This is a final patchset in the series created in the framework of
my Baikal-T1 PCIe/eDMA-related work:
[1: Done v5] PCI: dwc: Various fixes and cleanups
Link: https://lore.kernel.org/linux-pci/[email protected]/
Merged: kernel 6.0-rc1
[2: Done v4] PCI: dwc: Add hw version and dma-ranges support
Link: https://lore.kernel.org/linux-pci/[email protected]
Merged: kernel 6.0-rc1
[3: In-review v5] PCI: dwc: Add generic resources and Baikal-T1 support
Link: https://lore.kernel.org/linux-pci/[email protected]/
[4: Done v5] dmaengine: dw-edma: Add RP/EP local DMA support
Link: ---you are looking at it---
Note it is very recommended to merge the patchsets in the same order as
they are listed in the set above in order to have them applied smoothly.
Nothing prevents them from being reviewed synchronously though.
Please note originally this series was self content, but due to Frank
being a bit faster in his work submission I had to rebase my patchset onto
his one. So now this patchset turns to be dependent on the Frank' work:
Link: https://lore.kernel.org/linux-pci/[email protected]/
Here is a short summary regarding this patchset. The series starts with
fixes patches. We discovered that the dw-edma-pcie.c driver incorrectly
initializes the LL/DT base addresses for the platforms with not matching
CPU and PCIe memory spaces. It is fixed by using the pci_bus_address()
method to get a correct base address. After that you can find a series of
the interleaved xfers fixes. It turned out the interleaved transfers
implementation didn't work quite correctly from the very beginning for
instance missing src/dst addresses initialization, etc. In the framework
of the next two patches we suggest to add a new platform-specific
callback - pci_address() and use it to convert the CPU address to the PCIe
space address. It is at least required for the DW eDMA remote End-point
setup on the platforms with not-matching CPU/PCIe address spaces. In case
of the DW eDMA local RP/EP setup the conversion will be done automatically
by the outbound iATU (if no DMA-bypass flag is specified for the
corresponding iATU window). Then we introduce a set of the patches to make
the DebugFS part of the code supporting the multi-eDMA controllers
platforms. It starts with several cleanup patches and is closed joining
the Read/Write channels into a single DMA-device as they originally should
have been. After that you can find the patches with adding the non-atomic
io-64 methods usage, dropping DT-region descriptors allocation, replacing
chip IDs with the device name. In addition to that in order to have the
eDMA embedded into the DW PCIe RP/EP supported we need to bypass the
dma-ranges-based memory ranges mapping since in case of the root port DT
node it's applicable for the peripheral PCIe devices only. Finally at the
series closure we introduce a generic DW eDMA controller support being
available in the DW PCIe Root Port/Endpoint driver.
Link: https://lore.kernel.org/linux-pci/[email protected]/
Changelog v2:
- Drop the patches:
[PATCH 1/25] dmaengine: dw-edma: Drop dma_slave_config.direction field usage
[PATCH 2/25] dmaengine: dw-edma: Fix eDMA Rd/Wr-channels and DMA-direction semantics
since they are going to be merged in in the framework of the
Frank's patchset.
- Add a new patch: "dmaengine: dw-edma: Release requested IRQs on
failure."
- Drop __iomem qualifier from the struct dw_edma_debugfs_entry instance
definition in the dw_edma_debugfs_u32_get() method. (@Manivannan)
- Add a new patch: "dmaengine: dw-edma: Rename DebugFS dentry variables to
'dent'." (@Manivannan)
- Slightly extend the eDMA name array size. (@Manivannan)
- Change the specific DMA mapping comment a bit to being
clearer. (@Manivannan)
- Add a new patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
method."
- Don't fail eDMA detection procedure if the DW eDMA driver couldn't probe
device. That happens if the driver is disabled. (@Manivannan)
- Add "dma" registers resource mapping procedure. (@Manivannan)
- Move the eDMA CSRs space detection into the dw_pcie_map_detect() method.
- Remove eDMA on the dw_pcie_ep_init() internal errors. (@Manivannan)
- Remove eDMA in the dw_pcie_ep_exit() method.
- Move the dw_pcie_edma_detect() method execution to the tail of the
dw_pcie_ep_init() function.
Link: https://lore.kernel.org/linux-pci/[email protected]/
Changelog v3:
- Conditionally set dchan->dev->device.dma_coherent field since it can
be missing on some platforms. (@Manivannan)
- Drop the patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
method". A similar modification has been done in another patchset.
- Add more comprehensive and less regression prune eDMA block detection
procedure.
- Drop the patch: "dma-direct: take dma-ranges/offsets into account in
resource mapping". It will be separately reviewed.
- Remove Manivannan tb tag from the modified patches.
- Rebase onto the kernel v5.18.
Link: https://lore.kernel.org/linux-pci/[email protected]
Changelog v4:
- Rabase onto the laters Frank Li series:
Link: https://lore.kernel.org/all/[email protected]/
- Add Vinod' Ab-tag.
- Rebase onto the kernel v5.19-rcX.
Link: https://lore.kernel.org/linux-pci/[email protected]
Changelog v5:
- Just resend.
- Rebase onto the kernel v6.0-rc2.
Signed-off-by: Serge Semin <[email protected]>
Reviewed-by: Manivannan Sadhasivam <[email protected]>
Acked-By: Vinod Koul <[email protected]>
Cc: Alexey Malahov <[email protected]>
Cc: Pavel Parkhomenko <[email protected]>
Cc: "Krzysztof Wilczyński" <[email protected]>
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Serge Semin (24):
dmaengine: Fix dma_slave_config.dst_addr description
dmaengine: dw-edma: Release requested IRQs on failure
dmaengine: dw-edma: Convert ll/dt phys-address to PCIe bus/DMA address
dmaengine: dw-edma: Fix missing src/dst address of the interleaved
xfers
dmaengine: dw-edma: Don't permit non-inc interleaved xfers
dmaengine: dw-edma: Fix invalid interleaved xfers semantics
dmaengine: dw-edma: Add CPU to PCIe bus address translation
dmaengine: dw-edma: Add PCIe bus address getter to the remote EP
glue-driver
dmaengine: dw-edma: Drop chancnt initialization
dmaengine: dw-edma: Fix DebugFS reg entry type
dmaengine: dw-edma: Stop checking debugfs_create_*() return value
dmaengine: dw-edma: Add dw_edma prefix to the DebugFS nodes descriptor
dmaengine: dw-edma: Convert DebugFS descs to being kz-allocated
dmaengine: dw-edma: Rename DebugFS dentry variables to 'dent'
dmaengine: dw-edma: Simplify the DebugFS context CSRs init procedure
dmaengine: dw-edma: Move eDMA data pointer to DebugFS node descriptor
dmaengine: dw-edma: Join Write/Read channels into a single device
dmaengine: dw-edma: Use DMA-engine device DebugFS subdirectory
dmaengine: dw-edma: Use non-atomic io-64 methods
dmaengine: dw-edma: Drop DT-region allocation
dmaengine: dw-edma: Replace chip ID number with device name
dmaengine: dw-edma: Bypass dma-ranges mapping for the local setup
dmaengine: dw-edma: Skip cleanup procedure if no private data found
PCI: dwc: Add DW eDMA engine support
drivers/dma/dw-edma/dw-edma-core.c | 216 +++++-----
drivers/dma/dw-edma/dw-edma-core.h | 10 +-
drivers/dma/dw-edma/dw-edma-pcie.c | 24 +-
drivers/dma/dw-edma/dw-edma-v0-core.c | 60 +--
drivers/dma/dw-edma/dw-edma-v0-core.h | 1 -
drivers/dma/dw-edma/dw-edma-v0-debugfs.c | 372 ++++++++----------
drivers/dma/dw-edma/dw-edma-v0-debugfs.h | 5 -
.../pci/controller/dwc/pcie-designware-ep.c | 12 +-
.../pci/controller/dwc/pcie-designware-host.c | 13 +-
drivers/pci/controller/dwc/pcie-designware.c | 186 +++++++++
drivers/pci/controller/dwc/pcie-designware.h | 20 +
include/linux/dma/edma.h | 18 +-
include/linux/dmaengine.h | 2 +-
13 files changed, 583 insertions(+), 356 deletions(-)
--
2.35.1
On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
> This is a final patchset in the series created in the framework of
> my Baikal-T1 PCIe/eDMA-related work:
>
> [1: Done v5] PCI: dwc: Various fixes and cleanups
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> Merged: kernel 6.0-rc1
> [2: Done v4] PCI: dwc: Add hw version and dma-ranges support
> Link: https://lore.kernel.org/linux-pci/[email protected]
> Merged: kernel 6.0-rc1
> [3: In-review v5] PCI: dwc: Add generic resources and Baikal-T1 support
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> [4: Done v5] dmaengine: dw-edma: Add RP/EP local DMA support
> Link: ---you are looking at it---
>
> Note it is very recommended to merge the patchsets in the same order as
> they are listed in the set above in order to have them applied smoothly.
