From: Gustavo Romero <gustavo.romero@linaro.org>
To: eric.auger@redhat.com, qemu-devel@nongnu.org, philmd@linaro.org,
mst@redhat.com
Cc: qemu-arm@nongnu.org, alex.bennee@linaro.org, udo@hypervisor.org,
ajones@ventanamicro.com, peter.maydell@linaro.org,
imammedo@redhat.com, anisinha@redhat.com
Subject: Re: [PATCH v5 8/9] hw/arm/virt-acpi-build: Fix ACPI IORT and MADT tables when its=off
Date: Sat, 28 Jun 2025 16:44:13 -0300 [thread overview]
Message-ID: <85259079-e40c-4550-b251-c35110557d3f@linaro.org> (raw)
In-Reply-To: <5e0e0aa6-cc09-4aa3-b48c-65de8273549c@redhat.com>
Hi Eric,
On 6/27/25 12:44, Eric Auger wrote:
> Hi Gustavo,
>
> On 6/23/25 3:57 PM, Gustavo Romero wrote:
>> Currently, the ITS Group nodes in the IORT table and the GIC ITS Struct
>> in the MADT table are always generated, even if GIC ITS is not available
>> on the machine.
>>
>> This commit fixes it by not generating the ITS Group nodes, not mapping
>> any other node to them, and not advertising the GIC ITS in the MADT
>> table, when GIC ITS is not available on the machine.
>>
>> Since the fix changes the MADT and IORT tables, add the blobs for the
>> "its=off" test to the allow list and update them in the next commit.
>>
>> Reported-by: Udo Steinberg <udo@hypervisor.org>
>> Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2886
>> Signed-off-by: Gustavo Romero <gustavo.romero@linaro.org>
>> Co-authored-by: Philippe Mathieu-Daudé <philmd@linaro.org>
>> ---
>> hw/arm/virt-acpi-build.c | 128 ++++++++++++--------
>> tests/qtest/bios-tables-test-allowed-diff.h | 2 +
>> 2 files changed, 80 insertions(+), 50 deletions(-)
>>
>> diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c
>> index 40a782a498..17ae46804a 100644
>> --- a/hw/arm/virt-acpi-build.c
>> +++ b/hw/arm/virt-acpi-build.c
>> @@ -328,17 +328,27 @@ build_iort(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
>> /* Sort the smmu idmap by input_base */
>> g_array_sort(smmu_idmaps, iort_idmap_compare);
>>
>> - /*
>> - * Knowing the ID ranges from the RC to the SMMU, it's possible to
>> - * determine the ID ranges from RC that go directly to ITS.
>> - */
>> - create_its_idmaps(its_idmaps, smmu_idmaps);
>> -
>> - nb_nodes = 3; /* RC, ITS, SMMUv3 */
>> - rc_mapping_count = smmu_idmaps->len + its_idmaps->len;
>> + nb_nodes = 2; /* RC and SMMUv3 */
>> + rc_mapping_count = smmu_idmaps->len;
>> +
>> + if (vms->its) {
>> + /*
>> + * Knowing the ID ranges from the RC to the SMMU, it's possible to
>> + * determine the ID ranges from RC that go directly to ITS.
>> + */
>> + create_its_idmaps(its_idmaps, smmu_idmaps);
>> +
>> + nb_nodes++; /* ITS */
>> + rc_mapping_count += its_idmaps->len;
>> + }
>> } else {
>> - nb_nodes = 2; /* RC, ITS */
>> - rc_mapping_count = 1;
>> + if (vms->its) {
>> + nb_nodes = 2; /* RC and ITS */
>> + rc_mapping_count = 1; /* Direct map to ITS */
>> + } else {
>> + nb_nodes = 1; /* RC only */
>> + rc_mapping_count = 0; /* No output mapping */
>> + }
>> }
>> /* Number of IORT Nodes */
>> build_append_int_noprefix(table_data, nb_nodes, 4);
>> @@ -347,31 +357,43 @@ build_iort(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
>> build_append_int_noprefix(table_data, IORT_NODE_OFFSET, 4);
>> build_append_int_noprefix(table_data, 0, 4); /* Reserved */
>>
>> - /* Table 12 ITS Group Format */
>> - build_append_int_noprefix(table_data, 0 /* ITS Group */, 1); /* Type */
>> - node_size = 20 /* fixed header size */ + 4 /* 1 GIC ITS Identifier */;
>> - build_append_int_noprefix(table_data, node_size, 2); /* Length */
>> - build_append_int_noprefix(table_data, 1, 1); /* Revision */
>> - build_append_int_noprefix(table_data, id++, 4); /* Identifier */
>> - build_append_int_noprefix(table_data, 0, 4); /* Number of ID mappings */
>> - build_append_int_noprefix(table_data, 0, 4); /* Reference to ID Array */
>> - build_append_int_noprefix(table_data, 1, 4); /* Number of ITSs */
>> - /* GIC ITS Identifier Array */
>> - build_append_int_noprefix(table_data, 0 /* MADT translation_id */, 4);
>> + if (vms->its) {
>> + /* Table 12 ITS Group Format */
>> + build_append_int_noprefix(table_data, 0 /* ITS Group */, 1); /* Type */
>> + node_size = 20 /* fixed header size */ + 4 /* 1 GIC ITS Identifier */;
>> + build_append_int_noprefix(table_data, node_size, 2); /* Length */
>> + build_append_int_noprefix(table_data, 1, 1); /* Revision */
>> + build_append_int_noprefix(table_data, id++, 4); /* Identifier */
>> + build_append_int_noprefix(table_data, 0, 4); /* Number of ID mappings */
>> + build_append_int_noprefix(table_data, 0, 4); /* Reference to ID Array */
>> + build_append_int_noprefix(table_data, 1, 4); /* Number of ITSs */
>> + /* GIC ITS Identifier Array */
>> + build_append_int_noprefix(table_data, 0 /* MADT translation_id */, 4);
>> + }
>>
>> if (vms->iommu == VIRT_IOMMU_SMMUV3) {
>> int irq = vms->irqmap[VIRT_SMMU] + ARM_SPI_BASE;
>> -
>> + int num_id_mappings, offset_to_id_array;
>> +
>> + if (vms->its) {
>> + num_id_mappings = 1; /* ITS Group node */
> While at it, I would suggest to rename num_id_mappings into
> smmu_mapping_count for consistency with
>
> rc_mapping_count.
>
> Also I would recommend to rename smmu_idmaps and its_idmaps into rc_smmu_idmaps and rc_its_idmaps to make things clearer. But that be done outside of thise series if you prefer.
Thanks, that's indeed a nice change for clarification. I've done it in v6.
> Otherwise this looks good.
I was not sure if I had to add your R-b in v6 after I addressed your comments so
I left v6 without a R-b.
