Re: [PATCH v7 0/9] ACPI/IORT: Support for IORT RMR node

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2022-01-25 13:00, Shameerali Kolothum Thodi wrote:
Hi Robin/Lorenzo,

-----Original Message-----
From: iommu [mailto:iommu-bounces@xxxxxxxxxxxxxxxxxxxxxxxxxx] On Behalf
Of Shameer Kolothum
Sent: 05 August 2021 09:07
To: linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; linux-acpi@xxxxxxxxxxxxxxx;
iommu@xxxxxxxxxxxxxxxxxxxxxxxxxx
Cc: robin.murphy@xxxxxxx; jon@xxxxxxxxxxxxx; Linuxarm
<linuxarm@xxxxxxxxxx>; steven.price@xxxxxxx; Guohanjun (Hanjun Guo)
<guohanjun@xxxxxxxxxx>; yangyicong <yangyicong@xxxxxxxxxx>;
Sami.Mujawar@xxxxxxx; will@xxxxxxxxxx; wanghuiqiang
<wanghuiqiang@xxxxxxxxxx>
Subject: [PATCH v7 0/9] ACPI/IORT: Support for IORT RMR node

Hi,

The series adds support to IORT RMR nodes specified in IORT
Revision E.b -ARM DEN 0049E[0]. RMR nodes are used to describe
memory ranges that are used by endpoints and require a unity
mapping in SMMU.

We have faced issues with 3408iMR RAID controller cards which
fail to boot when SMMU is enabled. This is because these
controllers make use of host memory for various caching related
purposes and when SMMU is enabled the iMR firmware fails to
access these memory regions as there is no mapping for them.
IORT RMR provides a way for UEFI to describe and report these
memory regions so that the kernel can make a unity mapping for
these in SMMU.

Change History:

v6 --> v7

The only change from v6 is the fix pointed out by Steve to
the SMMUv2 SMR bypass install in patch #8.

Thanks to the Tested-by tags by Laurentiu with SMMUv2 and
Hanjun/Huiqiang with SMMUv3 for v6. I haven't added the tags
yet as the series still needs more review[1].

Feedback and tests on this series is very much appreciated.

Since we have an update to IORT spec(E.c) now[1] and includes additional
attributes/flags for the RMR node, I am planning to respin this series soon.

Going through the new spec, I have a few queries,

The memory range attributes can now be described as one of the following,

0x00: Device-nGnRnE memory
0x01: Device-nGnRE memory
0x02: Device-nGRE memory
0x03: Device-GRE memory
0x04: Normal Inner Non-cacheable Outer Non-cacheable
0x05: Normal Inner Write-back Outer Write-back Inner Shareable

I am not sure how this needs to be captured and used in the kernel. Is there
any intention of using these fine-grained attributes in the kernel now
or a generic mapping of the above to the struct iommu_rev_region prot field
is enough? i.e., something like,

{
     ....
     prot = IOMMU_READ | IOMMU_WRITE;

     if (rmr_attr == normal_mem) // 0x05
         prot |= IOMMU_CACHE;
if (rmr_attr == device_mem) { //0x00 - 0x03
         prot |= IOMMU_MMIO;
         prot |= IOMMU_NOEXEC;
     }
     ....
}

Yup, pretty much that, except don't bother with IOMMU_NOEXEC. We can't reliably infer it - e.g. on an AXI-based interconnect AxCACHE and AxPROT are entirely orthogonal, so a Device-type read with the "Instruction access" hint is perfectly legal - and in the common IORT code we're not in a position to second-guess what any given RMR might represent for whatever agent is accessing it.

All we can reasonably do here is map the Device types to IOMMU_MMIO and Write-back to IOMMU_CACHE, and if anyone ever does want to insist that that's not sufficient, then they're welcome to send patches to make the IOMMU API more expressive :)

Similarly for the 'flags' field, the new 'Access Privilege' is intended to set the
IOMMU_PRIV ?

Yes, exactly!

Cheers,
Robin.

Please let me know.

Thanks,
Shameer

[1] https://developer.arm.com/documentation/den0049/ec/?lang=en




[Index of Archives]     [Linux IBM ACPI]     [Linux Power Management]     [Linux Kernel]     [Linux Laptop]     [Kernel Newbies]     [Share Photos]     [Security]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Samba]     [Video 4 Linux]     [Device Mapper]     [Linux Resources]

  Powered by Linux