| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
| |
Enable the PCI-to-AHB bridge from the host-side.
Note: On my test system, this didn't seem to be required, it's possible
the bit was stuck on when enabled.
Tested: Not tested.
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: Idbea0e9d378ee108f24240d64b43b6a3c0eea799
|
|
|
|
|
|
|
|
| |
The read method returns bool, therefore if this returns 0, it's false,
and therefore the read failed.
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: I365f9edf49a79ca5ea3c13090d6b8a11b9598e53
|
|
|
|
|
| |
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: I9da1674d6cbc688efc7bab0e033788d6ee4694f7
|
|
|
|
|
|
|
|
|
|
| |
The host tool depended on the BMC header: firmware_handler.hpp because
it defined the flags. This header depends on phosphor-ipmi-blobs,
therefore the host-tool depends on this. Move the flags into a separate
common header file and snip this dependency.
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: Id2ad33a1611c02605a1ed5c695429d0451eb98d4
|
|
|
|
|
|
|
|
|
| |
(style) The scope of the variable 'bytesRead' can be reduced.
[tools/bt.cpp:21]
[tools/p2a.cpp:111]
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: I6b96247a02cd35d073b424501bb161e80aedf3b4
|
|
|
|
|
|
|
|
| |
The blobs namespace is really owned by phosphor-ipmi-blobs. Move all
phosphor-ipmi-flash objects into ipmi_flash namespace.
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: I51b721ad4f3cdc4cc43846c942967ed2e5f16589
|
|
|
|
|
|
|
|
|
| |
Tool worked by declaring its own pci utility object instead of using the
one provided. There are not yet unit-tests associated with the
P2aDataHandler that would have caught this.
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: I62cdeb17d298f1db51d8f513c59bfd299fc2884a
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Implement sending firmware image over p2a via the aspeed-p2a-ctrl
driver.
Test configuration:
# Image is static, uses the PCI bridge, and requires the ASPEED
# PCI-to-AHB hardware implementation.
EXTRA_OECONF_append_quanta-q71l = " --enable-static-layout"
EXTRA_OECONF_append_quanta-q71l = " --enable-pci-bridge"
EXTRA_OECONF_append_quanta-q71l = " --enable-aspeed-p2a"
EXTRA_OECONF_append_quanta-q71l = " MAPPED_ADDRESS=0x47FF0000"
Tested: Verified via md5sum the image-bmc file sent from the host via
this tool matches the hash of /run/initramfs/bmc-image. This code can
be used to send any file down, but was only tested with a "static"
layout build.
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: I430bc7444f06f4f93a63e3bf2646bd195eaa2e52
|
|
|
|
|
|
|
|
|
|
|
| |
Read the PCI configuration from the blob handler itself. Don't use this
yet, but read it.
Tested: Verified with a bugfix in ipmi-blob-tool that it was able to
read back and print the value configured into the BMC image.
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: I49a8c93508e64a65fd37b50ec2b437f0016272f2
|
|
|
|
|
|
|
|
|
|
|
| |
Tools bugfix where if the result isn't assigned, it's still used.
Therefore, add a proper exit condition. This was developed on a system
with the device, and interestingly, this compilation warning->error
wasn't hit during CI, only with bitbake. Therefore, different flags
must be involved between the two.
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: Ifca7046657b65a425b0c3bf4fe05dd42bd7a8a14
|
|
|
|
|
|
|
|
|
| |
Check whether the pci-to-ahb bridge is enabled. It should be enabled
because this comes after the firmware blob is opened for writing.
Tested: Verified it reports the correct state of the value.
Change-Id: I04bd21fdbf65938164f5845c0ec46e2231b17bd9
Signed-off-by: Patrick Venture <venture@google.com>
|
|
Add initial pci-to-ahb host-side support, by just listing the aspeed pci
devices.
Tested: This dumped the aspeed PCI device listed on my test platform.
Change-Id: I6dc4aeb3b38ec2f95bfd716bda2d8eb4786328ab
Signed-off-by: Patrick Venture <venture@google.com>
|