From: "Philippe Mathieu-Daudé" <philmd@linaro.org>
To: Thomas Huth <thuth@redhat.com>, qemu-devel@nongnu.org
Cc: "Richard Henderson" <richard.henderson@linaro.org>,
"Eric Farman" <farman@linux.ibm.com>,
"Peter Xu" <peterx@redhat.com>,
"Michael S. Tsirkin" <mst@redhat.com>,
"Halil Pasic" <pasic@linux.ibm.com>,
"Jiaxun Yang" <jiaxun.yang@flygoat.com>,
"Strahinja Jankovic" <strahinja.p.jankovic@gmail.com>,
"Eduardo Habkost" <eduardo@habkost.net>,
"Sergio Lopez" <slp@redhat.com>,
"Peter Maydell" <peter.maydell@linaro.org>,
"Marcel Apfelbaum" <marcel.apfelbaum@gmail.com>,
"Jason Wang" <jasowang@redhat.com>,
qemu-arm@nongnu.org, qemu-s390x@nongnu.org,
"Ilya Leoshkevich" <iii@linux.ibm.com>,
"Song Gao" <gaosong@loongson.cn>,
"Huacai Chen" <chenhuacai@kernel.org>,
"Beniamino Galvani" <b.galvani@gmail.com>,
"Christian Borntraeger" <borntraeger@linux.ibm.com>,
"David Hildenbrand" <david@redhat.com>,
"Daniel P. Berrangé" <berrange@redhat.com>,
"Paolo Bonzini" <pbonzini@redhat.com>
Subject: Re: [PATCH 05/12] hw/pci-host/bonito: Do not use SysBus API to map local MMIO region
Date: Wed, 18 Oct 2023 20:23:54 +0200 [thread overview]
Message-ID: <227c8351-75d7-db1d-aeab-1e6fb2db22c0@linaro.org> (raw)
In-Reply-To: <6020a0cd-19a8-4cf7-a99d-781854464887@redhat.com>
On 18/10/23 18:05, Thomas Huth wrote:
> On 18/10/2023 16.11, Philippe Mathieu-Daudé wrote:
>> There is no point in exposing an internal MMIO region via
>> SysBus and directly mapping it in the very same device.
>>
>> Just map it without using the SysBus API.
>>
>> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
>> ---
>> hw/pci-host/bonito.c | 29 ++++++++++++++---------------
>> 1 file changed, 14 insertions(+), 15 deletions(-)
>>
>> diff --git a/hw/pci-host/bonito.c b/hw/pci-host/bonito.c
>> index ee6cb85e97..3b803bcad3 100644
>> --- a/hw/pci-host/bonito.c
>> +++ b/hw/pci-host/bonito.c
>> @@ -654,7 +654,6 @@ static void bonito_host_realize(DeviceState *dev,
>> Error **errp)
>> static void bonito_pci_realize(PCIDevice *dev, Error **errp)
>> {
>> PCIBonitoState *s = PCI_BONITO(dev);
>> - SysBusDevice *sysbus = SYS_BUS_DEVICE(s->pcihost);
>> PCIHostState *phb = PCI_HOST_BRIDGE(s->pcihost);
>> BonitoState *bs = s->pcihost;
>> MemoryRegion *pcimem_alias = g_new(MemoryRegion, 1);
>> @@ -668,48 +667,48 @@ static void bonito_pci_realize(PCIDevice *dev,
>> Error **errp)
>> /* set the north bridge register mapping */
>> memory_region_init_io(&s->iomem, OBJECT(s), &bonito_ops, s,
>> "north-bridge-register",
>> BONITO_INTERNAL_REG_SIZE);
>> - sysbus_init_mmio(sysbus, &s->iomem);
>> - sysbus_mmio_map(sysbus, 0, BONITO_INTERNAL_REG_BASE);
