From 8100eef0826e7fb182434979a9b139716b3637e1 Mon Sep 17 00:00:00 2001 From: Kimplul Date: Fri, 1 Nov 2024 10:12:02 +0200 Subject: speed up creating threads quite a bit --- TODO | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'TODO') diff --git a/TODO b/TODO index 73b8040..c80e9b0 100644 --- a/TODO +++ b/TODO @@ -51,7 +51,6 @@ associated program. Easy. (started) + Start working on proper context swaps and RPC swaps. (which thread ID should be visible from cur_tcb() when in an RPC? should I use some kind of effective tid instead of cur_tcb()?) -+ Export VM flags to userspace, separate header or something? + Get core count? It's technically speaking available from the fdt but could be limited by the NR_CPUS macro, so it might be a good idea to export as a CONF_* + The node subsystem might benefit from larger-than-base-page sizes, for example @@ -62,6 +61,9 @@ limited by the NR_CPUS macro, so it might be a good idea to export as a CONF_* just check if the last message is OK, and we only really need more complex checking if we need to calculate like memory leaks etc. ++ grow_rpc() shouldn't exceed limits ++ handle_pagefault() should kill processes that have no good reason to pagefault + !!! FUTURE: + Cache locality? + When mapping some other address space into your own, I could keep a maximum -- cgit v1.3