On 17.11.2022 17:41, Alexander Motin wrote:
On 17.11.2022 17:07, Alison Schofield wrote:
On Thu, Nov 17, 2022 at 04:09:36PM -0500, Alexander Motin wrote:
Previous code reported "smart" state based on number of bits
set in the module health field. But actually any single bit
set there already means critical failure. Rework the logic
according to specification, properly reporting non-critical
state in case of warning threshold reached, critical in case
of any module health bit set or error threshold reached and
fatal if NVDIMM exhausted its life time. In attempt to
report the cause of failure in absence of better methods,
report reached thresholds as more or less matching alarms.
While there clean up the code, making it more uniform with
others and allowing more methods to be implemented later.
Hi Alexander,
Perhaps this would be better presented in 2 patches:
1)the cleanup and then 2) improvement of smart state reporting.
Done.
Pardon my ignorance about the processes here, but what's about review on
on the merits after two weeks? Any more thoughts after I fixed the
"procedural issue":
https://lore.kernel.org/nvdimm/[email protected]/T/#t
Thanks.
--
Alexander Motin
