SPARC32: First pass at getting this platform functional under 2.5.x

arch/sparc/Makefile

@@ -41,15 +41,16 @@ SUBDIRS += arch/sparc/kernel arch/sparc/lib arch/sparc/prom \
 CORE_FILES := arch/sparc/kernel/kernel.o arch/sparc/mm/mm.o $(CORE_FILES) \
 	arch/sparc/math-emu/math-emu.o
 
-LIBS := $(TOPDIR)/lib/lib.a $(LIBS) $(TOPDIR)/arch/sparc/prom/promlib.a \
-	$(TOPDIR)/arch/sparc/lib/lib.a
+LIBS := $(LIBS) arch/sparc/prom/promlib.a arch/sparc/lib/lib.a
 
 # This one has to come last
 SUBDIRS += arch/sparc/boot
 CORE_FILES_NO_BTFIX := $(CORE_FILES)
 CORE_FILES += arch/sparc/boot/btfix.o
 
+# Export what is needed by arch/sparc/boot/Makefile
 export	CORE_FILES_NO_BTFIX
+export INIT
 
 archclean:
 	rm -f $(TOPDIR)/vmlinux.aout
@@ -60,7 +61,7 @@ archmrproper:
 
 prepare: check_asm
 
-check_asm: include/linux/version.h include/linux/asm include/config/MARKER
+check_asm: include/linux/version.h include/asm include/config/MARKER
 	$(MAKE) -C arch/sparc/kernel check_asm
 
 tftpboot.img:

arch/sparc/boot/Makefile

@@ -22,9 +22,12 @@ btfixupprep: btfixupprep.c
 clean:
 	rm -f btfixupprep piggyback tftpboot.img btfix.o btfix.s
 
-BTOBJS := $(HEAD) init/main.o init/version.o
-BTLIBS := $(CORE_FILES_NO_BTFIX) $(FILESYSTEMS) \
-	$(DRIVERS) $(NETWORKS)
+#BTOBJS := $(HEAD) init/main.o init/version.o
+BTOBJS := $(HEAD) $(INIT)
+#BTLIBS := $(CORE_FILES_NO_BTFIX) $(FILESYSTEMS) \
+#	$(DRIVERS) $(NETWORKS)
+# Threw away drivers because they must not have btfixup definitions.
+BTLIBS := $(CORE_FILES_NO_BTFIX) $(LIBS)
 
 # I wanted to make this depend upon BTOBJS so that a parallel
 # build would work, but this fails because $(HEAD) cannot work
@@ -34,7 +37,6 @@ vmlinux.o: FORCE
 	$(LD) $(LDFLAGS) -r $(patsubst %,$(TOPDIR)/%,$(BTOBJS)) \
 		--start-group \
 		$(patsubst %,$(TOPDIR)/%,$(BTLIBS)) \
-		$(LIBS) \
 		--end-group -o vmlinux.o
 
 btfix.s: btfixupprep vmlinux.o

arch/sparc/config.in

@@ -239,7 +239,10 @@ endmenu
 mainmenu_option next_comment
 comment 'Kernel hacking'
 
+bool 'Debug memory allocations' CONFIG_DEBUG_SLAB
 bool 'Magic SysRq key' CONFIG_MAGIC_SYSRQ
+bool 'Spinlock debugging' CONFIG_DEBUG_SPINLOCK
+
 endmenu
 
 source lib/Config.in

arch/sparc/kernel/entry.S

@@ -31,6 +31,7 @@
 #include <asm/signal.h>
 #include <asm/obio.h>
 #include <asm/mxcc.h>
+#include <asm/thread_info.h>
 
 #include <asm/asmmacro.h>
 
@@ -1234,7 +1235,8 @@ C_LABEL(sys_ptrace):
 	call	C_LABEL(do_ptrace)
 	 add	%sp, REGWIN_SZ, %o0
 
-	ld	[%curptr + AOFF_task_ptrace], %l5
+	ld	[%curptr + TI_TASK], %l5
+	ld	[%l5 + AOFF_task_ptrace], %l5
 	andcc	%l5, 0x02, %g0
 	be	1f
 	 nop
@@ -1284,7 +1286,8 @@ C_LABEL(sys_sigpause):
 	call	C_LABEL(do_sigpause)
 	 add	%sp, REGWIN_SZ, %o1
 
-	ld	[%curptr + AOFF_task_ptrace], %l5
+	ld	[%curptr + TI_TASK], %l5
+	ld	[%l5 + AOFF_task_ptrace], %l5
 	andcc	%l5, 0x02, %g0
 	be	1f
 	 nop
@@ -1302,7 +1305,8 @@ C_LABEL(sys_sigsuspend):
 	call	C_LABEL(do_sigsuspend)
 	 add	%sp, REGWIN_SZ, %o0
 
