nodefaults -display none -machine none -S -qmp stdio {"QMP": {"version": {"qemu": {"micro": 50, "minor": 4, "major": 2}, "package": ""}, "capabilities": []}} { "execute": "qmp_capabilities" } {"return": {}} { "execute": "device-list-properties", "arguments": { "typename": "pxa2xx-pcmcia" } } qemu-system-aarch64: /home/armbru/work/qemu/memory.c:1307: memory_region_finalize: Assertion `((&mr->subregions)->tqh_first == ((void *)0))' failed. Aborted (core dumped) [Exit 134 (SIGABRT)] Unfortunately, I can't fix the problems in these devices right now. Instead, add DeviceClass member cannot_destroy_with_object_finalize_yet to mark them: * Hang during cleanup (didn't debug, so I can't say why): "realview_pci", "versatile_pci". * Dangling pointer in cpus: most CPUs, plus "allwinner-a10", "digic", "fsl,imx25", "fsl,imx31", "xlnx,zynqmp", because they create such CPUs * Assert kvm_enabled(): "host-x86_64-cpu", host-i386-cpu", "host-powerpc64-cpu", "host-embedded-powerpc-cpu", "host-powerpc-cpu" (the powerpc ones can't currently reach the assertion, because the CPUs are only registered when KVM is enabled, but the assertion is arguably in the wrong place all the same) Make qmp_device_list_properties() fail cleanly when the device is so marked. This improves device-list-properties from "crashes, hangs or leaves dangling pointers behind" to "fails". Not a complete fix, just a better-than-nothing work-around. In the above reproducer, device-list-properties now fails with "Can't list properties of device 'pxa2xx-pcmcia'". This also protects -device FOO,help, which uses the same machinery since commit ef52358 "qdev-monitor: include QOM properties in -device FOO, help output", v2.2. Example reproducer: $ qemu-system-aarch64 -machine none -device pxa2xx-pcmcia,help Before: qemu-system-aarch64: .../memory.c:1307: memory_region_finalize: Assertion `((&mr->subregions)->tqh_first == ((void *)0))' failed. After: Can't list properties of device 'pxa2xx-pcmcia' Cc: "Andreas Färber" Cc: "Edgar E. Iglesias" Cc: Alexander Graf Cc: Anthony Green Cc: Aurelien Jarno Cc: Bastian Koppelmann Cc: Blue Swirl Cc: Eduardo Habkost Cc: Guan Xuetao Cc: Jia Liu Cc: Leon Alrae Cc: Mark Cave-Ayland Cc: Max Filippov Cc: Michael Walle Cc: Paolo Bonzini Cc: Peter Maydell Cc: Richard Henderson Cc: qemu-ppc@nongnu.org Cc: qemu-stable@nongnu.org Signed-off-by: Markus Armbruster Reviewed-by: Eduardo Habkost Message-Id: <1443689999-12182-10-git-send-email-armbru@redhat.com> --- target-unicore32/cpu.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'target-unicore32') diff --git a/target-unicore32/cpu.c b/target-unicore32/cpu.c index fc451a1a35..e5252ebaf8 100644 --- a/target-unicore32/cpu.c +++ b/target-unicore32/cpu.c @@ -155,6 +155,13 @@ static void uc32_cpu_class_init(ObjectClass *oc, void *data) cc->get_phys_page_debug = uc32_cpu_get_phys_page_debug; #endif dc->vmsd = &vmstate_uc32_cpu; + + /* + * Reason: uc32_cpu_initfn() calls cpu_exec_init(), which saves + * the object in cpus -> dangling pointer after final + * object_unref(). + */ + dc->cannot_destroy_with_object_finalize_yet = true; } static void uc32_register_cpu_type(const UniCore32CPUInfo *info) -- cgit v1.2.3