>> + memory_region_add_subregion(get_system_memory(),
>> BONITO_INTERNAL_REG_BASE,
>> + &s->iomem);
>> /* set the north bridge pci configure mapping */
>> memory_region_init_io(&phb->conf_mem, OBJECT(s),
>> &bonito_pciconf_ops, s,
>> "north-bridge-pci-config",
>> BONITO_PCICONFIG_SIZE);
>> - sysbus_init_mmio(sysbus, &phb->conf_mem);
>> - sysbus_mmio_map(sysbus, 1, BONITO_PCICONFIG_BASE);
>> + memory_region_add_subregion(get_system_memory(),
>> BONITO_PCICONFIG_BASE,
>> + &phb->conf_mem);
>> /* set the south bridge pci configure mapping */
>> memory_region_init_io(&phb->data_mem, OBJECT(s),
>> &bonito_spciconf_ops, s,
>> "south-bridge-pci-config",
>> BONITO_SPCICONFIG_SIZE);
>> - sysbus_init_mmio(sysbus, &phb->data_mem);
>> - sysbus_mmio_map(sysbus, 2, BONITO_SPCICONFIG_BASE);
>> + memory_region_add_subregion(get_system_memory(),
>> BONITO_SPCICONFIG_BASE,
>> + &phb->data_mem);
>> create_unimplemented_device("bonito", BONITO_REG_BASE,
>> BONITO_REG_SIZE);
>> memory_region_init_io(&s->iomem_ldma, OBJECT(s),
>> &bonito_ldma_ops, s,
>> "ldma", 0x100);
>> - sysbus_init_mmio(sysbus, &s->iomem_ldma);
>> - sysbus_mmio_map(sysbus, 3, 0x1fe00200);
>> + memory_region_add_subregion(get_system_memory(), 0x1fe00200,
>> + &s->iomem_ldma);
>> /* PCI copier */
>> memory_region_init_io(&s->iomem_cop, OBJECT(s), &bonito_cop_ops, s,
>> "cop", 0x100);
>> - sysbus_init_mmio(sysbus, &s->iomem_cop);
>> - sysbus_mmio_map(sysbus, 4, 0x1fe00300);
>> + memory_region_add_subregion(get_system_memory(), 0x1fe00300,
>> + &s->iomem_cop);
>> create_unimplemented_device("ROMCS", BONITO_FLASH_BASE, 60 * MiB);
>> /* Map PCI IO Space 0x1fd0 0000 - 0x1fd1 0000 */
>> memory_region_init_alias(&s->bonito_pciio, OBJECT(s), "isa_mmio",
>> get_system_io(), 0, BONITO_PCIIO_SIZE);
>> - sysbus_init_mmio(sysbus, &s->bonito_pciio);
>> - sysbus_mmio_map(sysbus, 5, BONITO_PCIIO_BASE);
>> + memory_region_add_subregion(get_system_memory(), BONITO_PCIIO_BASE,
>> + &s->bonito_pciio);
>> /* add pci local io mapping */
>> memory_region_init_alias(&s->bonito_localio, OBJECT(s), "IOCS[0]",
>> get_system_io(), 0, 256 * KiB);
>> - sysbus_init_mmio(sysbus, &s->bonito_localio);
>> - sysbus_mmio_map(sysbus, 6, BONITO_DEV_BASE);
>> + memory_region_add_subregion(get_system_memory(), BONITO_DEV_BASE,
>> + &s->bonito_localio);
>> create_unimplemented_device("IOCS[1]", BONITO_DEV_BASE + 1 * 256
>> * KiB,
>> 256 * KiB);
>> create_unimplemented_device("IOCS[2]", BONITO_DEV_BASE + 2 * 256
>> * KiB,
>
> Would it make sense to cache the return value of get_system_memory() in
> a local variable instead of calling it again and again ...?