-	ld	[%curptr + AOFF_task_ptrace], %l5
+	ld	[%curptr + TI_TASK], %l5
+	ld	[%l5 + AOFF_task_ptrace], %l5
 	andcc	%l5, 0x02, %g0
 	be	1f
 	 nop
@@ -1321,7 +1325,8 @@ C_LABEL(sys_rt_sigsuspend):
 	call	C_LABEL(do_rt_sigsuspend)
 	 add	%sp, REGWIN_SZ, %o2
 
-	ld	[%curptr + AOFF_task_ptrace], %l5
+	ld	[%curptr + TI_TASK], %l5
+	ld	[%l5 + AOFF_task_ptrace], %l5
 	andcc	%l5, 0x02, %g0
 	be	1f
 	 nop
@@ -1339,7 +1344,8 @@ C_LABEL(sys_sigreturn):
 	call	C_LABEL(do_sigreturn)
 	 add	%sp, REGWIN_SZ, %o0
 
-	ld	[%curptr + AOFF_task_ptrace], %l5
+	ld	[%curptr + TI_TASK], %l5
+	ld	[%l5 + AOFF_task_ptrace], %l5
 	andcc	%l5, 0x02, %g0
 	be	1f
 	 nop
@@ -1359,7 +1365,8 @@ C_LABEL(sys_rt_sigreturn):
 	call	C_LABEL(do_rt_sigreturn)
 	 add	%sp, REGWIN_SZ, %o0
 
-	ld	[%curptr + AOFF_task_ptrace], %l5
+	ld	[%curptr + TI_TASK], %l5
+	ld	[%l5 + AOFF_task_ptrace], %l5
 	andcc	%l5, 0x02, %g0
 	be	1f
 	 nop
@@ -1384,16 +1391,17 @@ C_LABEL(sys_fork):
 	mov	%o7, %l5
 flush_patch_two:
 	FLUSH_ALL_KERNEL_WINDOWS;
+	ld	[%curptr + TI_TASK], %o4
 	rd	%psr, %g4
 	WRITE_PAUSE
 	mov	SIGCHLD, %o0			! arg0:	clone flags
 	rd	%wim, %g5
 	WRITE_PAUSE
 	mov	%fp, %o1			! arg1:	usp
-	std	%g4, [%curptr + AOFF_task_thread + AOFF_thread_fork_kpsr]
+	std	%g4, [%o4 + AOFF_task_thread + AOFF_thread_fork_kpsr]
 	add	%sp, REGWIN_SZ, %o2		! arg2:	pt_regs ptr
 	mov	0, %o3
-	call	C_LABEL(do_fork_FIXME_NOW_RETURNS_TASK_STRUCT)
+	call	C_LABEL(sparc_do_fork)
 	 mov	%l5, %o7
 
 	/* Whee, kernel threads! */
@@ -1402,6 +1410,7 @@ C_LABEL(sys_clone):
 	mov	%o7, %l5
 flush_patch_three:
 	FLUSH_ALL_KERNEL_WINDOWS;
+	ld	[%curptr + TI_TASK], %o4
 	rd	%psr, %g4
 	WRITE_PAUSE
 
@@ -1413,11 +1422,10 @@ flush_patch_three:
 	 mov	%fp, %o1			! yes, use callers usp
 	andn	%o1, 7, %o1			! no, align to 8 bytes
 1:
-	std	%g4, [%curptr + AOFF_task_thread + AOFF_thread_fork_kpsr]
+	std	%g4, [%o4 + AOFF_task_thread + AOFF_thread_fork_kpsr]
 	add	%sp, REGWIN_SZ, %o2		! arg2:	pt_regs ptr
 	mov	0, %o3
-	/* FIXME:	remove CLONE_IDLETASK from flags first */
-	call	C_LABEL(do_fork_WITHOUT_CLONE_IDLETASK)
+	call	C_LABEL(sparc_do_fork)
 	 mov	%l5, %o7
 
