Fix a longstanding "freeze right after enabling interrupt" problem.

With this fix, finally NetBSD/emips on Giano is fully functional.
See PR/45080 for more details.

Should be pulled up to netbsd-9.
This commit is contained in:
tsutsui 2019-12-11 16:16:13 +00:00
parent b2a9c378fb
commit 7bc4024944

View File

@ -1,4 +1,4 @@
/* $NetBSD: interrupt.c,v 1.7 2019/12/09 16:19:11 tsutsui Exp $ */
/* $NetBSD: interrupt.c,v 1.8 2019/12/11 16:16:13 tsutsui Exp $ */
/*-
* Copyright (c) 2010 The NetBSD Foundation, Inc.
@ -31,7 +31,7 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: interrupt.c,v 1.7 2019/12/09 16:19:11 tsutsui Exp $");
__KERNEL_RCSID(0, "$NetBSD: interrupt.c,v 1.8 2019/12/11 16:16:13 tsutsui Exp $");
#include <sys/param.h>
#include <sys/device.h>
@ -102,6 +102,19 @@ cpu_intr(int ppl, vaddr_t pc, uint32_t status)
curcpu()->ci_data.cpu_nintr++;
#if 0
/*
* According to Giano simulator sources (Cpus/mips_cpu.cpp),
* interrupt register bits in CAUSE register are updated
* only when the exception is triggered. This means checking
* CAUSE register via splintr() in a while loop in this
* interrupt handler doesn't work as expected on Giano.
*
* I don't know whether the real FPGA eMIPS has the same
* design as the Giano simulator, but for now I'd like to
* choose 'call only one handler per each interrupt' strategy,
* as the original NetBSD/emips implementation.
*/
while (ppl < (ipl = splintr(&ipending))) {
splx(ipl);
/* device interrupts */
@ -110,6 +123,14 @@ cpu_intr(int ppl, vaddr_t pc, uint32_t status)
}
(void)splhigh();
}
#else
ipl = splintr(&ipending);
__USE(ipl);
/* device interrupts */
if (ipending & MIPS_INT_MASK_5) {
(*platform.iointr)(status, pc, ipending);
}
#endif
}
/*