First draft of macOS + BSD plugins #81
Conversation
It makes it very easy to have 2 implementation of plugins. Only the correct one gets compiled and linked to the global executable.
|
awesome 👌 i need to fix the tests suite it seems 😅 |
|
And, thanks for keeping my |
| printf("speculative.value %lu\n", stats.speculative_count * pagesize); | ||
|
|
||
| int64_t freePages = stats.free_count - stats.speculative_count; | ||
| printf("other.value %lu\n", MAX(0, (int64_t)totalPages - stats.wire_count - stats.active_count - stats.inactive_count - freePages) * pagesize); |
There was a problem hiding this comment.
Do not hesitate to use temporary named variables.
They get optimized away and help understanding and debugging: 😉
There was a problem hiding this comment.
Oh, yes, definitely. I think the problem here is that I don't understand how these numbers are supposed to add up, and I haven't been able to find complete documentation on them. I found https://github.com/apple/darwin-xnu/blob/2ff845c2e033bd0ff64b5b6aa6063a1f8f65aa32/osfmk/mach/vm_statistics.h#L108-L111 which says:
speculative pages are already accounted for in "free_count", so "speculative_count" is the number of "free" pages that are used to hold data that was read speculatively from disk but haven't actually been used by anyone so far.
…so I've defined freePages = stats.free_count - stats.speculative_count.
But often the values in vm_statistics don't sum to exactly vm.pages — sometimes the sum is less than vm.pages (represented by the other.value I added); sometimes the sum is more than vm.pages (thus the need for MAX(0, …) to keep it from wrapping).
Here's what the latest revision (7b2914a) looks like for an 8 GB system:
@steveschnepp, I've made some progress on your work in PR #47:
masterbranch