qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Eric Auger <eric.auger@redhat.com>
To: "Cédric Le Goater" <clg@redhat.com>,
	eric.auger.pro@gmail.com, qemu-devel@nongnu.org,
	qemu-arm@nongnu.org, jean-philippe@linaro.org, mst@redhat.com,
	peter.maydell@linaro.org, zhenzhong.duan@intel.com,
	yanghliu@redhat.com
Cc: alex.williamson@redhat.com, jasowang@redhat.com
Subject: Re: [PATCH v5 3/4] hw: Set virtio-iommu aw-bits default value on pc_q35 and arm virt
Date: Tue, 27 Feb 2024 18:39:15 +0100	[thread overview]
Message-ID: <7797a7a8-1754-40e1-a580-895bbb4546ed@redhat.com> (raw)
In-Reply-To: <2259fba1-4c57-420b-bcab-ecf202d10819@redhat.com>

Hi Cédric,

On 2/27/24 18:17, Cédric Le Goater wrote:
> Hello Eric,
>
> On 2/15/24 09:42, Eric Auger wrote:
>> Currently the default input range can extend to 64 bits. On x86,
>> when the virtio-iommu protects vfio devices, the physical iommu
>> may support only 39 bits. Let's set the default to 39, as done
>> for the intel-iommu. On ARM we set 48b as a default (matching
>> SMMUv3 SMMU_IDR5.VAX == 0).
>>
>> We use hw_compat_8_2 to handle the compatibility for machines
>> before 9.0 which used to have a virtio-iommu default input range
>> of 64 bits.
>>
>> Of course if aw-bits is set from the command line, the default
>> is overriden.
>>
>> Signed-off-by: Eric Auger <eric.auger@redhat.com>
>> Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
>> Tested-by: Yanghang Liu<yanghliu@redhat.com>
>
> We need a property fixup for pseries also:
>
> $ build/ppc64-softmmu/qemu-system-ppc64 -M pseries  -device
> virtio-iommu-pci,addr=04.0
> qemu-system-ppc64: -device virtio-iommu-pci,addr=04.0: aw-bits must be
> within [32,64]

as stated in the doc (now! see [PATCH v5 4/4] qemu-options.hx: Add an
entry for virtio-iommu-pci and document aw-bits) , the virtio-iommu-pci
device only is supported in q35 and arm virt.

If you want to support virtio-iommu on other machines, you need to take
care of this aw-bits settings and also of the exposed reserved regions.