This is done later, at this point I'm just doing mechanical
transformation with:
@@
expression sbdev;
expression index;
expression addr;
expression subregion;
@@
- sysbus_init_mmio(sbdev, subregion);
... when != sbdev
- sysbus_mmio_map(sbdev, index, addr);
+ memory_region_add_subregion(get_system_memory(), addr, subregion);
@@
expression sbdev;
expression index;
expression addr;
expression subregion;
expression priority;
@@
- sysbus_init_mmio(sbdev, subregion);
... when != sbdev
- sysbus_mmio_map_overlap(sbdev, index, addr, priority);
+ memory_region_add_subregion_overlap(get_system_memory(), addr,
+ subregion, priority);
next prev parent reply other threads:[~2023-10-18 18:24 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-18 14:11 [PATCH 00/12] hw: Strengthen SysBus & QBus API Philippe Mathieu-Daudé
2023-10-18 14:11 ` [PATCH 01/12] hw/i386/amd_iommu: Do not use SysBus API to map local MMIO region Philippe Mathieu-Daudé
2023-10-18 14:11 ` [PATCH 02/12] hw/i386/intel_iommu: " Philippe Mathieu-Daudé
2023-10-18 14:11 ` [PATCH 03/12] hw/misc/allwinner-dramc: Move sysbus_mmio_map call from init -> realize Philippe Mathieu-Daudé
2023-10-18 14:11 ` [PATCH 04/12] hw/misc/allwinner-dramc: Do not use SysBus API to map local MMIO region Philippe Mathieu-Daudé
2023-10-18 14:11 ` [PATCH 05/12] hw/pci-host/bonito: " Philippe Mathieu-Daudé
2023-10-18 16:05 ` Thomas Huth
2023-10-18 18:23 ` Philippe Mathieu-Daudé [this message]
2023-10-18 14:11 ` [PATCH 06/12] hw/acpi: Realize ACPI_GED sysbus device before accessing it Philippe Mathieu-Daudé
2023-10-18 16:00 ` Thomas Huth
2023-10-18 14:11 ` [PATCH 07/12] hw/arm/virt: Realize ARM_GICV2M " Philippe Mathieu-Daudé
2023-10-18 15:58 ` Thomas Huth
2023-10-18 14:11 ` [PATCH 08/12] hw/isa: Realize ISA BUS " Philippe Mathieu-Daudé
2023-10-18 15:57 ` Thomas Huth
2023-10-18 18:26 ` Philippe Mathieu-Daudé
2023-10-18 14:11 ` [PATCH 09/12] hw/s390x/css-bridge: Realize " Philippe Mathieu-Daudé
2023-10-18 15:06 ` Thomas Huth
2023-10-18 18:27 ` Philippe Mathieu-Daudé
2023-10-18 14:11 ` [PATCH 10/12] hw/qdev: Ensure parent device is not realized before adding bus Philippe Mathieu-Daudé
2023-10-18 16:08 ` Thomas Huth
2023-10-18 14:11 ` [PATCH 11/12] hw/sysbus: Ensure device is not realized before adding MMIO region Philippe Mathieu-Daudé
2023-10-18 16:10 ` Thomas Huth
2023-10-18 14:11 ` [PATCH 12/12] hw/sysbus: Ensure device is realized before mapping it Philippe Mathieu-Daudé
2023-10-18 16:13 ` Thomas Huth
2023-10-18 18:24 ` Philippe Mathieu-Daudé
2023-10-18 16:11 ` [PATCH 00/12] hw: Strengthen SysBus & QBus API Michael S. Tsirkin
2023-10-18 16:24 ` Thomas Huth
2023-10-18 18:25 ` Philippe Mathieu-Daudé
2023-10-18 18:32 ` Philippe Mathieu-Daudé
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=227c8351-75d7-db1d-aeab-1e6fb2db22c0@linaro.org \
--to=philmd@linaro.org \
--cc=b.galvani@gmail.com \
--cc=berrange@redhat.com \
--cc=borntraeger@linux.ibm.com \
--cc=chenhuacai@kernel.org \
--cc=david@redhat.com \
--cc=eduardo@habkost.net \
--cc=farman@linux.ibm.com \
--cc=gaosong@loongson.cn \
--cc=iii@linux.ibm.com \
--cc=jasowang@redhat.com \
--cc=jiaxun.yang@flygoat.com \
--cc=marcel.apfelbaum@gmail.com \
--cc=mst@redhat.com \
--cc=pasic@linux.ibm.com \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=peterx@redhat.com \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=qemu-s390x@nongnu.org \
--cc=richard.henderson@linaro.org \
--cc=slp@redhat.com \
--cc=strahinja.p.jankovic@gmail.com \
--cc=thuth@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).