Upstream lwaftr ljsyscall changes #203
Conversation
Calling get_mempolicy() will return an object with "mode" and "mask" keys, corresponding to the two output arguments of the get_mempolicy function. The mask is implemented along the lines of cpu_set, except that it can hold any number of bits, defaulting to the size of a long. set_mempolicy(mode, mask) imposes a mode and possibly a mask as well.
* lib/ljsyscall/syscall/linux/c.lua: * lib/ljsyscall/syscall/linux/syscalls.lua: Add support for the migrate_pages Linux syscall.
|
Thanks! I think the CI is failing because of LuaJIT 2.0 vs 2.1 differences. I don't really want to break 2.0 yet, even though most people use 2.1. I think it should be possible to work around the differences, will take a look. |
| local ulong_bit_count = ffi.sizeof('unsigned long') * 8 | ||
| local function ulong_index_and_bit(n) | ||
| local i = math.floor(n / ulong_bit_count) | ||
| local b = bit.lshift(1ULL, n - i * ulong_bit_count) |
There was a problem hiding this comment.
Need to use 1 instead of 1ULL so that luajit 2.0 doesn't fail
|
That won't give the right result though... On 28 Oct 2016 8:15 p.m., "daurnimator" [email protected] wrote:
|
|
This has significant (but not total) overlap with #193. The numa/mempolicy code is the same, but this one also includes the migrate_pages patch. |
|
@kbara its not just the syntax, 2.0 does not have 64 bit shifts. There are some compatibility functions for 64 bit shifts here https://github.com/justincormack/ljsyscall/blob/master/syscall/bit.lua |
|
Yeah - that was a hastily-written patch based on some Jabber discussion, which I should revert. Sorry for the noise. |
kbara
force-pushed
the
upstream-lwaftr-ljsyscall-changes
branch
from
0b6c6ec to
7122cf4
Compare
Hi @justincormack ,
@wingo wrote a couple of patches to ljsyscall, which we've been using in the lwaftr. As per discussion with him, here's a PR upstreaming them. They deal with mempolicy/numa/page migration.