Cheers,
Gustavo
> Thanks
>
> Eric
>
>> + offset_to_id_array = SMMU_V3_ENTRY_SIZE; /* Just after the header */
>> + } else {
>> + num_id_mappings = 0; /* No ID mappings */
>> + offset_to_id_array = 0; /* No ID mappings array */
>> + }
>> smmu_offset = table_data->len - table.table_offset;
>> /* Table 9 SMMUv3 Format */
>> build_append_int_noprefix(table_data, 4 /* SMMUv3 */, 1); /* Type */
>> - node_size = SMMU_V3_ENTRY_SIZE + ID_MAPPING_ENTRY_SIZE;
>> + node_size = SMMU_V3_ENTRY_SIZE +
>> + (ID_MAPPING_ENTRY_SIZE * num_id_mappings);
>> build_append_int_noprefix(table_data, node_size, 2); /* Length */
>> build_append_int_noprefix(table_data, 4, 1); /* Revision */
>> build_append_int_noprefix(table_data, id++, 4); /* Identifier */
>> - build_append_int_noprefix(table_data, 1, 4); /* Number of ID mappings */
>> + /* Number of ID mappings */
>> + build_append_int_noprefix(table_data, num_id_mappings, 4);
>> /* Reference to ID Array */
>> - build_append_int_noprefix(table_data, SMMU_V3_ENTRY_SIZE, 4);
>> + build_append_int_noprefix(table_data, offset_to_id_array, 4);
>> /* Base address */
>> build_append_int_noprefix(table_data, vms->memmap[VIRT_SMMU].base, 8);
>> /* Flags */
>> @@ -387,9 +409,11 @@ build_iort(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
>> build_append_int_noprefix(table_data, 0, 4); /* Proximity domain */
>> /* DeviceID mapping index (ignored since interrupts are GSIV based) */
>> build_append_int_noprefix(table_data, 0, 4);
>> -
>> - /* Output IORT node is the ITS Group node (the first node) */
>> - build_iort_id_mapping(table_data, 0, 0x10000, IORT_NODE_OFFSET);
>> + /* Array of ID mappings */
>> + if (num_id_mappings) {
>> + /* Output IORT node is the ITS Group node (the first node). */
>> + build_iort_id_mapping(table_data, 0, 0x10000, IORT_NODE_OFFSET);
>> + }
>> }
>>
>> /* Table 17 Root Complex Node */
>> @@ -430,7 +454,7 @@ build_iort(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
>> *
>> * N.B.: The mapping from SMMUv3 to ITS Group node (SMMUv3 -> ITS) is
>> * defined in the SMMUv3 table, where all SMMUv3 IDs are mapped to the
>> - * ITS Group node.
>> + * ITS Group node, if ITS is available.
>> */
>> for (i = 0; i < smmu_idmaps->len; i++) {
>> range = &g_array_index(smmu_idmaps, AcpiIortIdMapping, i);
>> @@ -439,15 +463,17 @@ build_iort(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
>> range->id_count, smmu_offset);
>> }
>>
>> - /*
>> - * Map bypassed (don't go throught the SMMU) RIDs (input) to ITS Group
>> - * node directly: RC -> ITS.
>> - */
>> - for (i = 0; i < its_idmaps->len; i++) {
>> - range = &g_array_index(its_idmaps, AcpiIortIdMapping, i);
>> - /* Output IORT node is the ITS Group node (the first node). */
>> - build_iort_id_mapping(table_data, range->input_base,
>> - range->id_count, IORT_NODE_OFFSET);
>> + if (vms->its) {
>> + /*
>> + * Map bypassed (don't go throught the SMMU) RIDs (input) to ITS Group
>> + * node directly: RC -> ITS.
>> + */
>> + for (i = 0; i < its_idmaps->len; i++) {
>> + range = &g_array_index(its_idmaps, AcpiIortIdMapping, i);
>> + /* Output IORT node is the ITS Group node (the first node). */
>> + build_iort_id_mapping(table_data, range->input_base,
>> + range->id_count, IORT_NODE_OFFSET);
>> + }
>> }
>> } else {
>> /*
>> @@ -768,18 +794,20 @@ build_madt(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
>> memmap[VIRT_HIGH_GIC_REDIST2].size);
>> }
>>
>> - /*
>> - * ACPI spec, Revision 6.0 Errata A
>> - * (original 6.0 definition has invalid Length)
>> - * 5.2.12.18 GIC ITS Structure
>> - */
