Date: Thu, 17 Sep 1998 15:36:08 -0400
From: Andrew Heybey <ath@niksun.com>
To: freebsd-hackers@FreeBSD.ORG
Subject: masking of hardware interrupts in {,icu_}vector.s
Message-ID: <199809171936.PAA00480@stiegl.niksun.com>
next in thread | raw e-mail | index | archive | help
[I'm working with -STABLE though the vectors in -current look unchanged.]
The interrupt vectors in sys/i386/isa/vector.s (icu_vector.s in -current) mask
the current interrupt in the hardware while the interrupt handler runs. Is
this done for any reason besides efficiency? For example, say I wanted to
change the interrupt vector to count hardware interrupts even if the handler
was already executing. If I dike out the parts of the INTR macro in vector.s
as follows:
#if 0
movb _imen + IRQ_BYTE(irq_num),%al ; \
orb $IRQ_BIT(irq_num),%al ; \
movb %al,_imen + IRQ_BYTE(irq_num) ; \
outb %al,$icu+ICU_IMR_OFFSET ; \
#endif
enable_icus ; \
/* Following line is new code to count HW interrupts. */ \
incl _interrupt_count + (irq_num) * 4 ; /* Count this HW intr */ \
movl _cpl,%eax ; \
testb $IRQ_BIT(irq_num),%reg ; \
jne 2f ; \
incb _intr_nesting_level ; \
__CONCAT(Xresume,irq_num): ; \
FAKE_MCOUNT(12*4(%esp)) ; /* XXX late to avoid double count */ \
incl _cnt+V_INTR ; /* tally interrupts */ \
movl _intr_countp + (irq_num) * 4,%eax ; \
incl (%eax) ; \
movl _cpl,%eax ; \
pushl %eax ; \
pushl _intr_unit + (irq_num) * 4 ; \
orl _intr_mask + (irq_num) * 4,%eax ; \
movl %eax,_cpl ; \
sti ; \
call *_intr_handler + (irq_num) * 4 ; \
cli ; /* must unmask _imen and icu atomically */ \
#if 0
movb _imen + IRQ_BYTE(irq_num),%al ; \
andb $~IRQ_BIT(irq_num),%al ; \
movb %al,_imen + IRQ_BYTE(irq_num) ; \
outb %al,$icu+ICU_IMR_OFFSET ; \
#endif
What I think this means is that if the same interrupt happens again, then the
handler won't run (because the CPL is set too high), but the vector will run
and count the extra interrupt.
Is there anything I'm going to break by doing this? I am going to try it to
see what happens, but I thought I would ask just in case there is magic about
the ICU and/or devices that I don't understand that would make this fail.
thanks,
andrew
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199809171936.PAA00480>