> Nothing prevents them from being reviewed synchronously though.
>
> Please note originally this series was self content, but due to Frank
> being a bit faster in his work submission I had to rebase my patchset onto
> his one. So now this patchset turns to be dependent on the Frank' work:
>
> Link: https://lore.kernel.org/linux-pci/[email protected]/
>
> Here is a short summary regarding this patchset. The series starts with
> fixes patches. We discovered that the dw-edma-pcie.c driver incorrectly
> initializes the LL/DT base addresses for the platforms with not matching
> CPU and PCIe memory spaces. It is fixed by using the pci_bus_address()
> method to get a correct base address. After that you can find a series of
> the interleaved xfers fixes. It turned out the interleaved transfers
> implementation didn't work quite correctly from the very beginning for
> instance missing src/dst addresses initialization, etc. In the framework
> of the next two patches we suggest to add a new platform-specific
> callback - pci_address() and use it to convert the CPU address to the PCIe
> space address. It is at least required for the DW eDMA remote End-point
> setup on the platforms with not-matching CPU/PCIe address spaces. In case
> of the DW eDMA local RP/EP setup the conversion will be done automatically
> by the outbound iATU (if no DMA-bypass flag is specified for the
> corresponding iATU window). Then we introduce a set of the patches to make
> the DebugFS part of the code supporting the multi-eDMA controllers
> platforms. It starts with several cleanup patches and is closed joining
> the Read/Write channels into a single DMA-device as they originally should
> have been. After that you can find the patches with adding the non-atomic
> io-64 methods usage, dropping DT-region descriptors allocation, replacing
> chip IDs with the device name. In addition to that in order to have the
> eDMA embedded into the DW PCIe RP/EP supported we need to bypass the
> dma-ranges-based memory ranges mapping since in case of the root port DT
> node it's applicable for the peripheral PCIe devices only. Finally at the
> series closure we introduce a generic DW eDMA controller support being
> available in the DW PCIe Root Port/Endpoint driver.
>
I've tested this series on Qualcomm SM8450 SoC based dev board. So,
Tested-by: Manivannan Sadhasivam <[email protected]>
Not sure what is the merging strategy for this one but this series should get
merged into a single tree. Since the PCI patch is touching the designware
driver, merging the series into dmaengine tree might result in conflict later.
Thanks,
Mani
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> Changelog v2:
> - Drop the patches:
> [PATCH 1/25] dmaengine: dw-edma: Drop dma_slave_config.direction field usage
> [PATCH 2/25] dmaengine: dw-edma: Fix eDMA Rd/Wr-channels and DMA-direction semantics
> since they are going to be merged in in the framework of the
> Frank's patchset.
> - Add a new patch: "dmaengine: dw-edma: Release requested IRQs on
> failure."
> - Drop __iomem qualifier from the struct dw_edma_debugfs_entry instance
> definition in the dw_edma_debugfs_u32_get() method. (@Manivannan)
> - Add a new patch: "dmaengine: dw-edma: Rename DebugFS dentry variables to
> 'dent'." (@Manivannan)
> - Slightly extend the eDMA name array size. (@Manivannan)
> - Change the specific DMA mapping comment a bit to being
> clearer. (@Manivannan)
> - Add a new patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
> method."
> - Don't fail eDMA detection procedure if the DW eDMA driver couldn't probe
> device. That happens if the driver is disabled. (@Manivannan)
> - Add "dma" registers resource mapping procedure. (@Manivannan)
> - Move the eDMA CSRs space detection into the dw_pcie_map_detect() method.
> - Remove eDMA on the dw_pcie_ep_init() internal errors. (@Manivannan)
> - Remove eDMA in the dw_pcie_ep_exit() method.
> - Move the dw_pcie_edma_detect() method execution to the tail of the
> dw_pcie_ep_init() function.
>
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> Changelog v3:
> - Conditionally set dchan->dev->device.dma_coherent field since it can
> be missing on some platforms. (@Manivannan)
> - Drop the patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
> method". A similar modification has been done in another patchset.
> - Add more comprehensive and less regression prune eDMA block detection
> procedure.
> - Drop the patch: "dma-direct: take dma-ranges/offsets into account in
> resource mapping". It will be separately reviewed.
> - Remove Manivannan tb tag from the modified patches.
> - Rebase onto the kernel v5.18.
>
> Link: https://lore.kernel.org/linux-pci/[email protected]
> Changelog v4:
> - Rabase onto the laters Frank Li series:
> Link: https://lore.kernel.org/all/[email protected]/
> - Add Vinod' Ab-tag.
> - Rebase onto the kernel v5.19-rcX.
>
> Link: https://lore.kernel.org/linux-pci/[email protected]
> Changelog v5:
> - Just resend.
> - Rebase onto the kernel v6.0-rc2.
