From: "Michael S. Tsirkin" <mst@redhat.com>
To: Bernhard Beschow <shentey@gmail.com>
Cc: qemu-devel@nongnu.org,
Richard Henderson <richard.henderson@linaro.org>,
Marcel Apfelbaum <marcel.apfelbaum@gmail.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Eduardo Habkost <eduardo@habkost.net>
Subject: Re: [PATCH 06/12] hw/pci-host/q35: Initialize properties just once
Date: Wed, 1 Mar 2023 16:47:20 -0500 [thread overview]
Message-ID: <20230301164701-mutt-send-email-mst@kernel.org> (raw)
In-Reply-To: <20230214131441.101760-7-shentey@gmail.com>
On Tue, Feb 14, 2023 at 02:14:35PM +0100, Bernhard Beschow wrote:
> Although not used there, the attributes for Q35's "pci-hole64-size" and
> "short_root_bus" properties currently reside in its child device. This
> causes the default values to be overwritten during the child's
> object_initialize() phase.
pls add explanation why this is a problem.
> Avoid this by moving both attributes into the
> host device.
>
> Signed-off-by: Bernhard Beschow <shentey@gmail.com>
> ---
> include/hw/pci-host/q35.h | 5 +++--
> hw/pci-host/q35.c | 20 +++++---------------
> 2 files changed, 8 insertions(+), 17 deletions(-)
>
> diff --git a/include/hw/pci-host/q35.h b/include/hw/pci-host/q35.h
> index fcbe57b42d..93e41ffbee 100644
> --- a/include/hw/pci-host/q35.h
> +++ b/include/hw/pci-host/q35.h
> @@ -54,8 +54,6 @@ struct MCHPCIState {
> Range pci_hole;
> uint64_t below_4g_mem_size;
> uint64_t above_4g_mem_size;
> - uint64_t pci_hole64_size;
> - uint32_t short_root_bus;
> uint16_t ext_tseg_mbytes;
> };
>
> @@ -64,7 +62,10 @@ struct Q35PCIHost {
> PCIExpressHost parent_obj;
> /*< public >*/
>
> + uint64_t pci_hole64_size;
> + uint32_t short_root_bus;
> bool pci_hole64_fix;
> +
> MCHPCIState mch;
> };
>
> diff --git a/hw/pci-host/q35.c b/hw/pci-host/q35.c
> index 0e198f97a7..03aa08dae5 100644
> --- a/hw/pci-host/q35.c
> +++ b/hw/pci-host/q35.c
> @@ -76,7 +76,7 @@ static const char *q35_host_root_bus_path(PCIHostState *host_bridge,
> Q35PCIHost *s = Q35_HOST_DEVICE(host_bridge);
>
> /* For backwards compat with old device paths */
> - if (s->mch.short_root_bus) {
> + if (s->short_root_bus) {
> return "0000";
> }
> return "0000:00";
> @@ -161,27 +161,19 @@ static void q35_host_get_pci_hole64_end(Object *obj, Visitor *v,
>
> pci_bus_get_w64_range(h->bus, &w64);
> value = range_is_empty(&w64) ? 0 : range_upb(&w64) + 1;
> - hole64_end = ROUND_UP(hole64_start + s->mch.pci_hole64_size, 1ULL << 30);
> + hole64_end = ROUND_UP(hole64_start + s->pci_hole64_size, 1ULL << 30);
> if (s->pci_hole64_fix && value < hole64_end) {
> value = hole64_end;
> }
> visit_type_uint64(v, name, &value, errp);
> }
>
> -/*
> - * NOTE: setting defaults for the mch.* fields in this table
> - * doesn't work, because mch is a separate QOM object that is
> - * zeroed by the object_initialize(&s->mch, ...) call inside
> - * q35_host_initfn(). The default values for those
> - * properties need to be initialized manually by
> - * q35_host_initfn() after the object_initialize() call.