 	/* Whee, real vfork! */
@@ -1425,17 +1433,18 @@ flush_patch_three:
 C_LABEL(sys_vfork):
 flush_patch_four:
 	FLUSH_ALL_KERNEL_WINDOWS;
+	ld	[%curptr + TI_TASK], %o4
 	rd	%psr, %g4
 	WRITE_PAUSE
 	rd	%wim, %g5
 	WRITE_PAUSE
-	std	%g4, [%curptr + AOFF_task_thread + AOFF_thread_fork_kpsr]
+	std	%g4, [%o4 + AOFF_task_thread + AOFF_thread_fork_kpsr]
 	sethi	%hi(0x4000 | 0x0100 | SIGCHLD), %o0
 	mov	%fp, %o1
 	or	%o0, %lo(0x4000 | 0x0100 | SIGCHLD), %o0
-	sethi	%hi(C_LABEL(do_fork_FIXME_NOW_RETURNS_TASK_STRUCT)), %l1
+	sethi	%hi(C_LABEL(sparc_do_fork)), %l1
 	mov	0, %o3
-	jmpl	%l1 + %lo(C_LABEL(do_fork_FIXME_NOW_RETURNS_TASK_STRUCT)), %g0
+	jmpl	%l1 + %lo(C_LABEL(sparc_do_fork)), %g0
 	 add	%sp, REGWIN_SZ, %o2
 
         .align  4
@@ -1464,8 +1473,11 @@ linux_syscall_trace:
 
 	.globl	C_LABEL(ret_from_fork)
 C_LABEL(ret_from_fork):
+#if CONFIG_SMP || CONFIG_PREEMPT
+/* XXX Wrong location: call schedule_tail in every ret_sys_call. */
 	call	schedule_tail
 	 mov	%g3, %o0
+#endif
 	b	C_LABEL(ret_sys_call)
 	 ld	[%sp + REGWIN_SZ + PT_I0], %o0
 
@@ -1492,7 +1504,8 @@ syscall_is_too_hard:
 	mov	%i1, %o1
 	mov	%i2, %o2
 
-	ld	[%curptr + AOFF_task_ptrace], %l5
+	ld	[%curptr + TI_TASK], %l5
+	ld	[%l5 + AOFF_task_ptrace], %l5
 	mov	%i3, %o3
 	andcc	%l5, 0x02, %g0
 	mov	%i4, %o4
@@ -1506,7 +1519,8 @@ syscall_is_too_hard:
 
 	.globl	C_LABEL(ret_sys_call)
 C_LABEL(ret_sys_call):
-	ld	[%curptr + AOFF_task_ptrace], %l6
+	ld	[%curptr + TI_TASK], %l6
+	ld	[%l6 + AOFF_task_ptrace], %l6
 	cmp	%o0, -ENOIOCTLCMD
 	ld	[%sp + REGWIN_SZ + PT_PSR], %g3
 	set	PSR_C, %g2
@@ -1858,7 +1872,7 @@ kuw_patch1_7win:	sll	%o3, 6, %o3
 	 * traps with the old method of just doing flush_user_windows().
 	 */
 C_LABEL(kill_user_windows):
-	ld	[%g6 + AOFF_task_thread + AOFF_thread_uwinmask], %o0	! get current umask
+	ld	[%g6 + TI_UWINMASK], %o0	! get current umask
 	orcc	%g0, %o0, %g0			! if no bits set, we are done
 	be	3f				! nothing to do
 	 rd	%psr, %o5			! must clear interrupts
@@ -1866,7 +1880,7 @@ C_LABEL(kill_user_windows):
 	wr	%o4, 0x0, %psr			! the uwinmask state
 	WRITE_PAUSE				! burn them cycles
 1:
-	ld	[%g6 + AOFF_task_thread + AOFF_thread_uwinmask], %o0	! get consistant state
+	ld	[%g6 + TI_UWINMASK], %o0	! get consistant state
 	orcc	%g0, %o0, %g0			! did an interrupt come in?
 	be	4f				! yep, we are done
 	 rd	%wim, %o3			! get current wim
@@ -1878,13 +1892,14 @@ kuw_patch1:
 	bne	kuw_patch1			! not done yet
 	 srl	%o3, 1, %o4			! begin another save simulation
 	wr	%o3, 0x0, %wim			! set the new wim
-	st	%g0, [%g6 + AOFF_task_thread + AOFF_thread_uwinmask]	! clear uwinmask
+	st	%g0, [%g6 + TI_UWINMASK]	! clear uwinmask
 4:
 	wr	%o5, 0x0, %psr			! re-enable interrupts
 	WRITE_PAUSE				! burn baby burn
 3:
+	ld	[%g6 + TI_TASK], %o4
 	retl					! return
-	 st	%g0, [%g6 + AOFF_task_thread + AOFF_thread_w_saved]	! no windows saved
+	 st	%g0, [%o4 + AOFF_task_thread + AOFF_thread_w_saved]	! no windows saved
 