>
> Signed-off-by: Serge Semin <[email protected]>
> Reviewed-by: Manivannan Sadhasivam <[email protected]>
> Acked-By: Vinod Koul <[email protected]>
> Cc: Alexey Malahov <[email protected]>
> Cc: Pavel Parkhomenko <[email protected]>
> Cc: "Krzysztof Wilczyński" <[email protected]>
> Cc: [email protected]
> Cc: [email protected]
> Cc: [email protected]
>
> Serge Semin (24):
> dmaengine: Fix dma_slave_config.dst_addr description
> dmaengine: dw-edma: Release requested IRQs on failure
> dmaengine: dw-edma: Convert ll/dt phys-address to PCIe bus/DMA address
> dmaengine: dw-edma: Fix missing src/dst address of the interleaved
> xfers
> dmaengine: dw-edma: Don't permit non-inc interleaved xfers
> dmaengine: dw-edma: Fix invalid interleaved xfers semantics
> dmaengine: dw-edma: Add CPU to PCIe bus address translation
> dmaengine: dw-edma: Add PCIe bus address getter to the remote EP
> glue-driver
> dmaengine: dw-edma: Drop chancnt initialization
> dmaengine: dw-edma: Fix DebugFS reg entry type
> dmaengine: dw-edma: Stop checking debugfs_create_*() return value
> dmaengine: dw-edma: Add dw_edma prefix to the DebugFS nodes descriptor
> dmaengine: dw-edma: Convert DebugFS descs to being kz-allocated
> dmaengine: dw-edma: Rename DebugFS dentry variables to 'dent'
> dmaengine: dw-edma: Simplify the DebugFS context CSRs init procedure
> dmaengine: dw-edma: Move eDMA data pointer to DebugFS node descriptor
> dmaengine: dw-edma: Join Write/Read channels into a single device
> dmaengine: dw-edma: Use DMA-engine device DebugFS subdirectory
> dmaengine: dw-edma: Use non-atomic io-64 methods
> dmaengine: dw-edma: Drop DT-region allocation
> dmaengine: dw-edma: Replace chip ID number with device name
> dmaengine: dw-edma: Bypass dma-ranges mapping for the local setup
> dmaengine: dw-edma: Skip cleanup procedure if no private data found
> PCI: dwc: Add DW eDMA engine support
>
> drivers/dma/dw-edma/dw-edma-core.c | 216 +++++-----
> drivers/dma/dw-edma/dw-edma-core.h | 10 +-
> drivers/dma/dw-edma/dw-edma-pcie.c | 24 +-
> drivers/dma/dw-edma/dw-edma-v0-core.c | 60 +--
> drivers/dma/dw-edma/dw-edma-v0-core.h | 1 -
> drivers/dma/dw-edma/dw-edma-v0-debugfs.c | 372 ++++++++----------
> drivers/dma/dw-edma/dw-edma-v0-debugfs.h | 5 -
> .../pci/controller/dwc/pcie-designware-ep.c | 12 +-
> .../pci/controller/dwc/pcie-designware-host.c | 13 +-
> drivers/pci/controller/dwc/pcie-designware.c | 186 +++++++++
> drivers/pci/controller/dwc/pcie-designware.h | 20 +
> include/linux/dma/edma.h | 18 +-
> include/linux/dmaengine.h | 2 +-
> 13 files changed, 583 insertions(+), 356 deletions(-)
>
> --
> 2.35.1
>
--
மணிவண்ணன் சதாசிவம்
On Tue, Aug 23, 2022 at 09:15:26PM +0530, Manivannan Sadhasivam wrote:
> On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
> > This is a final patchset in the series created in the framework of
> > my Baikal-T1 PCIe/eDMA-related work:
> >
> > [1: Done v5] PCI: dwc: Various fixes and cleanups
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > Merged: kernel 6.0-rc1
> > [2: Done v4] PCI: dwc: Add hw version and dma-ranges support
> > Link: https://lore.kernel.org/linux-pci/[email protected]
> > Merged: kernel 6.0-rc1
> > [3: In-review v5] PCI: dwc: Add generic resources and Baikal-T1 support
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > [4: Done v5] dmaengine: dw-edma: Add RP/EP local DMA support
> > Link: ---you are looking at it---
> >
> > Note it is very recommended to merge the patchsets in the same order as
> > they are listed in the set above in order to have them applied smoothly.
> > Nothing prevents them from being reviewed synchronously though.
> >
> > Please note originally this series was self content, but due to Frank
> > being a bit faster in his work submission I had to rebase my patchset onto
> > his one. So now this patchset turns to be dependent on the Frank' work:
> >
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> >
> > Here is a short summary regarding this patchset. The series starts with
> > fixes patches. We discovered that the dw-edma-pcie.c driver incorrectly
> > initializes the LL/DT base addresses for the platforms with not matching
> > CPU and PCIe memory spaces. It is fixed by using the pci_bus_address()
> > method to get a correct base address. After that you can find a series of
> > the interleaved xfers fixes. It turned out the interleaved transfers
> > implementation didn't work quite correctly from the very beginning for
> > instance missing src/dst addresses initialization, etc. In the framework
> > of the next two patches we suggest to add a new platform-specific
> > callback - pci_address() and use it to convert the CPU address to the PCIe
> > space address. It is at least required for the DW eDMA remote End-point
> > setup on the platforms with not-matching CPU/PCIe address spaces. In case
> > of the DW eDMA local RP/EP setup the conversion will be done automatically
> > by the outbound iATU (if no DMA-bypass flag is specified for the
> > corresponding iATU window). Then we introduce a set of the patches to make
> > the DebugFS part of the code supporting the multi-eDMA controllers
> > platforms. It starts with several cleanup patches and is closed joining
> > the Read/Write channels into a single DMA-device as they originally should
> > have been. After that you can find the patches with adding the non-atomic
> > io-64 methods usage, dropping DT-region descriptors allocation, replacing
> > chip IDs with the device name. In addition to that in order to have the
> > eDMA embedded into the DW PCIe RP/EP supported we need to bypass the
> > dma-ranges-based memory ranges mapping since in case of the root port DT
> > node it's applicable for the peripheral PCIe devices only. Finally at the
> > series closure we introduce a generic DW eDMA controller support being
> > available in the DW PCIe Root Port/Endpoint driver.
> >
>
> I've tested this series on Qualcomm SM8450 SoC based dev board. So,
>
> Tested-by: Manivannan Sadhasivam <[email protected]>
>
Thanks.
> Not sure what is the merging strategy for this one but this series should get
> merged into a single tree. Since the PCI patch is touching the designware
> driver, merging the series into dmaengine tree might result in conflict later.
Right, the series
[PATCH v5 00/20] PCI: dwc: Add generic resources and Baikal-T1 support
is supposed to be merged in first. Then this one will get to be
applied with no conflicts. That's what I imply in the head of the
cover-letter.
-Sergey
>
> Thanks,
> Mani
>
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > Changelog v2:
> > - Drop the patches:
> > [PATCH 1/25] dmaengine: dw-edma: Drop dma_slave_config.direction field usage
> > [PATCH 2/25] dmaengine: dw-edma: Fix eDMA Rd/Wr-channels and DMA-direction semantics
> > since they are going to be merged in in the framework of the
> > Frank's patchset.
> > - Add a new patch: "dmaengine: dw-edma: Release requested IRQs on
> > failure."
> > - Drop __iomem qualifier from the struct dw_edma_debugfs_entry instance
> > definition in the dw_edma_debugfs_u32_get() method. (@Manivannan)
> > - Add a new patch: "dmaengine: dw-edma: Rename DebugFS dentry variables to
> > 'dent'." (@Manivannan)
> > - Slightly extend the eDMA name array size. (@Manivannan)
> > - Change the specific DMA mapping comment a bit to being
> > clearer. (@Manivannan)
> > - Add a new patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
> > method."
