ultra.s

这个linux源代码是很全面的~基本完整了~使用c编译的~由于时间问题我没有亲自测试~但就算用来做参考资料也是非常好的

	mov		PRIMARY_CONTEXT, %o2
	ldxa		[%o2] ASI_DMMU, %g2
	stxa		%o0, [%o2] ASI_DMMU
	stxa		%g0, [%o1] ASI_DMMU_DEMAP
	stxa		%g0, [%o1] ASI_IMMU_DEMAP
	stxa		%g2, [%o2] ASI_DMMU
	flush		%g6
	wrpr		%g0, 0, %tl
	retl
	 wrpr		%g5, 0x0, %pstate

__cheetah_flush_tlb_mm: /* 15 insns */
	rdpr		%pstate, %g5
	andn		%g5, PSTATE_IE, %g2
	wrpr		%g2, 0x0, %pstate
	wrpr		%g0, 1, %tl
	mov		PRIMARY_CONTEXT, %o2
	mov		0x40, %g3
	ldxa		[%o2] ASI_DMMU, %g2
	stxa		%o0, [%o2] ASI_DMMU
	stxa		%g0, [%g3] ASI_DMMU_DEMAP
	stxa		%g0, [%g3] ASI_IMMU_DEMAP
	stxa		%g2, [%o2] ASI_DMMU
	flush		%g6
	wrpr		%g0, 0, %tl
	retl
	 wrpr		%g5, 0x0, %pstate

__cheetah_flush_tlb_range: /* 20 insns */
	cmp		%o5, %o4
	blu,pt		%xcc, 9f
	 rdpr		%pstate, %g5
	andn		%g5, PSTATE_IE, %g2
	wrpr		%g2, 0x0, %pstate
	wrpr		%g0, 1, %tl
	mov		PRIMARY_CONTEXT, %o2
	sub		%o5, %o4, %o5
	ldxa		[%o2] ASI_DMMU, %g2
	stxa		%o0, [%o2] ASI_DMMU
1:	stxa		%g0, [%o1 + %o5] ASI_DMMU_DEMAP
	stxa		%g0, [%o1 + %o5] ASI_IMMU_DEMAP
	membar		#Sync
	brnz,pt		%o5, 1b
	 sub		%o5, %o4, %o5
	stxa		%g2, [%o2] ASI_DMMU
	flush		%g6
	wrpr		%g0, 0, %tl
9:	retl
	 wrpr		%g5, 0x0, %pstate

flush_dcpage_cheetah: /* 9 insns */
	sub		%o0, %g4, %o0
	sethi		%hi(PAGE_SIZE), %o4
1:	subcc		%o4, (1 << 5), %o4
	stxa		%g0, [%o0 + %o4] ASI_DCACHE_INVALIDATE
	membar		#Sync
	bne,pt		%icc, 1b
	 nop
	retl		/* I-cache flush never needed on Cheetah, see callers. */
	 nop

cheetah_patch_one:
1:	lduw		[%o1], %g1
	stw		%g1, [%o0]
	flush		%o0
	subcc		%o2, 1, %o2
	add		%o1, 4, %o1
	bne,pt		%icc, 1b
	 add		%o0, 4, %o0
	retl
	 nop

	.globl		cheetah_patch_cachetlbops
cheetah_patch_cachetlbops:
	save		%sp, -128, %sp

	sethi		%hi(__flush_tlb_page), %o0
	or		%o0, %lo(__flush_tlb_page), %o0
	sethi		%hi(__cheetah_flush_tlb_page), %o1
	or		%o1, %lo(__cheetah_flush_tlb_page), %o1
	call		cheetah_patch_one
	 mov		14, %o2

	sethi		%hi(__flush_tlb_mm), %o0
	or		%o0, %lo(__flush_tlb_mm), %o0
	sethi		%hi(__cheetah_flush_tlb_mm), %o1
	or		%o1, %lo(__cheetah_flush_tlb_mm), %o1
	call		cheetah_patch_one
	 mov		15, %o2

	sethi		%hi(__flush_tlb_range), %o0
	or		%o0, %lo(__flush_tlb_range), %o0
	sethi		%hi(__cheetah_flush_tlb_range), %o1
	or		%o1, %lo(__cheetah_flush_tlb_range), %o1
	call		cheetah_patch_one
	 mov		20, %o2

	sethi		%hi(__flush_dcache_page), %o0
	or		%o0, %lo(__flush_dcache_page), %o0
	sethi		%hi(flush_dcpage_cheetah), %o1
	or		%o1, %lo(flush_dcpage_cheetah), %o1
	call		cheetah_patch_one
	 mov		9, %o2