> - */
> static Property q35_host_props[] = {
> DEFINE_PROP_UINT64(PCIE_HOST_MCFG_BASE, Q35PCIHost, parent_obj.base_addr,
> MCH_HOST_BRIDGE_PCIEXBAR_DEFAULT),
> DEFINE_PROP_SIZE(PCI_HOST_PROP_PCI_HOLE64_SIZE, Q35PCIHost,
> - mch.pci_hole64_size, Q35_PCI_HOST_HOLE64_SIZE_DEFAULT),
> - DEFINE_PROP_UINT32("short_root_bus", Q35PCIHost, mch.short_root_bus, 0),
> + pci_hole64_size, Q35_PCI_HOST_HOLE64_SIZE_DEFAULT),
> + DEFINE_PROP_UINT32("short_root_bus", Q35PCIHost, short_root_bus, 0),
> DEFINE_PROP_SIZE(PCI_HOST_BELOW_4G_MEM_SIZE, Q35PCIHost,
> mch.below_4g_mem_size, 0),
> DEFINE_PROP_SIZE(PCI_HOST_ABOVE_4G_MEM_SIZE, Q35PCIHost,
> @@ -218,9 +210,7 @@ static void q35_host_initfn(Object *obj)
> object_initialize_child(OBJECT(s), "mch", &s->mch, TYPE_MCH_PCI_DEVICE);
> qdev_prop_set_int32(DEVICE(&s->mch), "addr", PCI_DEVFN(0, 0));
> qdev_prop_set_bit(DEVICE(&s->mch), "multifunction", false);
> - /* mch's object_initialize resets the default value, set it again */
> - qdev_prop_set_uint64(DEVICE(s), PCI_HOST_PROP_PCI_HOLE64_SIZE,
> - Q35_PCI_HOST_HOLE64_SIZE_DEFAULT);
> +
> object_property_add(obj, PCI_HOST_PROP_PCI_HOLE_START, "uint32",
> q35_host_get_pci_hole_start,
> NULL, NULL, NULL);
> --
> 2.39.1
next prev parent reply other threads:[~2023-03-01 21:47 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-14 13:14 [PATCH 00/12] Q35 PCI host fixes and QOM cleanup Bernhard Beschow
2023-02-14 13:14 ` [PATCH 01/12] hw/i386/pc_q35: Resolve redundant q35_host variable Bernhard Beschow
2023-02-14 13:14 ` [PATCH 02/12] hw/pci-host/q35: Fix contradicting .endianness assignment Bernhard Beschow
2023-03-01 21:43 ` Michael S. Tsirkin
2023-03-01 21:44 ` Michael S. Tsirkin
2023-02-14 13:14 ` [PATCH 03/12] hw/pci-host/q35: Use memory_region_set_address() also for tseg_blackhole Bernhard Beschow
2023-03-01 21:45 ` Michael S. Tsirkin
2023-02-14 13:14 ` [PATCH 04/12] hw/pci-host/q35: Initialize PCMachineState::bus in board code Bernhard Beschow
2023-02-14 13:14 ` [PATCH 05/12] hw/pci-host/q35: Initialize "bypass-iommu" property from " Bernhard Beschow
2023-03-01 21:46 ` Michael S. Tsirkin
2023-02-14 13:14 ` [PATCH 06/12] hw/pci-host/q35: Initialize properties just once Bernhard Beschow
2023-03-01 21:47 ` Michael S. Tsirkin [this message]
2023-02-14 13:14 ` [PATCH 07/12] hw/pci-host/q35: Initialize PCI hole boundaries " Bernhard Beschow
2023-02-14 13:14 ` [PATCH 08/12] hw/pci-host/q35: Turn PCI hole properties into class properties Bernhard Beschow
2023-02-14 13:14 ` [PATCH 09/12] hw/pci-host/q35: Rename local variable to more idiomatic "phb" Bernhard Beschow
2023-02-14 13:14 ` [PATCH 10/12] hw/pci-host/q35: Propagate to errp rather than doing error_fatal Bernhard Beschow
2023-02-14 13:14 ` [PATCH 11/12] hw/pci-host/q35: Merge mch_realize() into q35_host_realize() Bernhard Beschow
2023-02-14 13:14 ` [PATCH 12/12] hw/pci-host/q35: Move MemoryRegion pointers to host device Bernhard Beschow
2023-02-21 15:39 ` [PATCH 00/12] Q35 PCI host fixes and QOM cleanup Bernhard Beschow
2023-03-01 21:49 ` Michael S. Tsirkin
2023-03-02 21:54 ` Bernhard Beschow
2023-03-05 7:41 ` Bernhard Beschow
2023-03-01 21:50 ` Michael S. Tsirkin
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=20230301164701-mutt-send-email-mst@kernel.org \
--to=mst@redhat.com \
--cc=eduardo@habkost.net \
--cc=marcel.apfelbaum@gmail.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=richard.henderson@linaro.org \
--cc=shentey@gmail.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).