> > - Don't fail eDMA detection procedure if the DW eDMA driver couldn't probe
> > device. That happens if the driver is disabled. (@Manivannan)
> > - Add "dma" registers resource mapping procedure. (@Manivannan)
> > - Move the eDMA CSRs space detection into the dw_pcie_map_detect() method.
> > - Remove eDMA on the dw_pcie_ep_init() internal errors. (@Manivannan)
> > - Remove eDMA in the dw_pcie_ep_exit() method.
> > - Move the dw_pcie_edma_detect() method execution to the tail of the
> > dw_pcie_ep_init() function.
> >
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > Changelog v3:
> > - Conditionally set dchan->dev->device.dma_coherent field since it can
> > be missing on some platforms. (@Manivannan)
> > - Drop the patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
> > method". A similar modification has been done in another patchset.
> > - Add more comprehensive and less regression prune eDMA block detection
> > procedure.
> > - Drop the patch: "dma-direct: take dma-ranges/offsets into account in
> > resource mapping". It will be separately reviewed.
> > - Remove Manivannan tb tag from the modified patches.
> > - Rebase onto the kernel v5.18.
> >
> > Link: https://lore.kernel.org/linux-pci/[email protected]
> > Changelog v4:
> > - Rabase onto the laters Frank Li series:
> > Link: https://lore.kernel.org/all/[email protected]/
> > - Add Vinod' Ab-tag.
> > - Rebase onto the kernel v5.19-rcX.
> >
> > Link: https://lore.kernel.org/linux-pci/[email protected]
> > Changelog v5:
> > - Just resend.
> > - Rebase onto the kernel v6.0-rc2.
> >
> > Signed-off-by: Serge Semin <[email protected]>
> > Reviewed-by: Manivannan Sadhasivam <[email protected]>
> > Acked-By: Vinod Koul <[email protected]>
> > Cc: Alexey Malahov <[email protected]>
> > Cc: Pavel Parkhomenko <[email protected]>
> > Cc: "Krzysztof Wilczyński" <[email protected]>
> > Cc: [email protected]
> > Cc: [email protected]
> > Cc: [email protected]
> >
> > Serge Semin (24):
> > dmaengine: Fix dma_slave_config.dst_addr description
> > dmaengine: dw-edma: Release requested IRQs on failure
> > dmaengine: dw-edma: Convert ll/dt phys-address to PCIe bus/DMA address
> > dmaengine: dw-edma: Fix missing src/dst address of the interleaved
> > xfers
> > dmaengine: dw-edma: Don't permit non-inc interleaved xfers
> > dmaengine: dw-edma: Fix invalid interleaved xfers semantics
> > dmaengine: dw-edma: Add CPU to PCIe bus address translation
> > dmaengine: dw-edma: Add PCIe bus address getter to the remote EP
> > glue-driver
> > dmaengine: dw-edma: Drop chancnt initialization
> > dmaengine: dw-edma: Fix DebugFS reg entry type
> > dmaengine: dw-edma: Stop checking debugfs_create_*() return value
> > dmaengine: dw-edma: Add dw_edma prefix to the DebugFS nodes descriptor
> > dmaengine: dw-edma: Convert DebugFS descs to being kz-allocated
> > dmaengine: dw-edma: Rename DebugFS dentry variables to 'dent'
> > dmaengine: dw-edma: Simplify the DebugFS context CSRs init procedure
> > dmaengine: dw-edma: Move eDMA data pointer to DebugFS node descriptor
> > dmaengine: dw-edma: Join Write/Read channels into a single device
> > dmaengine: dw-edma: Use DMA-engine device DebugFS subdirectory
> > dmaengine: dw-edma: Use non-atomic io-64 methods
> > dmaengine: dw-edma: Drop DT-region allocation
> > dmaengine: dw-edma: Replace chip ID number with device name
> > dmaengine: dw-edma: Bypass dma-ranges mapping for the local setup
> > dmaengine: dw-edma: Skip cleanup procedure if no private data found
> > PCI: dwc: Add DW eDMA engine support
> >
> > drivers/dma/dw-edma/dw-edma-core.c | 216 +++++-----
> > drivers/dma/dw-edma/dw-edma-core.h | 10 +-
> > drivers/dma/dw-edma/dw-edma-pcie.c | 24 +-
> > drivers/dma/dw-edma/dw-edma-v0-core.c | 60 +--
> > drivers/dma/dw-edma/dw-edma-v0-core.h | 1 -
> > drivers/dma/dw-edma/dw-edma-v0-debugfs.c | 372 ++++++++----------
> > drivers/dma/dw-edma/dw-edma-v0-debugfs.h | 5 -
> > .../pci/controller/dwc/pcie-designware-ep.c | 12 +-
> > .../pci/controller/dwc/pcie-designware-host.c | 13 +-
> > drivers/pci/controller/dwc/pcie-designware.c | 186 +++++++++
> > drivers/pci/controller/dwc/pcie-designware.h | 20 +
> > include/linux/dma/edma.h | 18 +-
> > include/linux/dmaengine.h | 2 +-
> > 13 files changed, 583 insertions(+), 356 deletions(-)
> >
> > --
> > 2.35.1
> >
>
> --
> மணிவண்ணன் சதாசிவம்
On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
> This is a final patchset in the series created in the framework of
> my Baikal-T1 PCIe/eDMA-related work:
>
> [1: Done v5] PCI: dwc: Various fixes and cleanups
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> Merged: kernel 6.0-rc1
> [2: Done v4] PCI: dwc: Add hw version and dma-ranges support
> Link: https://lore.kernel.org/linux-pci/[email protected]
> Merged: kernel 6.0-rc1
> [3: In-review v5] PCI: dwc: Add generic resources and Baikal-T1 support
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> [4: Done v5] dmaengine: dw-edma: Add RP/EP local DMA support
> Link: ---you are looking at it---
> ...
> Please note originally this series was self content, but due to Frank
> being a bit faster in his work submission I had to rebase my patchset onto
> his one. So now this patchset turns to be dependent on the Frank' work:
>
> Link: https://lore.kernel.org/linux-pci/[email protected]/
I think this paragraph is obsolete, since the "Enable designware PCI
EP EDMA locally" series you reference is already upstream:
https://git.kernel.org/linus/94d13317bef3
What remains are items 3 and 4.
3 is mostly drivers/pci/ and DT bindings (Lorenzo). 4 is mostly
drivers/dma/dw-edma/ stuff (Gustavo). I guess Lorenzo and Gustavo can
figure out where it makes the most sense to merge it.
Bjorn
On Wed, Aug 24, 2022 at 11:39:16AM -0500, Bjorn Helgaas wrote:
> On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
> > This is a final patchset in the series created in the framework of
> > my Baikal-T1 PCIe/eDMA-related work:
> >
> > [1: Done v5] PCI: dwc: Various fixes and cleanups
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > Merged: kernel 6.0-rc1
> > [2: Done v4] PCI: dwc: Add hw version and dma-ranges support
> > Link: https://lore.kernel.org/linux-pci/[email protected]
> > Merged: kernel 6.0-rc1
> > [3: In-review v5] PCI: dwc: Add generic resources and Baikal-T1 support
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > [4: Done v5] dmaengine: dw-edma: Add RP/EP local DMA support
> > Link: ---you are looking at it---
> > ...
