Existing exynos-iommu driver only supports SysMMU versions up to v5. But
it's pretty much ready for basic usage with SysMMU v7, only small
changes have to be done. As SysMMU version is tested dynamically (by
reading the corresponding register), there is no need to introduce new
compatible string.
One major change is that SysMMU v7 can have different register layouts:
- with Virtual Machine support
- without Virtual Machine support
That can be checked by reading the capability registers. In case the
SysMMU IP-core is VM-capable, the VM registers have to be used, and some
additional initialization is needed. That is the case on E850-96 board,
which non-secure SysMMU (v7.4) implements VM-capable register set.
The patch series was tested on E850-96 board. Because at the moment
there are no SysMMU users for that board, the testing was done using so
called "Emulated Translation" registers available on SysMMU v7. That
allows one to initiate the translation from CPU, by writing to those
registers, and then reading the corresponding TLB registers to find out
the translation result. The testing driver can be found in [1] tree.
[1] https://github.com/joe-skb7/linux/tree/e850-96-mainline-iommu
Sam Protsenko (4):
iommu/exynos: Set correct dma mask for SysMMU v5+
iommu/exynos: Check if SysMMU v7 has VM registers
iommu/exynos: Use lookup based approach to access v7 registers
iommu/exynos: Add minimal support for SysMMU v7 with VM registers
drivers/iommu/exynos-iommu.c | 112 ++++++++++++++++++++++++++++++++---
1 file changed, 104 insertions(+), 8 deletions(-)
--
2.30.2
On Sun, 3 Jul 2022 at 00:37, Sam Protsenko <[email protected]> wrote:
>
> Existing exynos-iommu driver only supports SysMMU versions up to v5. But
> it's pretty much ready for basic usage with SysMMU v7, only small
> changes have to be done. As SysMMU version is tested dynamically (by
> reading the corresponding register), there is no need to introduce new
> compatible string.
>
> One major change is that SysMMU v7 can have different register layouts:
> - with Virtual Machine support
> - without Virtual Machine support
>
> That can be checked by reading the capability registers. In case the
> SysMMU IP-core is VM-capable, the VM registers have to be used, and some
> additional initialization is needed. That is the case on E850-96 board,
> which non-secure SysMMU (v7.4) implements VM-capable register set.
>
> The patch series was tested on E850-96 board. Because at the moment
> there are no SysMMU users for that board, the testing was done using so
> called "Emulated Translation" registers available on SysMMU v7. That
> allows one to initiate the translation from CPU, by writing to those
> registers, and then reading the corresponding TLB registers to find out
> the translation result. The testing driver can be found in [1] tree.
>
> [1] https://github.com/joe-skb7/linux/tree/e850-96-mainline-iommu
>
> Sam Protsenko (4):
> iommu/exynos: Set correct dma mask for SysMMU v5+
> iommu/exynos: Check if SysMMU v7 has VM registers
> iommu/exynos: Use lookup based approach to access v7 registers
> iommu/exynos: Add minimal support for SysMMU v7 with VM registers
>
> drivers/iommu/exynos-iommu.c | 112 ++++++++++++++++++++++++++++++++---
> 1 file changed, 104 insertions(+), 8 deletions(-)
>
> --
> 2.30.2
>
Hi Marek,
As I understand, you have some board with SysMMU v7, which is not VM
capable (judging from the patches you shared earlier). Could you
please somehow verify if this series works fine for you? For example,
this testing driver [1] can be helpful.
Thanks!
[1] https://github.com/joe-skb7/linux/commit/bbadd46fa525fe1fef2ccbdfff81f7d29caf0506
On Sun, 2022-07-03 at 00:48 +0300, Sam Protsenko wrote:
[...]
> Hi Marek,
>
> As I understand, you have some board with SysMMU v7, which is not VM
> capable (judging from the patches you shared earlier). Could you
> please somehow verify if this series works fine for you? For example,
> this testing driver [1] can be helpful.
>
> Thanks!
>
> [1]
> https://github.com/joe-skb7/linux/commit/bbadd46fa525fe1fef2ccbdfff81f7d29caf0506
Hi Sam,
Not Marek here, but I wanted to try this on my jackpotlte (Exynos
7885). The driver reports it's DPU sysmmu as version 7.2, and manually
reading the capabilities registers it looks like it has the 2nd
capability register but not the VM capability.
After applying your patches, adding your test driver (with SYSMMU_BASE
corrected to 7885 value), and adding the sysmmu to dt, I tried to cat
the test file that it creates in debugfs and I got an SError kernel
panic.
I tried tracing where the SError happens and it looks like it's this
line:
/* Preload for emulation */
iowrite32(rw | vpn, obj->reg_base + MMU_EMU_PRELOAD);
Trying to read the EMU registers using devmem results in a "Bus error".
Could these emulation registers be missing from my SysMMU? Do you have
any info on what version should have it? Or maybe some capability bit?
I'll try testing it with DECON/DPP later and see if it works that way.
Best regards,
David
On Sun, 3 Jul 2022 at 13:47, David Virag <[email protected]> wrote:
>
> On Sun, 2022-07-03 at 00:48 +0300, Sam Protsenko wrote:
> [...]
> > Hi Marek,
> >
> > As I understand, you have some board with SysMMU v7, which is not VM
> > capable (judging from the patches you shared earlier). Could you
> > please somehow verify if this series works fine for you? For example,
> > this testing driver [1] can be helpful.
> >
> > Thanks!
