Tegra20's PCIe controller has a couple of quirks. There are workarounds in
the driver for these, but they don't work after the DM conversion:

1) The PCI_CLASS value is wrong in HW.

This is worked around in pci_tegra_read_config() by patching up the value
read from that register. Pre-DM, the PCIe core always read this via a
16-bit access to the 16-bit offset 0xa. With DM, 32-bit accesses are used,
so we need to check for offset 0x8 instead. Mask the offset value back to
32-bit alignment to make this work in all cases.

2) Accessing devices other than dev 1 causes a data abort.

Pre-DM, this was worked around in pci_skip_dev(), which the PCIe core code
called during enumeration while iterating over a bus. The DM PCIe core
doesn't use this function. Instead, enhance tegra_pcie_conf_address() to
validate the bdf being accessed, and refuse to access invalid devices.
Since pci_skip_dev() isn't used, delete it.

I've also validated that both these WARs are only needed for Tegra20, by
testing on Tegra30/Cardhu and Tegra124/Jetson TKx. So, compile them in
conditionally.

Fixes: e81ca88451cf ("dm: tegra: pci: Convert tegra boards to driver model for PCI")
Signed-off-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Warren <twarren@nvidia.com>

The terminal condition in the area where a PCI device is scanned is wrong,
and 1f.7 isn't scanned.

Signed-off-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>

msi-map properties are used to tell an OS how PCI requester IDs are
mapped to ARM SMMU stream IDs.

for all PCI devices discovered in a system:
  -allocate a LUT (look-up-table) entry in that PCI controller
  -allocate a stream ID for the device
  -program and enable a LUT entry (maps PCI requester id to stream ID)
  -set the msi-map property on the controller reflecting the
   LUT mapping

basic bus scanning loop/logic was taken from drivers/pci/pci.c
pci_hose_scan_bus().

Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Reviewed-by: York Sun <york.sun@nxp.com>

Remove stream ID partitioning support that has been made
obsolete by upstream device tree bindings that specify how
representing how PCI requester IDs are mapped to MSI specifiers
and SMMU stream IDs.

Signed-off-by: Stuart Yoder <stuart.yoder@nxp.com>
Reviewed-by: York Sun <york.sun@nxp.com>

Two comments are missing a parameter and there is an extra blank line. Also
two of the region access macros are misnamed. Correct these problems.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>

It is common to read a config register value, clear and set some bits, then
write back the updated value. Add functions to do this in one step, for
convenience.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>

Each region is displayed in almost the same way. Break out this common code
into its own function.

Signed-off-by: Simon Glass <sjg@chromium.org>
Tested-by: Stephen Warren <swarren@nvidia.com>

Use this new function in places where it simplifies the code.

Signed-off-by: Simon Glass <sjg@chromium.org>

Fix the following compiler warnings when DEBUG is on.

warning: 'bar_res' may be used uninitialized in this function.
drivers/pci/pci_auto.c:101:21:
   if (!enum_only && pciauto_region_allocate(bar_res, bar_size,
                        ^

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

Building pci_rom.c with my toolchain complains about may be used uninitialized
rom varaible:

---8<---
+drivers/pci/pci_rom.c:269:25: note: 'rom' was declared here
w+drivers/pci/pci_rom.c: In function 'dm_pci_run_vga_bios':
w+drivers/pci/pci_rom.c:154:14: warning: 'rom' may be used uninitialized in this function [-Wmaybe-uninitialized]
--->8---

Fix this as done in 55616b86c745fcac5a791268ab8e7cba36965c0f the ram variable.

Signed-off-by: Andreas Bießmann <andreas.devel@googlemail.com>
Acked-by: Anatolij Gustschin <agust@denx.de>