>
> > Please note originally this series was self content, but due to Frank
> > being a bit faster in his work submission I had to rebase my patchset onto
> > his one. So now this patchset turns to be dependent on the Frank' work:
> >
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
>
> I think this paragraph is obsolete, since the "Enable designware PCI
> EP EDMA locally" series you reference is already upstream:
> https://git.kernel.org/linus/94d13317bef3
Right.
>
> What remains are items 3 and 4.
>
Right.
> 3 is mostly drivers/pci/ and DT bindings (Lorenzo). 4 is mostly
> drivers/dma/dw-edma/ stuff (Gustavo). I guess Lorenzo and Gustavo can
> figure out where it makes the most sense to merge it.
As I already said to you two months ago:
Link: https://lore.kernel.org/linux-pci/20220617104143.yj2mlnj4twoxoeld@mobilestation/
Neither Gustavo Pimentel nor Jingoo Han have given any sign of
activity during all the time the patches being on review, discussed,
resubmitted while both of them are the DW PCIe and eDMA drivers
maintainers. Moreover none of them have been active in kernel for more
than a year:
Last Gustavo' tag could be tracked at Apr' 2021.
Last Han' tag could be found at Nov' 2020 commit.
So it's very much unluckily they'll just get back any time soon. We
can't wait for their opinion especially seeing that my patches have
been on review for almost half a year.
So as we already agreed with you the best solution would be to merge
this and the patchset #3 via your repo (or Lorenzo' repo if he is back
from his long-term absence). Vinod already acked the DMA-patches of
this series:
Link: https://lore.kernel.org/linux-pci/YtlDivjaXfSEK1Xg@matsya/
So the only thing left to settle is the DT-related part of the
patchset #3. Which I hope will be done before the next kernel merge
window. After it's done I'll ping you and Lorenzo. Ok?
-Sergey
>
> Bjorn
On 24-08-22, 17:07, Serge Semin wrote:
> On Tue, Aug 23, 2022 at 09:15:26PM +0530, Manivannan Sadhasivam wrote:
> > On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
> > I've tested this series on Qualcomm SM8450 SoC based dev board. So,
> >
> > Tested-by: Manivannan Sadhasivam <[email protected]>
> >
>
> Thanks.
>
> > Not sure what is the merging strategy for this one but this series should get
> > merged into a single tree. Since the PCI patch is touching the designware
> > driver, merging the series into dmaengine tree might result in conflict later.
>
> Right, the series
> [PATCH v5 00/20] PCI: dwc: Add generic resources and Baikal-T1 support
> is supposed to be merged in first. Then this one will get to be
> applied with no conflicts. That's what I imply in the head of the
> cover-letter.
I dont see a dependency of dma patches with PCIe patches? I guess they
could go thru the respective trees now..?
--
~Vinod
On Thu, Aug 25, 2022 at 10:12:23AM +0530, Vinod Koul wrote:
> On 24-08-22, 17:07, Serge Semin wrote:
> > On Tue, Aug 23, 2022 at 09:15:26PM +0530, Manivannan Sadhasivam wrote:
> > > On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
>
> > > I've tested this series on Qualcomm SM8450 SoC based dev board. So,
> > >
> > > Tested-by: Manivannan Sadhasivam <[email protected]>
> > >
> >
> > Thanks.
> >
> > > Not sure what is the merging strategy for this one but this series should get
> > > merged into a single tree. Since the PCI patch is touching the designware
> > > driver, merging the series into dmaengine tree might result in conflict later.
> >
> > Right, the series
> > [PATCH v5 00/20] PCI: dwc: Add generic resources and Baikal-T1 support
> > is supposed to be merged in first. Then this one will get to be
> > applied with no conflicts. That's what I imply in the head of the
> > cover-letter.
>
> I dont see a dependency of dma patches with PCIe patches? I guess they
> could go thru the respective trees now..?
There is a backward dependency: the PCIe patch in this series depends
on the eDMA patches and the patches in the patchset #3. So should you
merge the eDMA patches via your tree, the later patch in this series
and the patchset #3 would have needed to be applied in there too. So
the patches can't be split up between different branches. Seeing all
the changes (including the DW eDMA part) concern the PCIe device (DW
eDMA is a part of either DW PCIe End-point or Root Port) and we
already agreed to merge all the changes via the PCIe tree, I would
stick to the previous settled agreement.
-Sergey
>
> --
> ~Vinod
On 25-08-22, 08:04, Serge Semin wrote:
> On Thu, Aug 25, 2022 at 10:12:23AM +0530, Vinod Koul wrote:
> > On 24-08-22, 17:07, Serge Semin wrote:
> > > On Tue, Aug 23, 2022 at 09:15:26PM +0530, Manivannan Sadhasivam wrote:
> > > > On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
> >
> > > > I've tested this series on Qualcomm SM8450 SoC based dev board. So,
> > > >
> > > > Tested-by: Manivannan Sadhasivam <[email protected]>
> > > >
> > >
> > > Thanks.
> > >
> > > > Not sure what is the merging strategy for this one but this series should get
> > > > merged into a single tree. Since the PCI patch is touching the designware
> > > > driver, merging the series into dmaengine tree might result in conflict later.
> > >
> > > Right, the series
> > > [PATCH v5 00/20] PCI: dwc: Add generic resources and Baikal-T1 support
> > > is supposed to be merged in first. Then this one will get to be
> > > applied with no conflicts. That's what I imply in the head of the
> > > cover-letter.
> >
>
> > I dont see a dependency of dma patches with PCIe patches? I guess they
> > could go thru the respective trees now..?
>
> There is a backward dependency: the PCIe patch in this series depends
> on the eDMA patches and the patches in the patchset #3. So should you
What is the dependency...? Looking at the patches there does not seem to
be one...
> merge the eDMA patches via your tree, the later patch in this series
> and the patchset #3 would have needed to be applied in there too. So
> the patches can't be split up between different branches. Seeing all
> the changes (including the DW eDMA part) concern the PCIe device (DW
> eDMA is a part of either DW PCIe End-point or Root Port) and we
> already agreed to merge all the changes via the PCIe tree, I would
> stick to the previous settled agreement.
>
> -Sergey
>
> >
> > --
> > ~Vinod
--
~Vinod
On Thu, Aug 25, 2022 at 02:14:36PM +0530, Vinod Koul wrote:
> On 25-08-22, 08:04, Serge Semin wrote:
> > On Thu, Aug 25, 2022 at 10:12:23AM +0530, Vinod Koul wrote:
> > > On 24-08-22, 17:07, Serge Semin wrote:
> > > > On Tue, Aug 23, 2022 at 09:15:26PM +0530, Manivannan Sadhasivam wrote:
> > > > > On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
> > >
> > > > > I've tested this series on Qualcomm SM8450 SoC based dev board. So,
> > > > >
> > > > > Tested-by: Manivannan Sadhasivam <[email protected]>
> > > > >
> > > >
> > > > Thanks.