 	.align	4
 	.globl	C_LABEL(restore_current)

arch/sparc/kernel/etrap.S

@@ -14,6 +14,7 @@
 #include <asm/ptrace.h>
 #include <asm/winmacro.h>
 #include <asm/asmmacro.h>
+#include <asm/thread_info.h>
 
 /* Registers to not touch at all. */
 #define t_psr        l0 /* Set by caller */
@@ -101,7 +102,7 @@ trap_setup:
 	 mov	%t_kstack, %sp		! jump onto new stack
 
 trap_setup_kernel_spill:
-	ld	[%curptr + AOFF_task_thread + AOFF_thread_uwinmask], %g1
+	ld	[%curptr + TI_UWINMASK], %g1
 	orcc	%g0, %g1, %g0
 	bne	trap_setup_user_spill	! there are some user windows, yuck
 	/* Spill from kernel, but only kernel windows, adjust
@@ -127,7 +128,8 @@ tsetup_patch2:
 	jmpl	%t_retpc + 0x8, %g0	! return to caller
 	 mov	%t_kstack, %sp		! and onto new kernel stack
 
-#define STACK_OFFSET (TASK_UNION_SIZE - (TRACEREG_SZ + REGWIN_SZ))
+#define STACK_OFFSET (THREAD_SIZE - (TRACEREG_SZ + REGWIN_SZ))
+
 trap_setup_from_user:
 	/* We can't use %curptr yet. */
 	LOAD_CURRENT(t_kstack, t_twinmask)
@@ -143,18 +145,19 @@ trap_setup_from_user:
 	STORE_PT_ALL(t_kstack, t_psr, t_pc, t_npc, g2)
 
 #if 0
-	/* If we're sure every task_struct is TASK_UNION_SIZE aligned,
+	/* If we're sure every task_struct is THREAD_SIZE aligned,
 	   we can speed this up. */
 	sethi	%hi(STACK_OFFSET), %curptr
 	or	%curptr, %lo(STACK_OFFSET), %curptr
 	sub	%t_kstack, %curptr, %curptr
 #else
-	sethi	%hi(~(TASK_UNION_SIZE - 1)), %curptr
+	sethi	%hi(~(THREAD_SIZE - 1)), %curptr
 	and	%t_kstack, %curptr, %curptr
 #endif
 
 	/* Clear current->thread.w_saved */
-	st	%g0, [%curptr + AOFF_task_thread + AOFF_thread_w_saved]
+	ld	[%curptr + TI_TASK], %g2
+	st	%g0, [%g2 + AOFF_task_thread + AOFF_thread_w_saved]
 
 	/* See if we are in the trap window. */
 	andcc	%t_twinmask, %t_wim, %g0
@@ -185,7 +188,7 @@ trap_setup_from_user:
 	andn	%g2, %t_twinmask, %g2
 tsetup_patch3:
 	and	%g2, 0xff, %g2			! patched on 7win Sparcs
-	st	%g2, [%curptr + AOFF_task_thread + AOFF_thread_uwinmask]	! store new umask
+	st	%g2, [%curptr + TI_UWINMASK]	! store new umask
 
 	jmpl	%t_retpc + 0x8, %g0		! return to caller
 	 mov	%t_kstack, %sp			! and onto kernel stack
@@ -206,7 +209,7 @@ tsetup_patch5:
 tsetup_patch6:
 	and	%g2, 0xff, %g2		! patched on 7win Sparcs
 	andn	%g1, %g2, %g1		! clear this bit in %g1
-	st	%g1, [%curptr + AOFF_task_thread + AOFF_thread_uwinmask]
+	st	%g1, [%curptr + TI_UWINMASK]
 
 	save	%g0, %g0, %g0
 
@@ -288,7 +291,8 @@ trap_setup_user_stack_is_bolixed:
 	/* From user/kernel into invalid window w/bad user
 	 * stack. Save bad user stack, and return to caller.
 	 */
-	SAVE_BOLIXED_USER_STACK(curptr, g3)
+	ld	[%curptr + TI_TASK], %glob_tmp
+	SAVE_BOLIXED_USER_STACK(glob_tmp, g3)
 	restore	%g0, %g0, %g0
 
 	jmpl	%t_retpc + 0x8, %g0

arch/sparc/kernel/head.S

@@ -23,6 +23,7 @@
 #include <asm/page.h>
 #include <asm/kdebug.h>
 #include <asm/winmacro.h>
+#include <asm/thread_info.h>	/* TI_UWINMASK */
 #include <asm/errno.h>
 