	ret
	 restore

#ifdef CONFIG_SMP
	/* These are all called by the slaves of a cross call, at
	 * trap level 1, with interrupts fully disabled.
	 *
	 * Register usage:
	 *   %g5	mm->context	(all tlb flushes)
	 *   %g1	address arg 1	(tlb page and range flushes)
	 *   %g7	address arg 2	(tlb range flush only)
	 *
	 *   %g6	ivector table, don't touch
	 *   %g2	scratch 1
	 *   %g3	scratch 2
	 *   %g4	scratch 3
	 *
	 * TODO: Make xcall TLB range flushes use the tricks above... -DaveM
	 */
	.align		32
	.globl		xcall_flush_tlb_page, xcall_flush_tlb_mm, xcall_flush_tlb_range
xcall_flush_tlb_page:
	mov		PRIMARY_CONTEXT, %g2
	ldxa		[%g2] ASI_DMMU, %g3
	stxa		%g5, [%g2] ASI_DMMU
	stxa		%g0, [%g1] ASI_DMMU_DEMAP
	stxa		%g0, [%g1] ASI_IMMU_DEMAP
	stxa		%g3, [%g2] ASI_DMMU
	retry
	nop

xcall_flush_tlb_mm:
	mov		PRIMARY_CONTEXT, %g2
	mov		0x40, %g4
	ldxa		[%g2] ASI_DMMU, %g3
	stxa		%g5, [%g2] ASI_DMMU
	stxa		%g0, [%g4] ASI_DMMU_DEMAP
	stxa		%g0, [%g4] ASI_IMMU_DEMAP
	stxa		%g3, [%g2] ASI_DMMU
	retry

xcall_flush_tlb_range:
	sethi		%hi(PAGE_SIZE - 1), %g2
	or		%g2, %lo(PAGE_SIZE - 1), %g2
	andn		%g1, %g2, %g1
	andn		%g7, %g2, %g7
	sub		%g7, %g1, %g3
	add		%g2, 1, %g2
	srlx		%g3, PAGE_SHIFT, %g4
	cmp		%g4, 96

	bgu,pn		%icc, xcall_flush_tlb_mm
	 mov		PRIMARY_CONTEXT, %g4
	ldxa		[%g4] ASI_DMMU, %g7
	sub		%g3, %g2, %g3
	stxa		%g5, [%g4] ASI_DMMU
	nop
	nop
	nop

1:	stxa		%g0, [%g1 + %g3] ASI_DMMU_DEMAP
	stxa		%g0, [%g1 + %g3] ASI_IMMU_DEMAP
	membar		#Sync
	brnz,pt		%g3, 1b
	 sub		%g3, %g2, %g3
	stxa		%g7, [%g4] ASI_DMMU
	retry
	nop
	nop

	/* NOTE: This is SPECIAL!!  We do etrap/rtrap however
	 *       we choose to deal with the "BH's run with
	 *       %pil==15" problem (described in asm/pil.h)
	 *       by just invoking rtrap directly past where
	 *       BH's are checked for.
	 *
	 *       We do it like this because we do not want %pil==15
	 *       lockups to prevent regs being reported.
	 */
	.globl		xcall_report_regs
xcall_report_regs:
	rdpr		%pstate, %g2
	wrpr		%g2, PSTATE_IG | PSTATE_AG, %pstate
	rdpr		%pil, %g2
	wrpr		%g0, 15, %pil
	sethi		%hi(109f), %g7
	b,pt		%xcc, etrap_irq
109:	 or		%g7, %lo(109b), %g7
	call		__show_regs
	 add		%sp, STACK_BIAS + REGWIN_SZ, %o0
	clr		%l6
	/* Has to be a non-v9 branch due to the large distance. */
	b		rtrap_xcall
	 ldx		[%sp + PTREGS_OFF + PT_V9_TSTATE], %l1

	.align		32
	.globl		xcall_flush_dcache_page_cheetah
xcall_flush_dcache_page_cheetah: /* %g1 == physical page address */
	sethi		%hi(PAGE_SIZE), %g3
1:	subcc		%g3, (1 << 5), %g3
	stxa		%g0, [%g1 + %g3] ASI_DCACHE_INVALIDATE
	membar		#Sync
	bne,pt		%icc, 1b
	 nop
	retry
	nop

