| Message ID | 157650846660.354886.16810288202617432561.stgit@bahia.lan |
|---|---|
| Headers | show |
| Series | cpu: Clarify overloading of reset QOM methods | expand |
On 16/12/19 16:01, Greg Kurz wrote: > Each cpu subclass overloads the reset method of its parent class with > its own. But since it needs to call the parent method as well, it keeps > a parent_reset pointer to do so. This causes the same not very explicit > boiler plate to be duplicated all around the place: > > pcc->parent_reset = cc->reset; > cc->reset = ppc_cpu_reset; > > A similar concern was addressed some time back by Philippe Mathieu-Daudé > in qdev, with the addition of device_class_set_parent_reset() and friends: > > https://git.qemu.org/?p=qemu.git;a=commit;h=46795cf2e2f6 > https://git.qemu.org/?p=qemu.git;a=commit;h=bf853881690d > > Follow the same approach with cpus. > > Changes in v3: > - drop 'cpu: Introduce CPUReset callback typedef' patch which isn't needed > and makes the code less clear. This changes the declaration of the helper > in 'cpu: Introduce cpu_class_set_parent_reset()', but it is minor so I > keep the Reviewed-by and Acked-by tags. > > Changes in v2: > - added Reviewed-by and Acked-by tags > - rebased on top of https://github.com/cohuck/qemu.git s390-next > SHA1 dd6252f035a2 > > -- > Greg > > --- > > Greg Kurz (2): > cpu: Introduce cpu_class_set_parent_reset() > cpu: Use cpu_class_set_parent_reset() > > > hw/core/cpu.c | 8 ++++++++ > include/hw/core/cpu.h | 4 ++++ > target/arm/cpu.c | 3 +-- > target/cris/cpu.c | 3 +-- > target/i386/cpu.c | 3 +-- > target/lm32/cpu.c | 3 +-- > target/m68k/cpu.c | 3 +-- > target/microblaze/cpu.c | 3 +-- > target/mips/cpu.c | 3 +-- > target/moxie/cpu.c | 3 +-- > target/nios2/cpu.c | 3 +-- > target/openrisc/cpu.c | 3 +-- > target/ppc/translate_init.inc.c | 3 +-- > target/riscv/cpu.c | 3 +-- > target/s390x/cpu.c | 3 +-- > target/sh4/cpu.c | 3 +-- > target/sparc/cpu.c | 3 +-- > target/tilegx/cpu.c | 3 +-- > target/tricore/cpu.c | 3 +-- > target/xtensa/cpu.c | 3 +-- > 20 files changed, 30 insertions(+), 36 deletions(-) > Queued, thanks. Paolo
Each cpu subclass overloads the reset method of its parent class with its own. But since it needs to call the parent method as well, it keeps a parent_reset pointer to do so. This causes the same not very explicit boiler plate to be duplicated all around the place: pcc->parent_reset = cc->reset; cc->reset = ppc_cpu_reset; A similar concern was addressed some time back by Philippe Mathieu-Daudé in qdev, with the addition of device_class_set_parent_reset() and friends: https://git.qemu.org/?p=qemu.git;a=commit;h=46795cf2e2f6 https://git.qemu.org/?p=qemu.git;a=commit;h=bf853881690d Follow the same approach with cpus. Changes in v3: - drop 'cpu: Introduce CPUReset callback typedef' patch which isn't needed and makes the code less clear. This changes the declaration of the helper in 'cpu: Introduce cpu_class_set_parent_reset()', but it is minor so I keep the Reviewed-by and Acked-by tags. Changes in v2: - added Reviewed-by and Acked-by tags - rebased on top of https://github.com/cohuck/qemu.git s390-next SHA1 dd6252f035a2 -- Greg --- Greg Kurz (2): cpu: Introduce cpu_class_set_parent_reset() cpu: Use cpu_class_set_parent_reset() hw/core/cpu.c | 8 ++++++++ include/hw/core/cpu.h | 4 ++++ target/arm/cpu.c | 3 +-- target/cris/cpu.c | 3 +-- target/i386/cpu.c | 3 +-- target/lm32/cpu.c | 3 +-- target/m68k/cpu.c | 3 +-- target/microblaze/cpu.c | 3 +-- target/mips/cpu.c | 3 +-- target/moxie/cpu.c | 3 +-- target/nios2/cpu.c | 3 +-- target/openrisc/cpu.c | 3 +-- target/ppc/translate_init.inc.c | 3 +-- target/riscv/cpu.c | 3 +-- target/s390x/cpu.c | 3 +-- target/sh4/cpu.c | 3 +-- target/sparc/cpu.c | 3 +-- target/tilegx/cpu.c | 3 +-- target/tricore/cpu.c | 3 +-- target/xtensa/cpu.c | 3 +-- 20 files changed, 30 insertions(+), 36 deletions(-)