 	.data
@@ -749,7 +750,7 @@ go_to_highmem:
 		jmpl	%g1, %g0
 		 nop
 
-	/* This is to align init_task_union properly, be careful. -DaveM */
+	/* This is to align init_thread_union properly, be careful. -DaveM */
 	.align	8192
 
 /* The code above should be at beginning and we have to take care about
@@ -1010,8 +1011,8 @@ sun4c_continue_boot:
 		WRITE_PAUSE
 
 		/* I want a kernel stack NOW! */
-		set	C_LABEL(init_task_union), %g1
-		set	(TASK_UNION_SIZE - REGWIN_SZ), %g2
+		set	C_LABEL(init_thread_union), %g1
+		set	(THREAD_SIZE - REGWIN_SZ), %g2
 		add	%g1, %g2, %sp
 		mov	0, %fp			/* And for good luck */
 
@@ -1025,10 +1026,10 @@ sun4c_continue_boot:
 		bl	1b
 		 add	%o0, 0x1, %o0
 
-		/* Initialize the umask value for init_task just in case.
+		/* Initialize the uwinmask value for init task just in case.
 		 * But first make current_set[boot_cpu_id] point to something useful.
 		 */
-		set	C_LABEL(init_task_union), %g6
+		set	C_LABEL(init_thread_union), %g6
 		set	C_LABEL(current_set), %g2
 #ifdef CONFIG_SMP
 		sethi	%hi(C_LABEL(boot_cpu_id4)), %g3
@@ -1038,7 +1039,7 @@ sun4c_continue_boot:
 #endif
 		st	%g6, [%g2]
 
-		st	%g0, [%g6 + AOFF_task_thread + AOFF_thread_uwinmask]
+		st	%g0, [%g6 + TI_UWINMASK]
 
 /* Compute NWINDOWS and stash it away. Now uses %wim trick explained
  * in the V8 manual. Ok, this method seems to work, Sparc is cool...

arch/sparc/kernel/init_task.c

@@ -9,12 +9,12 @@ static struct fs_struct init_fs = INIT_FS;
 static struct files_struct init_files = INIT_FILES;
 static struct signal_struct init_signals = INIT_SIGNALS;
 struct mm_struct init_mm = INIT_MM(init_mm);
+struct task_struct init_task = INIT_TASK(init_task);
 
 /* .text section in head.S is aligned at 8k boundry and this gets linked
- * right after that so that the init_task_union is aligned properly as well.
+ * right after that so that the init_thread_union is aligned properly as well.
  * If this is not aligned on a 8k boundry, then you should change code
  * in etrap.S which assumes it.
  */
 __asm__(".section \".text\",#alloc\n");
-union task_union init_task_union =
-		{ INIT_TASK(init_task_union.task) };
+union thread_union init_thread_union = { INIT_THREAD_INFO(init_task) };

arch/sparc/kernel/ioport.c

@@ -530,7 +530,7 @@ void *pci_alloc_consistent(struct pci_dev *pdev, size_t len, dma_addr_t *pba)
 		}
 	}
 
-	*pba = virt_to_bus(va);
+	*pba = virt_to_phys(va); /* equals virt_to_bus (R.I.P.) for us. */
 	return (void *) res->start;
 }
 
@@ -565,7 +565,7 @@ void pci_free_consistent(struct pci_dev *pdev, size_t n, void *p, dma_addr_t ba)
 		return;
 	}
 
-	pgp = (unsigned long) bus_to_virt(ba);
+	pgp = (unsigned long) phys_to_virt(ba);	/* bus_to_virt actually */
 	mmu_inval_dma_area(pgp, n);
 	{
 		int x;
@@ -592,7 +592,7 @@ dma_addr_t pci_map_single(struct pci_dev *hwdev, void *ptr, size_t size,
 	if (direction == PCI_DMA_NONE)
 		BUG();
 	/* IIep is write-through, not flushing. */
-	return virt_to_bus(ptr);
+	return virt_to_phys(ptr);
 }
 