	.globl		xcall_flush_dcache_page_spitfire
xcall_flush_dcache_page_spitfire: /* %g1 == physical page address
				     %g7 == kernel page virtual address
				     %g5 == (page->mapping != NULL)  */
#if (L1DCACHE_SIZE > PAGE_SIZE)
	srlx		%g1, (13 - 2), %g1	! Form tag comparitor
	sethi		%hi(L1DCACHE_SIZE), %g3	! D$ size == 16K
	sub		%g3, (1 << 5), %g3	! D$ linesize == 32
1:	ldxa		[%g3] ASI_DCACHE_TAG, %g2
	andcc		%g2, 0x3, %g0
	be,pn		%xcc, 2f
	 andn		%g2, 0x3, %g2
	cmp		%g2, %g1

	bne,pt		%xcc, 2f
	 nop
	stxa		%g0, [%g3] ASI_DCACHE_TAG
	membar		#Sync
2:	cmp		%g3, 0
	bne,pt		%xcc, 1b
	 sub		%g3, (1 << 5), %g3

	brz,pn		%g5, 2f
#endif /* L1DCACHE_SIZE > PAGE_SIZE */
	 sethi		%hi(PAGE_SIZE), %g3

1:	flush		%g7
	subcc		%g3, (1 << 5), %g3
	bne,pt		%icc, 1b
	 add		%g7, (1 << 5), %g7

2:	retry
	nop
	nop

	.globl		xcall_promstop
xcall_promstop:
	rdpr		%pstate, %g2
	wrpr		%g2, PSTATE_IG | PSTATE_AG, %pstate
	rdpr		%pil, %g2
	wrpr		%g0, 15, %pil
	sethi		%hi(109f), %g7
	b,pt		%xcc, etrap_irq
109:	 or		%g7, %lo(109b), %g7
	flushw
	call		prom_stopself
	 nop
	/* We should not return, just spin if we do... */
1:	b,a,pt		%xcc, 1b
	nop

	.data

errata32_hwbug:
	.xword	0

	.text

	/* These two are not performance critical... */
	.globl		xcall_flush_tlb_all_spitfire
xcall_flush_tlb_all_spitfire:
	/* Spitfire Errata #32 workaround. */
	sethi		%hi(errata32_hwbug), %g4
	stx		%g0, [%g4 + %lo(errata32_hwbug)]

	clr		%g2
	clr		%g3
1:	ldxa		[%g3] ASI_DTLB_DATA_ACCESS, %g4
	and		%g4, _PAGE_L, %g5
	brnz,pn		%g5, 2f
	 mov		TLB_TAG_ACCESS, %g7

	stxa		%g0, [%g7] ASI_DMMU
	membar		#Sync
	stxa		%g0, [%g3] ASI_DTLB_DATA_ACCESS
	membar		#Sync

	/* Spitfire Errata #32 workaround. */
	sethi		%hi(errata32_hwbug), %g4
	stx		%g0, [%g4 + %lo(errata32_hwbug)]

2:	ldxa		[%g3] ASI_ITLB_DATA_ACCESS, %g4
	and		%g4, _PAGE_L, %g5
	brnz,pn		%g5, 2f
	 mov		TLB_TAG_ACCESS, %g7

	stxa		%g0, [%g7] ASI_IMMU
	membar		#Sync
	stxa		%g0, [%g3] ASI_ITLB_DATA_ACCESS
	membar		#Sync

	/* Spitfire Errata #32 workaround. */
	sethi		%hi(errata32_hwbug), %g4
	stx		%g0, [%g4 + %lo(errata32_hwbug)]

2:	add		%g2, 1, %g2
	cmp		%g2, SPITFIRE_HIGHEST_LOCKED_TLBENT
	ble,pt		%icc, 1b
	 sll		%g2, 3, %g3
	flush		%g6
	retry

	.globl		xcall_flush_tlb_all_cheetah
xcall_flush_tlb_all_cheetah:
	mov		0x80, %g2
	stxa		%g0, [%g2] ASI_DMMU_DEMAP
	stxa		%g0, [%g2] ASI_IMMU_DEMAP
	retry

	.globl		xcall_flush_cache_all_spitfire
xcall_flush_cache_all_spitfire:
	sethi		%hi(16383), %g2
	or		%g2, %lo(16383), %g2
	clr		%g3
1:	stxa		%g0, [%g3] ASI_IC_TAG
	membar		#Sync
	add		%g3, 32, %g3
	cmp		%g3, %g2
	bleu,pt		%xcc, 1b
	 nop
	flush		%g6
	retry

	/* These just get rescheduled to PIL vectors. */
	.globl		xcall_call_function
xcall_call_function:
	wr		%g0, (1 << PIL_SMP_CALL_FUNC), %set_softint
	retry

	.globl		xcall_receive_signal
xcall_receive_signal:
	wr		%g0, (1 << PIL_SMP_RECEIVE_SIGNAL), %set_softint
	retry

	.globl		xcall_capture
xcall_capture:
	wr		%g0, (1 << PIL_SMP_CAPTURE), %set_softint
	retry

#endif /* CONFIG_SMP */