> > > >
> > > > > Not sure what is the merging strategy for this one but this series should get
> > > > > merged into a single tree. Since the PCI patch is touching the designware
> > > > > driver, merging the series into dmaengine tree might result in conflict later.
> > > >
> > > > Right, the series
> > > > [PATCH v5 00/20] PCI: dwc: Add generic resources and Baikal-T1 support
> > > > is supposed to be merged in first. Then this one will get to be
> > > > applied with no conflicts. That's what I imply in the head of the
> > > > cover-letter.
> > >
> >
> > > I dont see a dependency of dma patches with PCIe patches? I guess they
> > > could go thru the respective trees now..?
> >
> > There is a backward dependency: the PCIe patch in this series depends
> > on the eDMA patches and the patches in the patchset #3. So should you
>
> What is the dependency...? Looking at the patches there does not seem to
> be one...
[PATCH RESEND v5 24/24] PCI: dwc: Add DW eDMA engine support:
|
+-> depends on the modifications done in the framework DW eDMA driver
| patchset, for instance the changes introduced in the patch
| [PATCH RESEND v5 22/24] dmaengine: dw-edma: Bypass dma-ranges mapping for the local setup
| make sure the dma-ranges property isn't taken into account for the
| Local CPU/Application setup (See it makes the DW_EDMA_CHIP_LOCAL flag
| used which is enabled for the eDMA embedded into the DW PCIe EP/RP).
| All the DebugFS-related and channels join updates are also required to
| make the DW eDMA driver working in the framework of the DW PCIe RP/EP
| device.
|
+-> at the very least depends on the changes introduced in the patchset #3:
| [PATCH v5 16/20] PCI: dwc: Introduce generic controller capabilities interface
| [PATCH v5 17/20] PCI: dwc: Introduce generic resources getter
| The patch at consideration adds CSR region request procedure in the
| method created and updated in these two patches.
There might be some other dependencies, but what I cited above must be
enough not to split the patchsets up between different branches
otherwise besides not properly working DW PCIe driver you'll have merge
conflicts.
-Sergey
>
> > merge the eDMA patches via your tree, the later patch in this series
> > and the patchset #3 would have needed to be applied in there too. So
> > the patches can't be split up between different branches. Seeing all
> > the changes (including the DW eDMA part) concern the PCIe device (DW
> > eDMA is a part of either DW PCIe End-point or Root Port) and we
> > already agreed to merge all the changes via the PCIe tree, I would
> > stick to the previous settled agreement.
> >
> > -Sergey
> >
> > >
> > > --
> > > ~Vinod
>
> --
> ~Vinod
On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
> This is a final patchset in the series created in the framework of
> my Baikal-T1 PCIe/eDMA-related work:
>
> [1: Done v5] PCI: dwc: Various fixes and cleanups
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> Merged: kernel 6.0-rc1
> [2: Done v4] PCI: dwc: Add hw version and dma-ranges support
> Link: https://lore.kernel.org/linux-pci/[email protected]
> Merged: kernel 6.0-rc1
> [3: In-review v5] PCI: dwc: Add generic resources and Baikal-T1 support
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> [4: Done v5] dmaengine: dw-edma: Add RP/EP local DMA support
> Link: ---you are looking at it---
>
> Note it is very recommended to merge the patchsets in the same order as
> they are listed in the set above in order to have them applied smoothly.
> Nothing prevents them from being reviewed synchronously though.
>
> Please note originally this series was self content, but due to Frank
> being a bit faster in his work submission I had to rebase my patchset onto
> his one. So now this patchset turns to be dependent on the Frank' work:
>
> Link: https://lore.kernel.org/linux-pci/[email protected]/
>
> Here is a short summary regarding this patchset. The series starts with
> fixes patches. We discovered that the dw-edma-pcie.c driver incorrectly
> initializes the LL/DT base addresses for the platforms with not matching
> CPU and PCIe memory spaces. It is fixed by using the pci_bus_address()
> method to get a correct base address. After that you can find a series of
> the interleaved xfers fixes. It turned out the interleaved transfers
> implementation didn't work quite correctly from the very beginning for
> instance missing src/dst addresses initialization, etc. In the framework
> of the next two patches we suggest to add a new platform-specific
> callback - pci_address() and use it to convert the CPU address to the PCIe
> space address. It is at least required for the DW eDMA remote End-point
> setup on the platforms with not-matching CPU/PCIe address spaces. In case
> of the DW eDMA local RP/EP setup the conversion will be done automatically
> by the outbound iATU (if no DMA-bypass flag is specified for the
> corresponding iATU window). Then we introduce a set of the patches to make
> the DebugFS part of the code supporting the multi-eDMA controllers
> platforms. It starts with several cleanup patches and is closed joining
> the Read/Write channels into a single DMA-device as they originally should
> have been. After that you can find the patches with adding the non-atomic
> io-64 methods usage, dropping DT-region descriptors allocation, replacing
> chip IDs with the device name. In addition to that in order to have the
> eDMA embedded into the DW PCIe RP/EP supported we need to bypass the
> dma-ranges-based memory ranges mapping since in case of the root port DT
> node it's applicable for the peripheral PCIe devices only. Finally at the
> series closure we introduce a generic DW eDMA controller support being
> available in the DW PCIe Root Port/Endpoint driver.
>
Looks like this series still got stuck :( And most of the discussion is with
patch 22/24 and it is for the Baikal platform (at present) I believe. There is
also a discussion on 24/24, but that's not a big deal as 22/24.
Can we exclude the patch 22/24 (even 24/24 if required) and merge rest in the
meantime. The patch(es) can still be submitted separately and merged after
reaching consensus.
Thanks,
Mani
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> Changelog v2:
> - Drop the patches:
> [PATCH 1/25] dmaengine: dw-edma: Drop dma_slave_config.direction field usage
> [PATCH 2/25] dmaengine: dw-edma: Fix eDMA Rd/Wr-channels and DMA-direction semantics
> since they are going to be merged in in the framework of the
> Frank's patchset.
> - Add a new patch: "dmaengine: dw-edma: Release requested IRQs on
> failure."
> - Drop __iomem qualifier from the struct dw_edma_debugfs_entry instance
> definition in the dw_edma_debugfs_u32_get() method. (@Manivannan)
> - Add a new patch: "dmaengine: dw-edma: Rename DebugFS dentry variables to
> 'dent'." (@Manivannan)
> - Slightly extend the eDMA name array size. (@Manivannan)
> - Change the specific DMA mapping comment a bit to being
> clearer. (@Manivannan)
> - Add a new patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
> method."
> - Don't fail eDMA detection procedure if the DW eDMA driver couldn't probe
> device. That happens if the driver is disabled. (@Manivannan)
> - Add "dma" registers resource mapping procedure. (@Manivannan)
> - Move the eDMA CSRs space detection into the dw_pcie_map_detect() method.
> - Remove eDMA on the dw_pcie_ep_init() internal errors. (@Manivannan)
> - Remove eDMA in the dw_pcie_ep_exit() method.
> - Move the dw_pcie_edma_detect() method execution to the tail of the
> dw_pcie_ep_init() function.
>
> Link: https://lore.kernel.org/linux-pci/[email protected]/
> Changelog v3:
> - Conditionally set dchan->dev->device.dma_coherent field since it can
> be missing on some platforms. (@Manivannan)
> - Drop the patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
> method". A similar modification has been done in another patchset.
