The i82975x_edac driver is completely broken. I got a bug report
related to an OOPS caused by the recent changes there
After fixing it, I noticed weird behaviours at the driver on my
test machine, especially after enabling the EDAC debug logic.
In the end, the entire logic there that fills memory data were
broken, and also, the error report logic.
The worse bug there, IMHO, is that, during Corrected Errors,
it will only report the right DIMM in 50% of the cases, on
the more standard memory configuration (Dual channel, interleaved).
That happens because the logic were checking for bit 1, instead of
bit 6, in order to decide if the error belongs to channel A or to
channel B.
The patches on this series are:
- patch 1: fixes the reported OOPS - it is a good candidate
for Kernel 3.7 and -stable;
- patch 2: convert the already existing debug code
to use the standard way for debug (instead of requiring someone
to edit the source code to uncomment the debug macro);
So, the third patch on this series;
- patch 3: rewrite the entire fill/report logic.
Patches got tested on a Dell Precision N390 and worked as expected.
Regards,
Mauro
Mauro Carvalho Chehab (3):
i82975x_edac: Fix dimm label initialization
i82975x_edac: Use the edac standard debug macro
i82975x_edac: rewrite the entire fill/report logic
drivers/edac/i82975x_edac.c | 457 ++++++++++++++++++++++++++------------------
1 file changed, 272 insertions(+), 185 deletions(-)
--
1.7.11.7
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