 /* Unmap a single streaming mode DMA translation.  The dma_addr and size
@@ -608,7 +608,7 @@ void pci_unmap_single(struct pci_dev *hwdev, dma_addr_t ba, size_t size,
 	if (direction == PCI_DMA_NONE)
 		BUG();
 	if (direction != PCI_DMA_TODEVICE) {
-		mmu_inval_dma_area((unsigned long)bus_to_virt(ba),
+		mmu_inval_dma_area((unsigned long)phys_to_virt(ba),
 		    (size + PAGE_SIZE-1) & PAGE_MASK);
 	}
 }
@@ -637,7 +637,8 @@ int pci_map_sg(struct pci_dev *hwdev, struct scatterlist *sg, int nents,
 		BUG();
 	/* IIep is write-through, not flushing. */
 	for (n = 0; n < nents; n++) {
-		sg->dvma_address = virt_to_bus(sg->address);
+		if (page_address(sg->page) == NULL) BUG();
+		sg->dvma_address = virt_to_phys(page_address(sg->page));
 		sg->dvma_length = sg->length;
 		sg++;
 	}
@@ -657,7 +658,9 @@ void pci_unmap_sg(struct pci_dev *hwdev, struct scatterlist *sg, int nents,
 		BUG();
 	if (direction != PCI_DMA_TODEVICE) {
 		for (n = 0; n < nents; n++) {
-			mmu_inval_dma_area((unsigned long)sg->address,
+			if (page_address(sg->page) == NULL) BUG();
+			mmu_inval_dma_area(
+			    (unsigned long) page_address(sg->page),
 			    (sg->length + PAGE_SIZE-1) & PAGE_MASK);
 			sg++;
 		}
@@ -678,7 +681,7 @@ void pci_dma_sync_single(struct pci_dev *hwdev, dma_addr_t ba, size_t size, int
 	if (direction == PCI_DMA_NONE)
 		BUG();
 	if (direction != PCI_DMA_TODEVICE) {
-		mmu_inval_dma_area((unsigned long)bus_to_virt(ba),
+		mmu_inval_dma_area((unsigned long)phys_to_virt(ba),
 		    (size + PAGE_SIZE-1) & PAGE_MASK);
 	}
 }
@@ -697,7 +700,9 @@ void pci_dma_sync_sg(struct pci_dev *hwdev, struct scatterlist *sg, int nents, i
 		BUG();
 	if (direction != PCI_DMA_TODEVICE) {
 		for (n = 0; n < nents; n++) {
-			mmu_inval_dma_area((unsigned long)sg->address,
+			if (page_address(sg->page) == NULL) BUG();
+			mmu_inval_dma_area(
+			    (unsigned long) page_address(sg->page),
 			    (sg->length + PAGE_SIZE-1) & PAGE_MASK);
 			sg++;
 		}

arch/sparc/kernel/irq.c

@@ -6,7 +6,7 @@
  *
  *  Copyright (C) 1995 David S. Miller (davem@caip.rutgers.edu)
  *  Copyright (C) 1995 Miguel de Icaza (miguel@nuclecu.unam.mx)
- *  Copyright (C) 1995 Pete A. Zaitcev (zaitcev@yahoo.com)
+ *  Copyright (C) 1995,2002 Pete A. Zaitcev (zaitcev@yahoo.com)
  *  Copyright (C) 1996 Dave Redman (djhr@tadpole.co.uk)
  *  Copyright (C) 1998-2000 Anton Blanchard (anton@samba.org)
  */
@@ -46,6 +46,7 @@
 #include <asm/hardirq.h>
 #include <asm/softirq.h>
 #include <asm/pcic.h>
+#include <asm/cacheflush.h>
 
 /*
  * Dave Redman (djhr@tadpole.co.uk)
@@ -119,9 +120,11 @@ int show_interrupts(struct seq_file *p, void *v)
 #ifndef CONFIG_SMP
 		seq_printf(p, "%10u ", kstat_irqs(i));
 #else
-		for (j = 0; j < smp_num_cpus; j++)
+		for (j = 0; j < NR_CPUS; j++) {
+			if (cpu_online(j))
 				seq_printf(p, "%10u ",
 				    kstat.irqs[cpu_logical_map(j)][i]);
+		}
 #endif
 		seq_printf(p, " %c %s",
 			(action->flags & SA_INTERRUPT) ? '+' : ' ',
@@ -214,12 +217,16 @@ static void show(char * str)
 
 	printk("\n%s, CPU %d:\n", str, cpu);
 	printk("irq:  %d [ ", irqs_running());
-	for (i = 0; i < smp_num_cpus; i++)
+	for (i = 0; i < NR_CPUS; i++) {
+		if (cpu_online(i))
 			printk("%u ", __brlock_array[i][BR_GLOBALIRQ_LOCK]);
+	}
 	printk("]\nbh:   %d [ ",
 	       (spin_is_locked(&global_bh_lock) ? 1 : 0));
-	for (i = 0; i < smp_num_cpus; i++)
+	for (i = 0; i < NR_CPUS; i++) {
+		if (cpu_online(i))
 			printk("%u ", local_bh_count(i));
+	}
 	printk("]\n");
 