> - Add more comprehensive and less regression prune eDMA block detection
> procedure.
> - Drop the patch: "dma-direct: take dma-ranges/offsets into account in
> resource mapping". It will be separately reviewed.
> - Remove Manivannan tb tag from the modified patches.
> - Rebase onto the kernel v5.18.
>
> Link: https://lore.kernel.org/linux-pci/[email protected]
> Changelog v4:
> - Rabase onto the laters Frank Li series:
> Link: https://lore.kernel.org/all/[email protected]/
> - Add Vinod' Ab-tag.
> - Rebase onto the kernel v5.19-rcX.
>
> Link: https://lore.kernel.org/linux-pci/[email protected]
> Changelog v5:
> - Just resend.
> - Rebase onto the kernel v6.0-rc2.
>
> Signed-off-by: Serge Semin <[email protected]>
> Reviewed-by: Manivannan Sadhasivam <[email protected]>
> Acked-By: Vinod Koul <[email protected]>
> Cc: Alexey Malahov <[email protected]>
> Cc: Pavel Parkhomenko <[email protected]>
> Cc: "Krzysztof Wilczyński" <[email protected]>
> Cc: [email protected]
> Cc: [email protected]
> Cc: [email protected]
>
> Serge Semin (24):
> dmaengine: Fix dma_slave_config.dst_addr description
> dmaengine: dw-edma: Release requested IRQs on failure
> dmaengine: dw-edma: Convert ll/dt phys-address to PCIe bus/DMA address
> dmaengine: dw-edma: Fix missing src/dst address of the interleaved
> xfers
> dmaengine: dw-edma: Don't permit non-inc interleaved xfers
> dmaengine: dw-edma: Fix invalid interleaved xfers semantics
> dmaengine: dw-edma: Add CPU to PCIe bus address translation
> dmaengine: dw-edma: Add PCIe bus address getter to the remote EP
> glue-driver
> dmaengine: dw-edma: Drop chancnt initialization
> dmaengine: dw-edma: Fix DebugFS reg entry type
> dmaengine: dw-edma: Stop checking debugfs_create_*() return value
> dmaengine: dw-edma: Add dw_edma prefix to the DebugFS nodes descriptor
> dmaengine: dw-edma: Convert DebugFS descs to being kz-allocated
> dmaengine: dw-edma: Rename DebugFS dentry variables to 'dent'
> dmaengine: dw-edma: Simplify the DebugFS context CSRs init procedure
> dmaengine: dw-edma: Move eDMA data pointer to DebugFS node descriptor
> dmaengine: dw-edma: Join Write/Read channels into a single device
> dmaengine: dw-edma: Use DMA-engine device DebugFS subdirectory
> dmaengine: dw-edma: Use non-atomic io-64 methods
> dmaengine: dw-edma: Drop DT-region allocation
> dmaengine: dw-edma: Replace chip ID number with device name
> dmaengine: dw-edma: Bypass dma-ranges mapping for the local setup
> dmaengine: dw-edma: Skip cleanup procedure if no private data found
> PCI: dwc: Add DW eDMA engine support
>
> drivers/dma/dw-edma/dw-edma-core.c | 216 +++++-----
> drivers/dma/dw-edma/dw-edma-core.h | 10 +-
> drivers/dma/dw-edma/dw-edma-pcie.c | 24 +-
> drivers/dma/dw-edma/dw-edma-v0-core.c | 60 +--
> drivers/dma/dw-edma/dw-edma-v0-core.h | 1 -
> drivers/dma/dw-edma/dw-edma-v0-debugfs.c | 372 ++++++++----------
> drivers/dma/dw-edma/dw-edma-v0-debugfs.h | 5 -
> .../pci/controller/dwc/pcie-designware-ep.c | 12 +-
> .../pci/controller/dwc/pcie-designware-host.c | 13 +-
> drivers/pci/controller/dwc/pcie-designware.c | 186 +++++++++
> drivers/pci/controller/dwc/pcie-designware.h | 20 +
> include/linux/dma/edma.h | 18 +-
> include/linux/dmaengine.h | 2 +-
> 13 files changed, 583 insertions(+), 356 deletions(-)
>
> --
> 2.35.1
>
--
மணிவண்ணன் சதாசிவம்
On Tue, Oct 25, 2022 at 01:29:30PM +0530, Manivannan Sadhasivam wrote:
> On Mon, Aug 22, 2022 at 09:53:08PM +0300, Serge Semin wrote:
> > This is a final patchset in the series created in the framework of
> > my Baikal-T1 PCIe/eDMA-related work:
> >
> > [1: Done v5] PCI: dwc: Various fixes and cleanups
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > Merged: kernel 6.0-rc1
> > [2: Done v4] PCI: dwc: Add hw version and dma-ranges support
> > Link: https://lore.kernel.org/linux-pci/[email protected]
> > Merged: kernel 6.0-rc1
> > [3: In-review v5] PCI: dwc: Add generic resources and Baikal-T1 support
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > [4: Done v5] dmaengine: dw-edma: Add RP/EP local DMA support
> > Link: ---you are looking at it---
> >
> > Note it is very recommended to merge the patchsets in the same order as
> > they are listed in the set above in order to have them applied smoothly.
> > Nothing prevents them from being reviewed synchronously though.
> >
> > Please note originally this series was self content, but due to Frank
> > being a bit faster in his work submission I had to rebase my patchset onto
> > his one. So now this patchset turns to be dependent on the Frank' work:
> >
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> >
> > Here is a short summary regarding this patchset. The series starts with
> > fixes patches. We discovered that the dw-edma-pcie.c driver incorrectly
> > initializes the LL/DT base addresses for the platforms with not matching
> > CPU and PCIe memory spaces. It is fixed by using the pci_bus_address()
> > method to get a correct base address. After that you can find a series of
> > the interleaved xfers fixes. It turned out the interleaved transfers
> > implementation didn't work quite correctly from the very beginning for
> > instance missing src/dst addresses initialization, etc. In the framework
> > of the next two patches we suggest to add a new platform-specific
> > callback - pci_address() and use it to convert the CPU address to the PCIe
> > space address. It is at least required for the DW eDMA remote End-point
> > setup on the platforms with not-matching CPU/PCIe address spaces. In case
> > of the DW eDMA local RP/EP setup the conversion will be done automatically
> > by the outbound iATU (if no DMA-bypass flag is specified for the
> > corresponding iATU window). Then we introduce a set of the patches to make
> > the DebugFS part of the code supporting the multi-eDMA controllers
> > platforms. It starts with several cleanup patches and is closed joining
> > the Read/Write channels into a single DMA-device as they originally should
> > have been. After that you can find the patches with adding the non-atomic
> > io-64 methods usage, dropping DT-region descriptors allocation, replacing
> > chip IDs with the device name. In addition to that in order to have the
> > eDMA embedded into the DW PCIe RP/EP supported we need to bypass the
> > dma-ranges-based memory ranges mapping since in case of the root port DT
> > node it's applicable for the peripheral PCIe devices only. Finally at the
> > series closure we introduce a generic DW eDMA controller support being
> > available in the DW PCIe Root Port/Endpoint driver.
