[Bug 111461] [CI][BAT] igt at runner@aborted - fail - Previous test: nothing, Unclaimed read from register 0x61204

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=111461

--- Comment #8 from Matt Roper <matthew.d.roper at intel.com> ---
The unclaimed registers in question are 0x6120X, which are the old PPS
registers that were used on pre-PCH platforms.  When we access PPS in the code,
we use macros that take the old-style PPS register address (e.g., 0x61204),
subtract off PPS_BASE (0x61200), and then add the proper base for PPS registers
on the current platform (0xC7200 for platforms with a PCH).  See
intel_pps_init() for where the PPS base address is determined and the
definitions of PP_CONTROL and such for the address math.

The only reason we should be trying to use these old register addresses is if
we're failing to detect the PCH.  There's so much dmesg spam that we've lost
the beginning of the boot dmesg where the driver prints out the PCH type so
it's not very obvious from the log output here.  However CI's lspci output for
this CML machine shows a device ID of 0x0681 for device 31, function 0 rather
than the 0x0280 that we define INTEL_PCH_CMP_DEVICE_ID_TYPE with in the driver.

Digging through the rest of the lspci for CML machines in CI, the CML-U's do
indeed have PCH ID's of 0x0284, whereas the CML-S has 0x0681 and CML-H has
0x068C.  I believe we need to add a #define INTEL_PCH_CMP2_DEVICE_ID_TYPE
0x0680 to capture the ID's used on the S and H sku's.  I'll send along a patch
for this shortly.

Note that although PPS is what we see obviously blowing up here, failure to
detect the PCH would likely also lead to other problems with display hotplug,
backlight, and display audio; I think the "highest" priority for this defect is
warranted given those impacts.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20190916/0f598126/attachment.html>