> >
> > [1]
> > https://github.com/joe-skb7/linux/commit/bbadd46fa525fe1fef2ccbdfff81f7d29caf0506
>
> Hi Sam,
>
> Not Marek here, but I wanted to try this on my jackpotlte (Exynos
> 7885). The driver reports it's DPU sysmmu as version 7.2, and manually
> reading the capabilities registers it looks like it has the 2nd
> capability register but not the VM capability.
>
> After applying your patches, adding your test driver (with SYSMMU_BASE
> corrected to 7885 value), and adding the sysmmu to dt, I tried to cat
> the test file that it creates in debugfs and I got an SError kernel
> panic.
>
> I tried tracing where the SError happens and it looks like it's this
> line:
> /* Preload for emulation */
> iowrite32(rw | vpn, obj->reg_base + MMU_EMU_PRELOAD);
>
> Trying to read the EMU registers using devmem results in a "Bus error".
>
> Could these emulation registers be missing from my SysMMU? Do you have
> any info on what version should have it? Or maybe some capability bit?
> I'll try testing it with DECON/DPP later and see if it works that way.
>
Hi Janghyuck,
Do you have by chance any info on SysMMU v7.2, which is present e.g.
on Exynos7885? David is trying to use emulation registers there with
no luck, so it would be nice if you can provide some details on
questions above.
Thanks!
> Best regards,
> David
On Sun, 3 Jul 2022 at 13:47, David Virag <[email protected]> wrote:
>
> On Sun, 2022-07-03 at 00:48 +0300, Sam Protsenko wrote:
> [...]
> > Hi Marek,
> >
> > As I understand, you have some board with SysMMU v7, which is not VM
> > capable (judging from the patches you shared earlier). Could you
> > please somehow verify if this series works fine for you? For example,
> > this testing driver [1] can be helpful.
> >
> > Thanks!
> >
> > [1]
> > https://github.com/joe-skb7/linux/commit/bbadd46fa525fe1fef2ccbdfff81f7d29caf0506
>
> Hi Sam,
>
> Not Marek here, but I wanted to try this on my jackpotlte (Exynos
> 7885). The driver reports it's DPU sysmmu as version 7.2, and manually
> reading the capabilities registers it looks like it has the 2nd
> capability register but not the VM capability.
>
> After applying your patches, adding your test driver (with SYSMMU_BASE
> corrected to 7885 value), and adding the sysmmu to dt, I tried to cat
> the test file that it creates in debugfs and I got an SError kernel
> panic.
>
> I tried tracing where the SError happens and it looks like it's this
> line:
> /* Preload for emulation */
> iowrite32(rw | vpn, obj->reg_base + MMU_EMU_PRELOAD);
>
> Trying to read the EMU registers using devmem results in a "Bus error".
>
> Could these emulation registers be missing from my SysMMU? Do you have
> any info on what version should have it? Or maybe some capability bit?
> I'll try testing it with DECON/DPP later and see if it works that way.
>
I don't have any manuals for v7.2, so I can only assume. Yes, it looks
to me very much like those EMU registers are missing in your SysMMU
IP-core: I remember seeing some similar SError messages while trying
to access some incorrect MMIO addresses. Good news is that once this
patch series is fixed and accepted, you can *probably* base your work
on top of it (as I only validated it with EMU registers for now). I
mean you can add some real IP-core users of that IOMMU, like graphics
(DPU), audio, camera, etc. Not sure though if it would be enough to
just add some DTS nodes, or your SoC support has to be added to some
drivers first.
> Best regards,
> David
On Sun, 3 Jul 2022 at 13:47, David Virag <[email protected]> wrote:
>
> On Sun, 2022-07-03 at 00:48 +0300, Sam Protsenko wrote:
> [...]
> > Hi Marek,
> >
> > As I understand, you have some board with SysMMU v7, which is not VM
> > capable (judging from the patches you shared earlier). Could you
> > please somehow verify if this series works fine for you? For example,
> > this testing driver [1] can be helpful.
> >
> > Thanks!
> >
> > [1]
> > https://github.com/joe-skb7/linux/commit/bbadd46fa525fe1fef2ccbdfff81f7d29caf0506
>
> Hi Sam,
>
> Not Marek here, but I wanted to try this on my jackpotlte (Exynos
> 7885). The driver reports it's DPU sysmmu as version 7.2, and manually
> reading the capabilities registers it looks like it has the 2nd
> capability register but not the VM capability.
>
> After applying your patches, adding your test driver (with SYSMMU_BASE
> corrected to 7885 value), and adding the sysmmu to dt, I tried to cat
> the test file that it creates in debugfs and I got an SError kernel
> panic.
>
> I tried tracing where the SError happens and it looks like it's this
> line:
> /* Preload for emulation */
> iowrite32(rw | vpn, obj->reg_base + MMU_EMU_PRELOAD);
>
> Trying to read the EMU registers using devmem results in a "Bus error".
>
> Could these emulation registers be missing from my SysMMU? Do you have
> any info on what version should have it? Or maybe some capability bit?
> I'll try testing it with DECON/DPP later and see if it works that way.
>
Hey David,
Can you please try out the [1] branch again? I've added support for
Exynos7885 there, using the info provided by Janghyuck Kim: apparently
Exynos7885 has EMU registers, but those have some other offsets than
on Exynos850:
#define MMU_EMU_PRELOAD 0x8040
#define MMU_EMU_SHOT 0x8044
I wonder if my changes to [1] do the trick.
[1] https://github.com/joe-skb7/linux/tree/e850-96-mainline-iommu
> Best regards,
> David