> >
>
> Looks like this series still got stuck :( And most of the discussion is with
> patch 22/24 and it is for the Baikal platform (at present) I believe. There is
> also a discussion on 24/24, but that's not a big deal as 22/24.
>
> Can we exclude the patch 22/24 (even 24/24 if required) and merge rest in the
> meantime. The patch(es) can still be submitted separately and merged after
> reaching consensus.
The discussion has come to some conclusion in the framework of another
series:
https://lore.kernel.org/linux-pci/20221007224515.sseyabdfa2phcsdz@mobilestation/
It includes solutions for several problems raised by @Robin Murphy.
Seeing no objection from @Robin received and the solutions were mainly
suggested by @Robin himself I'll implement them in the next series
re-spin. So to speak don't worry. The series isn't stuck. I'll
resubmit it for review on this week.
-Sergey
>
> Thanks,
> Mani
>
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > Changelog v2:
> > - Drop the patches:
> > [PATCH 1/25] dmaengine: dw-edma: Drop dma_slave_config.direction field usage
> > [PATCH 2/25] dmaengine: dw-edma: Fix eDMA Rd/Wr-channels and DMA-direction semantics
> > since they are going to be merged in in the framework of the
> > Frank's patchset.
> > - Add a new patch: "dmaengine: dw-edma: Release requested IRQs on
> > failure."
> > - Drop __iomem qualifier from the struct dw_edma_debugfs_entry instance
> > definition in the dw_edma_debugfs_u32_get() method. (@Manivannan)
> > - Add a new patch: "dmaengine: dw-edma: Rename DebugFS dentry variables to
> > 'dent'." (@Manivannan)
> > - Slightly extend the eDMA name array size. (@Manivannan)
> > - Change the specific DMA mapping comment a bit to being
> > clearer. (@Manivannan)
> > - Add a new patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
> > method."
> > - Don't fail eDMA detection procedure if the DW eDMA driver couldn't probe
> > device. That happens if the driver is disabled. (@Manivannan)
> > - Add "dma" registers resource mapping procedure. (@Manivannan)
> > - Move the eDMA CSRs space detection into the dw_pcie_map_detect() method.
> > - Remove eDMA on the dw_pcie_ep_init() internal errors. (@Manivannan)
> > - Remove eDMA in the dw_pcie_ep_exit() method.
> > - Move the dw_pcie_edma_detect() method execution to the tail of the
> > dw_pcie_ep_init() function.
> >
> > Link: https://lore.kernel.org/linux-pci/[email protected]/
> > Changelog v3:
> > - Conditionally set dchan->dev->device.dma_coherent field since it can
> > be missing on some platforms. (@Manivannan)
> > - Drop the patch: "PCI: dwc: Add generic iATU/eDMA CSRs space detection
> > method". A similar modification has been done in another patchset.
> > - Add more comprehensive and less regression prune eDMA block detection
> > procedure.
> > - Drop the patch: "dma-direct: take dma-ranges/offsets into account in
> > resource mapping". It will be separately reviewed.
> > - Remove Manivannan tb tag from the modified patches.
> > - Rebase onto the kernel v5.18.
> >
> > Link: https://lore.kernel.org/linux-pci/[email protected]
> > Changelog v4:
> > - Rabase onto the laters Frank Li series:
> > Link: https://lore.kernel.org/all/[email protected]/
> > - Add Vinod' Ab-tag.
> > - Rebase onto the kernel v5.19-rcX.
> >
> > Link: https://lore.kernel.org/linux-pci/[email protected]
> > Changelog v5:
> > - Just resend.
> > - Rebase onto the kernel v6.0-rc2.
> >
> > Signed-off-by: Serge Semin <[email protected]>
> > Reviewed-by: Manivannan Sadhasivam <[email protected]>
> > Acked-By: Vinod Koul <[email protected]>
> > Cc: Alexey Malahov <[email protected]>
> > Cc: Pavel Parkhomenko <[email protected]>
> > Cc: "Krzysztof Wilczyński" <[email protected]>
> > Cc: [email protected]
> > Cc: [email protected]
> > Cc: [email protected]
> >
> > Serge Semin (24):
> > dmaengine: Fix dma_slave_config.dst_addr description
> > dmaengine: dw-edma: Release requested IRQs on failure
> > dmaengine: dw-edma: Convert ll/dt phys-address to PCIe bus/DMA address
> > dmaengine: dw-edma: Fix missing src/dst address of the interleaved
> > xfers
> > dmaengine: dw-edma: Don't permit non-inc interleaved xfers
> > dmaengine: dw-edma: Fix invalid interleaved xfers semantics
> > dmaengine: dw-edma: Add CPU to PCIe bus address translation
> > dmaengine: dw-edma: Add PCIe bus address getter to the remote EP
> > glue-driver
> > dmaengine: dw-edma: Drop chancnt initialization
> > dmaengine: dw-edma: Fix DebugFS reg entry type
> > dmaengine: dw-edma: Stop checking debugfs_create_*() return value
> > dmaengine: dw-edma: Add dw_edma prefix to the DebugFS nodes descriptor
> > dmaengine: dw-edma: Convert DebugFS descs to being kz-allocated
> > dmaengine: dw-edma: Rename DebugFS dentry variables to 'dent'
> > dmaengine: dw-edma: Simplify the DebugFS context CSRs init procedure
> > dmaengine: dw-edma: Move eDMA data pointer to DebugFS node descriptor
> > dmaengine: dw-edma: Join Write/Read channels into a single device
> > dmaengine: dw-edma: Use DMA-engine device DebugFS subdirectory
> > dmaengine: dw-edma: Use non-atomic io-64 methods
> > dmaengine: dw-edma: Drop DT-region allocation
> > dmaengine: dw-edma: Replace chip ID number with device name
> > dmaengine: dw-edma: Bypass dma-ranges mapping for the local setup
> > dmaengine: dw-edma: Skip cleanup procedure if no private data found
> > PCI: dwc: Add DW eDMA engine support
> >
> > drivers/dma/dw-edma/dw-edma-core.c | 216 +++++-----
> > drivers/dma/dw-edma/dw-edma-core.h | 10 +-
> > drivers/dma/dw-edma/dw-edma-pcie.c | 24 +-
> > drivers/dma/dw-edma/dw-edma-v0-core.c | 60 +--
> > drivers/dma/dw-edma/dw-edma-v0-core.h | 1 -
> > drivers/dma/dw-edma/dw-edma-v0-debugfs.c | 372 ++++++++----------
> > drivers/dma/dw-edma/dw-edma-v0-debugfs.h | 5 -
> > .../pci/controller/dwc/pcie-designware-ep.c | 12 +-
> > .../pci/controller/dwc/pcie-designware-host.c | 13 +-
> > drivers/pci/controller/dwc/pcie-designware.c | 186 +++++++++
> > drivers/pci/controller/dwc/pcie-designware.h | 20 +
> > include/linux/dma/edma.h | 18 +-
> > include/linux/dmaengine.h | 2 +-
> > 13 files changed, 583 insertions(+), 356 deletions(-)
> >
> > --
> > 2.35.1
> >
>
> --
> மணிவண்ணன் சதாசிவம்