From 3212707a9ab549119ad0ad6441a07fe774c99f3b Mon Sep 17 00:00:00 2001 From: Kimplul Date: Thu, 27 Jul 2023 20:29:28 +0300 Subject: use nifty variable argument count trick + Found it in newlib. This allows the compiler to more effectively optimize away unnecessary register operations. Applied to both kernel and init.c for a slight speed increase. I really should move init.c to some other repository though, the binary files are starting to get annoying --- common/uapi/ipc.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) (limited to 'common/uapi/ipc.c') diff --git a/common/uapi/ipc.c b/common/uapi/ipc.c index c2548e9..f9b2817 100644 --- a/common/uapi/ipc.c +++ b/common/uapi/ipc.c @@ -134,7 +134,7 @@ static struct stack_diff enter_rpc(struct tcb *t, struct sys_ret a, /* try to get rid of args as fast as possible to free up registers for * later use */ - set_args(t, a); + set_args(t, 6, a); ctx->exec = t->exec; ctx->pid = t->pid; @@ -185,7 +185,7 @@ static void leave_rpc(struct tcb *t, struct sys_ret a) t->regs = ctx->regs; /* again, get rid of args as fast as possible */ - set_args(t, a); + set_args(t, 6, a); set_return(t, ctx->exec); /* if we're returning from a failed rpc, this should essentially be a @@ -230,7 +230,7 @@ static bool enough_rpc_stack(struct tcb *t) SYSCALL_DEFINE1(ipc_server)(struct tcb *t, sys_arg_t callback) { get_cproc(t)->callback = callback; - return_args(t, SYS_RET1(OK)); + return_args1(t, OK); } /** @@ -261,7 +261,7 @@ static void do_ipc(struct tcb *t, enum ipc_kind kind) { if (unlikely(!enough_rpc_stack(t))) - return_args(t, SYS_RET1(ERR_OOMEM)); + return_args1(t, ERR_OOMEM); struct stack_diff sd = enter_rpc(t, SYS_RET6(OK, t->eid, d0, d1, d2, d3), kind); @@ -283,6 +283,9 @@ static void do_ipc(struct tcb *t, t->eid = t->pid; finalize_rpc(t, r, sd); + /* I tested out passing the return values as arguments to + * ret_userspace_fast, but apparently that causes enough stack shuffling + * to be slower overall. */ ret_userspace_fast(); } /** @@ -354,7 +357,7 @@ SYSCALL_DEFINE4(ipc_resp)(struct tcb *t, sys_arg_t d0, sys_arg_t d1, /* if we're not in an rpc, the user messed something up. */ /** @todo choose or come up with more fitting error value. */ if (unlikely(!is_rpc(t))) - return_args(t, SYS_RET1(ERR_MISC)); + return_args1(t, ERR_MISC); /* we need the current proc before leaving the rpc */ struct tcb *r = get_cproc(t); @@ -373,20 +376,20 @@ SYSCALL_DEFINE4(ipc_resp)(struct tcb *t, sys_arg_t d0, sys_arg_t d1, */ SYSCALL_DEFINE1(ipc_notify)(struct tcb *t, sys_arg_t tid){ if (!has_cap(t->caps, CAP_CALL)) - return_args(t, SYS_RET1(ERR_PERM)); + return_args1(t, ERR_PERM); struct tcb *r = get_tcb(tid); if (r->notify_state == NOTIFY_QUEUED) - return_args(t, SYS_RET1(OK)); + return_args1(t, OK); if (r->notify_state == NOTIFY_RUNNING) { t->notify_state = NOTIFY_QUEUED; - return_args(t, SYS_RET1(OK)); + return_args1(t, OK); } r->notify_state = NOTIFY_QUEUED; if (running(r)) send_ipi(r); - return_args(t, SYS_RET1(OK)); + return_args1(t, OK); } -- cgit v1.3