[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v2 12/58] qdev: Convert uses of qdev_set_parent_bus() manuall
From: |
Markus Armbruster |
Subject: |
Re: [PATCH v2 12/58] qdev: Convert uses of qdev_set_parent_bus() manually |
Date: |
Tue, 09 Jun 2020 12:00:22 +0200 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/26.3 (gnu/linux) |
Philippe Mathieu-Daudé <philmd@redhat.com> writes:
> On 5/29/20 3:44 PM, Markus Armbruster wrote:
>> Same transformation as in the previous commit. Manual, because
>> convincing Coccinelle to transform these cases is somewhere between
>> not worthwhile and infeasible (at least for me).
>>
>> Signed-off-by: Markus Armbruster <armbru@redhat.com>
>> ---
>> hw/pci-host/prep.c | 3 +--
>> hw/ppc/pnv.c | 6 ++----
>> hw/s390x/sclp.c | 10 ++++------
>> 3 files changed, 7 insertions(+), 12 deletions(-)
>>
>> diff --git a/hw/pci-host/prep.c b/hw/pci-host/prep.c
>> index c821ef889d..42c7e63a60 100644
>> --- a/hw/pci-host/prep.c
>> +++ b/hw/pci-host/prep.c
>> @@ -268,7 +268,7 @@ static void raven_pcihost_realizefn(DeviceState *d,
>> Error **errp)
>> memory_region_add_subregion(address_space_mem, 0xbffffff0,
>> &s->pci_intack);
>>
>> /* TODO Remove once realize propagates to child devices. */
>> - object_property_set_bool(OBJECT(&s->pci_dev), true, "realized", errp);
>> + qdev_realize(DEVICE(&s->pci_dev), BUS(&s->pci_bus), errp);
>> }
>>
>> static void raven_pcihost_initfn(Object *obj)
>> @@ -308,7 +308,6 @@ static void raven_pcihost_initfn(Object *obj)
>>
>> object_initialize(&s->pci_dev, sizeof(s->pci_dev),
>> TYPE_RAVEN_PCI_DEVICE);
>> pci_dev = DEVICE(&s->pci_dev);
>> - qdev_set_parent_bus(pci_dev, BUS(&s->pci_bus));
>> object_property_set_int(OBJECT(&s->pci_dev), PCI_DEVFN(0, 0), "addr",
>> NULL);
>> qdev_prop_set_bit(pci_dev, "multifunction", false);
>> diff --git a/hw/ppc/pnv.c b/hw/ppc/pnv.c
>> index 8562af3fe0..e0588285a2 100644
>> --- a/hw/ppc/pnv.c
>> +++ b/hw/ppc/pnv.c
>> @@ -1212,12 +1212,11 @@ static void pnv_chip_power8_realize(DeviceState
>> *dev, Error **errp)
>> object_property_set_int(OBJECT(phb), i, "index", &error_fatal);
>> object_property_set_int(OBJECT(phb), chip->chip_id, "chip-id",
>> &error_fatal);
>> - object_property_set_bool(OBJECT(phb), true, "realized", &local_err);
>> + qdev_realize(DEVICE(phb), NULL, &local_err);
>> if (local_err) {
>> error_propagate(errp, local_err);
>> return;
>> }
>> - qdev_set_parent_bus(DEVICE(phb), sysbus_get_default());
>>
>> /* Populate the XSCOM address space. */
>> pnv_xscom_add_subregion(chip,
>> @@ -1422,12 +1421,11 @@ static void pnv_chip_power9_phb_realize(PnvChip
>> *chip, Error **errp)
>> object_property_set_int(obj, PNV_PHB4_DEVICE_ID, "device-id",
>> &error_fatal);
>> object_property_set_link(obj, OBJECT(stack), "stack",
>> &error_abort);
>> - object_property_set_bool(obj, true, "realized", &local_err);
>> + qdev_realize(DEVICE(obj), NULL, &local_err);
>> if (local_err) {
>> error_propagate(errp, local_err);
>> return;
>> }
>> - qdev_set_parent_bus(DEVICE(obj), sysbus_get_default());
>>
>> /* Populate the XSCOM address space. */
>> pnv_xscom_add_subregion(chip,
>> diff --git a/hw/s390x/sclp.c b/hw/s390x/sclp.c
>> index 20aca30ac4..40e27a8cb4 100644
>> --- a/hw/s390x/sclp.c
>> +++ b/hw/s390x/sclp.c
>> @@ -333,17 +333,15 @@ static void sclp_realize(DeviceState *dev, Error
>> **errp)
>> uint64_t hw_limit;
>> int ret;
>>
>> - object_property_set_bool(OBJECT(sclp->event_facility), true, "realized",
>> - &err);
>> - if (err) {
>> - goto out;
>> - }
>> /*
>> * qdev_device_add searches the sysbus for TYPE_SCLP_EVENTS_BUS. As long
>> * as we can't find a fitting bus via the qom tree, we have to add the
>> * event facility to the sysbus, so e.g. a sclp console can be created.
>> */
>> - qdev_set_parent_bus(DEVICE(sclp->event_facility), sysbus_get_default());
>> + qdev_realize(DEVICE(sclp->event_facility), NULL, &err);
>
> I agree with the API change, but I find using sysbus_get_default() way
> clearer than using NULL. I'v go the other way around, replace bus=NULL
> by an explicit bus=sysbus_get_default() in our codebase.
I agree with you on sysbus_get_default() being clearer than NULL. I
actually get rid of NULL in this series, but it takes me many more
patches to get there. Starts at PATCH 45.
Of course, I must not expect reviewers to look three dozen patches
ahead.
> Can be done later, I know... But we can also keep sysbus_get_default()
> in this patch ;)
This patch's purpose is getting rid of separate qdev_set_parent_bus(),
in part to ease the conversion from qdev_realize() to sysbus_realize().
> Whichever you prefer (mine is keeping it):
> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Thanks!
>
>> + if (err) {
>> + goto out;
>> + }
>>
>> ret = s390_set_memory_limit(machine->maxram_size, &hw_limit);
>> if (ret == -E2BIG) {
>>