Thanks

Eric
>
>
> Thanks,
>
> C.
>
>
>
>> ---
>>
>> v3 -> v4:
>> - update the qos test to relax the check on the max input IOVA
>>
>> v2 -> v3:
>> - collected Zhenzhong's R-b
>> - use &error_abort instead of NULL error handle
>>    on object_property_get_uint() call (Cédric)
>> - use VTD_HOST_AW_39BIT (Cédric)
>>
>> v1 -> v2:
>> - set aw-bits to 48b on ARM
>> - use hw_compat_8_2 to handle the compat for older machines
>>    which used 64b as a default
>> ---
>>   hw/arm/virt.c                   | 6 ++++++
>>   hw/core/machine.c               | 5 ++++-
>>   hw/i386/pc.c                    | 6 ++++++
>>   hw/virtio/virtio-iommu.c        | 2 +-
>>   tests/qtest/virtio-iommu-test.c | 2 +-
>>   5 files changed, 18 insertions(+), 3 deletions(-)
>>
>> diff --git a/hw/arm/virt.c b/hw/arm/virt.c
>> index 368c2a415a..0994f2a560 100644
>> --- a/hw/arm/virt.c
>> +++ b/hw/arm/virt.c
>> @@ -2716,10 +2716,16 @@ static void
>> virt_machine_device_pre_plug_cb(HotplugHandler *hotplug_dev,
>>       } else if (object_dynamic_cast(OBJECT(dev), TYPE_VIRTIO_MD_PCI)) {
>>           virtio_md_pci_pre_plug(VIRTIO_MD_PCI(dev),
>> MACHINE(hotplug_dev), errp);
>>       } else if (object_dynamic_cast(OBJECT(dev),
>> TYPE_VIRTIO_IOMMU_PCI)) {
>> +        uint8_t aw_bits = object_property_get_uint(OBJECT(dev),
>> +                                                   "aw-bits",
>> &error_abort);
>>           hwaddr db_start = 0, db_end = 0;
>>           QList *reserved_regions;
>>           char *resv_prop_str;
>>   +        if (!aw_bits) {
>> +            qdev_prop_set_uint8(dev, "aw-bits", 48);
>> +        }
>> +
>>           if (vms->iommu != VIRT_IOMMU_NONE) {
>>               error_setg(errp, "virt machine does not support
>> multiple IOMMUs");
>>               return;
>> diff --git a/hw/core/machine.c b/hw/core/machine.c
>> index fb5afdcae4..70ac96954c 100644
>> --- a/hw/core/machine.c
>> +++ b/hw/core/machine.c
>> @@ -30,9 +30,12 @@
>>   #include "exec/confidential-guest-support.h"
>>   #include "hw/virtio/virtio-pci.h"
>>   #include "hw/virtio/virtio-net.h"
>> +#include "hw/virtio/virtio-iommu.h"
>>   #include "audio/audio.h"
>>   -GlobalProperty hw_compat_8_2[] = {};
>> +GlobalProperty hw_compat_8_2[] = {
>> +    { TYPE_VIRTIO_IOMMU_PCI, "aw-bits", "64" },
>> +};
>>   const size_t hw_compat_8_2_len = G_N_ELEMENTS(hw_compat_8_2);
>>     GlobalProperty hw_compat_8_1[] = {
>> diff --git a/hw/i386/pc.c b/hw/i386/pc.c
>> index 196827531a..ee2d379c90 100644
>> --- a/hw/i386/pc.c
>> +++ b/hw/i386/pc.c
>> @@ -1456,6 +1456,8 @@ static void
>> pc_machine_device_pre_plug_cb(HotplugHandler *hotplug_dev,
>>       } else if (object_dynamic_cast(OBJECT(dev), TYPE_VIRTIO_MD_PCI)) {
>>           virtio_md_pci_pre_plug(VIRTIO_MD_PCI(dev),
>> MACHINE(hotplug_dev), errp);
>>       } else if (object_dynamic_cast(OBJECT(dev),
>> TYPE_VIRTIO_IOMMU_PCI)) {
>> +        uint8_t aw_bits = object_property_get_uint(OBJECT(dev),
>> +                                                   "aw-bits",
>> &error_abort);
>>           /* Declare the APIC range as the reserved MSI region */
>>           char *resv_prop_str =
>> g_strdup_printf("0xfee00000:0xfeefffff:%d",
>>                                                
>> VIRTIO_IOMMU_RESV_MEM_T_MSI);
>> @@ -1464,6 +1466,10 @@ static void
>> pc_machine_device_pre_plug_cb(HotplugHandler *hotplug_dev,
>>           qlist_append_str(reserved_regions, resv_prop_str);
>>           qdev_prop_set_array(dev, "reserved-regions",
>> reserved_regions);
>>   +        if (!aw_bits) {
>> +            qdev_prop_set_uint8(dev, "aw-bits", VTD_HOST_AW_39BIT);
>> +        }
>> +
>>           g_free(resv_prop_str);
>>       }
>>   diff --git a/hw/virtio/virtio-iommu.c b/hw/virtio/virtio-iommu.c
>> index 8b541de850..2ec5ef3cd1 100644
>> --- a/hw/virtio/virtio-iommu.c
>> +++ b/hw/virtio/virtio-iommu.c
>> @@ -1526,7 +1526,7 @@ static Property virtio_iommu_properties[] = {
>>       DEFINE_PROP_LINK("primary-bus", VirtIOIOMMU, primary_bus,
>>                        TYPE_PCI_BUS, PCIBus *),
>>       DEFINE_PROP_BOOL("boot-bypass", VirtIOIOMMU, boot_bypass, true),
>> -    DEFINE_PROP_UINT8("aw-bits", VirtIOIOMMU, aw_bits, 64),
>> +    DEFINE_PROP_UINT8("aw-bits", VirtIOIOMMU, aw_bits, 0),
>>       DEFINE_PROP_END_OF_LIST(),
>>   };
>>   diff --git a/tests/qtest/virtio-iommu-test.c
>> b/tests/qtest/virtio-iommu-test.c
>> index 068e7a9e6c..0f36381acb 100644
>> --- a/tests/qtest/virtio-iommu-test.c
>> +++ b/tests/qtest/virtio-iommu-test.c
>> @@ -34,7 +34,7 @@ static void pci_config(void *obj, void *data,
>> QGuestAllocator *t_alloc)
>>       uint8_t bypass = qvirtio_config_readb(dev, 36);
>>         g_assert_cmpint(input_range_start, ==, 0);
>> -    g_assert_cmphex(input_range_end, ==, UINT64_MAX);
>> +    g_assert_cmphex(input_range_end, >=, 32);
>>       g_assert_cmpint(domain_range_start, ==, 0);
>>       g_assert_cmpint(domain_range_end, ==, UINT32_MAX);
>>       g_assert_cmpint(bypass, ==, 1);
>



  reply	other threads:[~2024-02-27 17:39 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-02-15  8:42 [PATCH v5 0/4] VIRTIO-IOMMU: Introduce an aw-bits option Eric Auger
2024-02-15  8:42 ` [PATCH v5 1/4] virtio-iommu: Add an option to define the input range width Eric Auger
2024-02-15  8:42 ` [PATCH v5 2/4] virtio-iommu: Trace domain range limits as unsigned int Eric Auger
2024-02-15  8:42 ` [PATCH v5 3/4] hw: Set virtio-iommu aw-bits default value on pc_q35 and arm virt Eric Auger
2024-02-27 17:17   ` Cédric Le Goater
2024-02-27 17:39     ` Eric Auger [this message]
2024-02-27 18:26     ` Eric Auger
2024-02-29  9:20   ` Igor Mammedov
2024-03-04 17:37     ` Eric Auger
2024-02-15  8:42 ` [PATCH v5 4/4] qemu-options.hx: Add an entry for virtio-iommu-pci and document aw-bits Eric Auger
2024-02-15  9:33   ` Eric Auger

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=7797a7a8-1754-40e1-a580-895bbb4546ed@redhat.com \
    --to=eric.auger@redhat.com \
    --cc=alex.williamson@redhat.com \
    --cc=clg@redhat.com \
    --cc=eric.auger.pro@gmail.com \
    --cc=jasowang@redhat.com \
    --cc=jean-philippe@linaro.org \
    --cc=mst@redhat.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-arm@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=yanghliu@redhat.com \
    --cc=zhenzhong.duan@intel.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).