Itanium SI status update
My most recent full test run, last Wed., showed
52 failures on my Itanium LSB-si.
Of those I've addressed 27, explained what's
happening in another 12, and have 13 remaining
that I don't fully understand the reasons for the
failure.
"Addressed" is a pretty loose categorization
including manually removing a file that had been
left lying around which blocked one test from running
(this could perhaps be considered a test bug),
updating a couple of config file values, running
through telnet instead of "chroot ; login", and -
for most - changes I've made to the tests to
address what I think are test suite bugs.
Note that the test changes are *not* official,
there are a small handful of PRs filed against
these awaiting disposition; but I *have* been able
to compile and pass the tests with those fixes applied.
Of the 12 "explained" but not fixed, three require
determining one value for the config file, which I've
asked about on a different list, but not yet gotten
a value for (VSRT_ADDR_SPACE_PAGES). Another six
require a single fix to allow for timer resolutions
that are not a power-of-ten HZ - the PR on that one
has been recommended to accept. Two more require
resolution of picking up the wrong value of SHMLBA
(which is 1meg on Itanium, but that's not reflected
in the header files). The final one requires resolution
of why SSIZE_MAX on the system is defined as if ssize_t
was an "int" (32-bit on LP64 machines such as Itanium),
even though on a 64-bit platform it's a 64-bit quantity.
Mats
Reply to: