Hi,
On 09/11/2020 00:04, Jan Kiszka wrote:
> On 04.11.20 16:36, Andrea Bastoni wrote:
>> I.MX8QM fixed mask is factored-in into the stream_ids.
>>
>> Signed-off-by: Andrea Bastoni <[email protected]>
>> ---
>> configs/arm64/imx8qm.c | 6 ++++--
>> hypervisor/arch/arm64/smmu.c | 3 ---
>> include/jailhouse/cell-config.h | 4 ----
>> 3 files changed, 4 insertions(+), 9 deletions(-)
>>
>> diff --git a/configs/arm64/imx8qm.c b/configs/arm64/imx8qm.c
>> index d63c73cf..566fd0ad 100644
>> --- a/configs/arm64/imx8qm.c
>> +++ b/configs/arm64/imx8qm.c
>> @@ -54,7 +54,6 @@ struct {
>> .type = JAILHOUSE_IOMMU_ARM_MMU500,
>> .base = 0x51400000,
>> .size = 0x40000,
>> - .arm_mmu500.sid_mask = 0x7f80,
>> },
>> },
>>
>> @@ -209,6 +208,9 @@ struct {
>> },
>>
>> .stream_ids = {
>> - 0x11, 0x12, 0x13,
>> + /* bits 30 - 16: SMR MASK
>> + * bits 14 - 0 : SMR ID
>> + */
>> + 0x7f80011, 0x7f80012, 0x7f80013,
>> },
>> };
>> diff --git a/hypervisor/arch/arm64/smmu.c b/hypervisor/arch/arm64/smmu.c
>> index 191ff154..d2824985 100644
>> --- a/hypervisor/arch/arm64/smmu.c
>> +++ b/hypervisor/arch/arm64/smmu.c
>> @@ -157,7 +157,6 @@ struct arm_smmu_device {
>> unsigned long pgshift;
>> u32 num_context_banks;
>> u32 num_mapping_groups;
>> - u16 arm_sid_mask;
>> struct arm_smmu_smr *smrs;
>> };
>>
>> @@ -564,8 +563,6 @@ static int arm_smmu_init(void)
>> continue;
>>
>> smmu = &smmu_device[num_smmu_devices];
>> - smmu->arm_sid_mask = iommu->arm_mmu500.sid_mask;
>> -
>> smmu->base = paging_map_device(iommu->base, iommu->size);
>> if (!smmu->base) {
>> err = -ENOMEM;
>> diff --git a/include/jailhouse/cell-config.h
>> b/include/jailhouse/cell-config.h
>> index 472cb4bb..d6315489 100644
>> --- a/include/jailhouse/cell-config.h
>> +++ b/include/jailhouse/cell-config.h
>> @@ -279,10 +279,6 @@ struct jailhouse_iommu {
>> __u64 tlb_base;
>> __u32 tlb_size;
>> } __attribute__((packed)) tipvu;
>> -
>> - struct {
>> - __u32 sid_mask;
>> - } __attribute__((packed)) arm_mmu500;
>> };
>> } __attribute__((packed));
>>
>>
>
> This concept of masking it SMMUv2-specific? Or would it be worth to
> introduce it generically, e.g. by using a 32-bit mask?
To my understanding this is a specialty of the SMMUv2, but I couldn't find many
details on the PVU.
> If not, then we should model a stream ID entry via a union, something
> like this:
>
> union jailhouse_stream_id {
> u32 long_id;
> struct {
> u16 id;
> u16 mask;
> } __attribute__((packed)) mmu500;
> } __attribute__((packed));
This makes the SMMUv2 specification clearer, but it makes somehow longer the
configuration for v3 and PVU. (I assume you want to convert all u32 stream_ids[]
into union jailhouse_stream_id stream_ids[].)
The alternative struct jailhouse_mmu500_stream_id is even worse for maintenance
in the long term.
--
Thanks,
Andrea Bastoni
--
You received this message because you are subscribed to the Google Groups
"Jailhouse" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/jailhouse-dev/96d07bef-e970-c199-50fc-c9541790f1c3%40tum.de.