 #ifdef VERBOSE_DEBUG_IRQLOCK

arch/sparc/kernel/process.c

@@ -57,7 +57,14 @@ void (*pm_power_off)(void);
 extern void fpsave(unsigned long *, unsigned long *, void *, unsigned long *);
 
 struct task_struct *last_task_used_math = NULL;
-struct task_struct *current_set[NR_CPUS] = {&init_task, };
+struct thread_info *current_set[NR_CPUS];
+
+/*
+ * default_idle is new in 2.5. XXX Review, currently stolen from sparc64.
+ */
+void default_idle(void)
+{
+}
 
 #ifndef CONFIG_SMP
 
@@ -106,8 +113,8 @@ int cpu_idle(void)
 			restore_flags(flags);
 		}
 
-		while((!current->need_resched) && pm_idle) {
-				(*pm_idle)();
+		while((!need_resched()) && pm_idle) {
+			(*pm_idle)();		/* XXX Huh? On sparc?! */
 		}
 
 		schedule();
@@ -306,7 +313,7 @@ void show_trace_task(struct task_struct *tsk)
 	if (!tsk)
 		return;
 
-	fp = tsk->thread.ksp;
+	fp = tsk->thread_info->ksp;
 	do {
 		/* Bogus frame pointer? */
 		if (fp < (task_base + sizeof(struct task_struct)) ||
@@ -320,6 +327,14 @@ void show_trace_task(struct task_struct *tsk)
 	printk("\n");
 }
 
+/*
+ * Note: sparc64 has a pretty intricated thread_saved_pc, check it out.
+ */
+unsigned long thread_saved_pc(struct task_struct *tsk)
+{
+	return tsk->thread_info->kpc;
+}
+
 /*
  * Free current thread data structures etc..
  */
@@ -372,7 +387,7 @@ void flush_thread(void)
 		/* We must fixup kregs as well. */
 		current->thread.kregs = (struct pt_regs *)
 			(((unsigned long)current) +
-			 (TASK_UNION_SIZE - TRACEREG_SZ));
+			 (THREAD_SIZE - TRACEREG_SZ));
 	}
 }
 
@@ -445,6 +460,22 @@ clone_stackframe(struct sparc_stackf *dst, struct sparc_stackf *src)
 	return sp;
 }
 
+asmlinkage int sparc_do_fork(unsigned long clone_flags,
+                             unsigned long stack_start,
+                             struct pt_regs *regs,
+                             unsigned long stack_size)
+{
+	struct task_struct *p;
+
+	/* XXX This was spelled in DaveM's will and testament. Why? */
+	if (clone_flags & CLONE_IDLETASK) {
+		printk(KERN_DEBUG "Userland clone with CLONE_IDLETASK\n");
+		clone_flags &= ~CLONE_IDLETASK;
+	}
+
+	p = do_fork(clone_flags, stack_start, regs, stack_size);
+	return IS_ERR(p) ? PTR_ERR(p) : p->pid;
+}
 
 /* Copy a Sparc thread.  The fork() return value conventions
  * under SunOS are nothing short of bletcherous:
@@ -457,6 +488,7 @@ clone_stackframe(struct sparc_stackf *dst, struct sparc_stackf *src)
  *       if the parent should sleep while trying to
  *       allocate the task_struct and kernel stack in
  *       do_fork().
+ * XXX See comment above sys_vfork in sparc64. todo.
  */
 extern void ret_from_fork(void);
 
@@ -464,6 +496,7 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
 		unsigned long unused,
 		struct task_struct *p, struct pt_regs *regs)
 {
+	struct thread_info *ti = p->thread_info;
 	struct pt_regs *childregs;
 	struct reg_window *new_stack;
 	unsigned long stack_offset;
@@ -482,19 +515,19 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
 	}
 
 	/* Calculate offset to stack_frame & pt_regs */
-	stack_offset = TASK_UNION_SIZE - TRACEREG_SZ;
+	stack_offset = THREAD_SIZE - TRACEREG_SZ;
 
 	if(regs->psr & PSR_PS)
 		stack_offset -= REGWIN_SZ;
-	childregs = ((struct pt_regs *) (((unsigned long)p) + stack_offset));
+	childregs = ((struct pt_regs *) (((unsigned long)ti) + stack_offset));
 	copy_regs(childregs, regs);
 	new_stack = (((struct reg_window *) childregs) - 1);
 	copy_regwin(new_stack, (((struct reg_window *) regs) - 1));
 