>> - build_append_int_noprefix(table_data, 0xF, 1); /* Type */
>> - build_append_int_noprefix(table_data, 20, 1); /* Length */
>> - build_append_int_noprefix(table_data, 0, 2); /* Reserved */
>> - build_append_int_noprefix(table_data, 0, 4); /* GIC ITS ID */
>> - /* Physical Base Address */
>> - build_append_int_noprefix(table_data, memmap[VIRT_GIC_ITS].base, 8);
>> - build_append_int_noprefix(table_data, 0, 4); /* Reserved */
>> + if (vms->its) {
>> + /*
>> + * ACPI spec, Revision 6.0 Errata A
>> + * (original 6.0 definition has invalid Length)
>> + * 5.2.12.18 GIC ITS Structure
>> + */
>> + build_append_int_noprefix(table_data, 0xF, 1); /* Type */
>> + build_append_int_noprefix(table_data, 20, 1); /* Length */
>> + build_append_int_noprefix(table_data, 0, 2); /* Reserved */
>> + build_append_int_noprefix(table_data, 0, 4); /* GIC ITS ID */
>> + /* Physical Base Address */
>> + build_append_int_noprefix(table_data, memmap[VIRT_GIC_ITS].base, 8);
>> + build_append_int_noprefix(table_data, 0, 4); /* Reserved */
>> + }
>> } else {
>> const uint16_t spi_base = vms->irqmap[VIRT_GIC_V2M] + ARM_SPI_BASE;
>>
>> diff --git a/tests/qtest/bios-tables-test-allowed-diff.h b/tests/qtest/bios-tables-test-allowed-diff.h
>> index dfb8523c8b..a88198d5c2 100644
>> --- a/tests/qtest/bios-tables-test-allowed-diff.h
>> +++ b/tests/qtest/bios-tables-test-allowed-diff.h
>> @@ -1 +1,3 @@
>> /* List of comma-separated changed AML files to ignore */
>> +"tests/data/acpi/aarch64/virt/APIC.its_off",
>> +"tests/data/acpi/aarch64/virt/IORT.its_off",
>
next prev parent reply other threads:[~2025-06-28 19:44 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-06-23 13:57 [PATCH-for-10.1 v5 0/9] hw/arm: GIC 'its=off' ACPI table fixes Gustavo Romero
2025-06-23 13:57 ` [PATCH v5 1/9] hw/intc/gicv3_its: Do not check its_class_name() Gustavo Romero
2025-06-23 13:57 ` [PATCH v5 2/9] hw/arm/virt: Simplify logic for setting instance's 'tcg_its' variable Gustavo Romero
2025-06-23 13:57 ` [PATCH v5 3/9] hw/arm/virt: Simplify create_its() Gustavo Romero
2025-06-23 13:57 ` [PATCH v5 4/9] hw/arm/virt-acpi-build: Improve comment in build_iort Gustavo Romero
2025-06-27 13:57 ` Eric Auger
2025-06-27 14:05 ` Eric Auger via
2025-06-23 13:57 ` [PATCH v5 5/9] hw/arm/virt-acpi-build: Factor out create_its_idmaps Gustavo Romero
2025-06-27 15:28 ` Eric Auger
2025-06-28 19:43 ` Gustavo Romero
2025-06-23 13:57 ` [PATCH v5 6/9] qtest/bios-tables-test: Add test for when ITS is off on aarch64 Gustavo Romero
2025-06-27 15:30 ` Eric Auger
2025-06-23 13:57 ` [PATCH v5 7/9] qtest/bios-tables-test: Add blobs for its=off test " Gustavo Romero
2025-06-27 15:32 ` Eric Auger
2025-06-23 13:57 ` [PATCH v5 8/9] hw/arm/virt-acpi-build: Fix ACPI IORT and MADT tables when its=off Gustavo Romero
2025-06-27 15:44 ` Eric Auger
2025-06-28 19:44 ` Gustavo Romero [this message]
2025-06-27 15:45 ` Eric Auger
2025-06-27 15:46 ` Eric Auger via
2025-06-23 13:57 ` [PATCH v5 9/9] qtest/bios-tables-test: Update blobs for its=off test on aarch64 Gustavo Romero
2025-06-27 15:49 ` 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=85259079-e40c-4550-b251-c35110557d3f@linaro.org \
--to=gustavo.romero@linaro.org \
--cc=ajones@ventanamicro.com \
--cc=alex.bennee@linaro.org \
--cc=anisinha@redhat.com \
--cc=eric.auger@redhat.com \
--cc=imammedo@redhat.com \
--cc=mst@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=philmd@linaro.org \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=udo@hypervisor.org \
/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).