Fpgamatt/subdevid 0001

[fpgamatt]

PR Title should start with one of the following tags: [Fix]/[Feature]/[Style]/[Update]

[Fix]- Bug Fix

[Feature]- for new feature

[Style]- Grammar or branding fix

[Update]-For an update to an existing feature

------------- Keep everything below this line -------------------------

Description

Describe the issue, update, change or fix and why

Collateral (docs, reports, design examples, case IDs):

• Document Update Required? (Specify FIM/AFU/Scripts)

Tests added:

Tests run:

[pcolberg]

Is the moved functionality really common to all boards; or is this a subvariant of the N6000? Is there a more specific name for these BMC-less Agilex-based(?) devices?

[fpgamatt]

The functionality moved to board_common exists on any board that has qsfp(s) and hssi subsystems.

We could call these boards something like pci_jtag_devkit.

[pcolberg]

The build fails due to missing link dependencies:

opae-sdk/libraries/libboard/board_n6000/CMakeLists.txt

Line 35 in afb66cd

opaeuio

[ 86%] Linking CXX executable ../../bin/test_board_d5005_c
/usr/bin/ld: ../../lib/libboard-common-static.a(board_common.c.o): in function `print_common_phy_info':
/home/runner/work/opae-sdk/opae-sdk/libraries/libboard/board_common/board_common.c:1083: undefined reference to `opae_uio_open'
/usr/bin/ld: /home/runner/work/opae-sdk/opae-sdk/libraries/libboard/board_common/board_common.c:1089: undefined reference to `opae_uio_region_get'
/usr/bin/ld: /home/runner/work/opae-sdk/opae-sdk/libraries/libboard/board_common/board_common.c:1092: undefined reference to `opae_uio_close'
/usr/bin/ld: /home/runner/work/opae-sdk/opae-sdk/libraries/libboard/board_common/board_common.c:1101: undefined reference to `opae_uio_close'
collect2: error: ld returned 1 exit status
make[2]: *** [tests/board/CMakeFiles/test_board_d5005_c.dir/build.make:124: bin/test_board_d5005_c] Error 1
make[1]: *** [CMakeFiles/Makefile2:6683: tests/board/CMakeFiles/test_board_d5005_c.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
[ 86%] Building CXX object tests/board/CMakeFiles/test_board_c6100_c.dir/__/framework/mock/opae_mock.cpp.o
[ 87%] Building CXX object tests/board/CMakeFiles/test_board_n6000_c.dir/__/framework/mock/opae_mock.cpp.o
[ 87%] Building CXX object tests/board/CMakeFiles/test_board_n5010_c.dir/__/framework/mock/opae_mock.cpp.o
[ 87%] Linking CXX executable ../../bin/test_board_n5010_c
/usr/bin/ld: ../../lib/libboard-common-static.a(board_common.c.o): in function `print_common_phy_info':
/home/runner/work/opae-sdk/opae-sdk/libraries/libboard/board_common/board_common.c:1083: undefined reference to `opae_uio_open'
/usr/bin/ld: /home/runner/work/opae-sdk/opae-sdk/libraries/libboard/board_common/board_common.c:1089: undefined reference to `opae_uio_region_get'
/usr/bin/ld: /home/runner/work/opae-sdk/opae-sdk/libraries/libboard/board_common/board_common.c:1092: undefined reference to `opae_uio_close'
/usr/bin/ld: /home/runner/work/opae-sdk/opae-sdk/libraries/libboard/board_common/board_common.c:1101: undefined reference to `opae_uio_close'
collect2: error: ld returned 1 exit status
make[2]: *** [tests/board/CMakeFiles/test_board_n5010_c.dir/build.make:124: bin/test_board_n5010_c] Error 1
make[1]: *** [CMakeFiles/Makefile2:6713: tests/board/CMakeFiles/test_board_n5010_c.dir/all] Error 2

[michael-adler]

Can someone please explain why this very invasive change is needed and why something simpler won't work? I see complexity being added both here and to the HW side.

[fpgamatt]

I would not classify this change as invasive. We are just adding a new board to the board plug-in framework of opae. This new board is a generic PCI devkit without a board BMC and is jtag only. Without this change, boards were incorrectly identifying themselves, and then opae would in turn mishandle the misidentified board.

[pcolberg]

I would not classify this change as invasive. We are just adding a new board to the board plug-in framework of opae. This new board is a generic PCI devkit without a board BMC and is jtag only. Without this change, boards were incorrectly identifying themselves, and then opae would in turn mishandle the misidentified board.

Could you extend the commit message with the case?

Closes: https://hsdes.intel.com/appstore/article/#/16023271460

[michael-adler]

So is the rule: set subdevice ID to 1 if there is no BMC and to 0x1771 if there is a BMC?

[fpgamatt]

There are currently quite a few subdevice ids being used. The intent is help identify the board with just lspci, and to address concerns from the kernel community of having just a generic DFL Device id with out any subdevice info. The subdevice of 0x1771 is actually a n6001, and 0x1770 is a n6000. 0x17d4 is used for C6100.

[coveralls]

Pull Request Test Coverage Report for Build 8711165719

Details

• 16 of 88 (18.18%) changed or added relevant lines in 3 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.02%) to 64.65%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
libraries/libboard/board_jtag_pci_dk/board_jtag_pci_dk.c	0	10	0.0%
libraries/libboard/board_common/board_common.c	15	77	19.48%

Totals
Change from base Build 8654404228:	-0.02%
Covered Lines:	15825
Relevant Lines:	24478

---

💛 - Coveralls

[fpgamatt]

can I get an approve so this can be merged into master. The RTL has been merged.

[michael-adler]

can I get an approve so this can be merged into master. The RTL has been merged.

I can't help there. Not a code owner.