-	p->thread.ksp = (unsigned long) new_stack;
-	p->thread.kpc = (((unsigned long) ret_from_fork) - 0x8);
-	p->thread.kpsr = current->thread.fork_kpsr;
-	p->thread.kwim = current->thread.fork_kwim;
+	ti->ksp = (unsigned long) new_stack;
+	ti->kpc = (((unsigned long) ret_from_fork) - 0x8);
+	ti->kpsr = current->thread.fork_kpsr;
+	ti->kwim = current->thread.fork_kwim;
 
 	/* This is used for sun4c only */
 	atomic_set(&p->thread.refcount, 1);
@@ -504,16 +537,14 @@ int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
 
 		p->thread.kregs = &fake_swapper_regs;
 		new_stack = (struct reg_window *)
-			((((unsigned long)p) +
-			  (TASK_UNION_SIZE)) -
-			 (REGWIN_SZ));
+			((((unsigned long)ti) + (THREAD_SIZE)) - REGWIN_SZ);
 		childregs->u_regs[UREG_FP] = (unsigned long) new_stack;
 		p->thread.flags |= SPARC_FLAG_KTHREAD;
 		p->thread.current_ds = KERNEL_DS;
 		memcpy((void *)new_stack,
 		       (void *)regs->u_regs[UREG_FP],
 		       sizeof(struct reg_window));
-		childregs->u_regs[UREG_G6] = (unsigned long) p;
+		childregs->u_regs[UREG_G6] = (unsigned long) ti;
 	} else {
 		p->thread.kregs = childregs;
 		childregs->u_regs[UREG_FP] = sp;

arch/sparc/kernel/ptrace.c

@@ -92,16 +92,16 @@ static inline void read_sunos_user(struct pt_regs *regs, unsigned long offset,
 	}
 	switch(offset) {
 	case 0:
-		v = t->ksp;
+		v = tsk->thread_info->ksp;
 		break;
 	case 4:
-		v = t->kpc;
+		v = tsk->thread_info->kpc;
 		break;
 	case 8:
-		v = t->kpsr;
+		v = tsk->thread_info->kpsr;
 		break;
 	case 12:
-		v = t->uwinmask;
+		v = tsk->thread_info->uwinmask;
 		break;
 	case 832:
 		v = t->w_saved;
@@ -530,9 +530,9 @@ asmlinkage void do_ptrace(struct pt_regs *regs)
 		}
 
 		if (request == PTRACE_SYSCALL)
-			child->ptrace |= PT_TRACESYS;
+			set_tsk_thread_flag(child, TIF_SYSCALL_TRACE);
 		else
-			child->ptrace &= ~PT_TRACESYS;
+			clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE);
 
 		child->exit_code = data;
 #ifdef DEBUG_PTRACE
@@ -581,7 +581,7 @@ asmlinkage void do_ptrace(struct pt_regs *regs)
 	}
 out_tsk:
 	if (child)
-		free_task_struct(child);
+		put_task_struct(child);
 out:
 	unlock_kernel();
 }
@@ -591,8 +591,9 @@ asmlinkage void syscall_trace(void)
 #ifdef DEBUG_PTRACE
 	printk("%s [%d]: syscall_trace\n", current->comm, current->pid);
 #endif
-	if ((current->ptrace & (PT_PTRACED|PT_TRACESYS))
-			!= (PT_PTRACED|PT_TRACESYS))
+	if (!test_thread_flag(TIF_SYSCALL_TRACE))
+		return;
+	if (!(current->ptrace & PT_PTRACED))
 		return;
 	current->exit_code = SIGTRAP;
 	current->state = TASK_STOPPED;

arch/sparc/kernel/rtrap.S

@@ -13,6 +13,7 @@
 #include <asm/contregs.h>
 #include <asm/winmacro.h>
 #include <asm/asmmacro.h>
+#include <asm/thread_info.h>
 
 #define t_psr     l0
 #define t_pc      l1
@@ -58,17 +59,17 @@ C_LABEL(ret_trap_lockless_ipi):
 	 nop
 
 1:
-#error	ld	[%curptr + AOFF_task_need_resched], %g2
-	orcc	%g2, %g0, %g0
+	ld	[%curptr + TI_FLAGS], %g2
+	andcc	%g2, (_TIF_NEED_RESCHED), %g0
 	be	signal_p
-#error	 ld	[%curptr + AOFF_task_sigpending], %g2
+	 nop
 
 	call	C_LABEL(schedule)
 	 nop
 
-#error	ld	[%curptr + AOFF_task_sigpending], %g2
+	ld	[%curptr + TI_FLAGS], %g2
 signal_p:
-	cmp	%g2, 0
+	andcc	%g2, (_TIF_NOTIFY_RESUME|_TIF_SIGPENDING), %g0
 	bz,a	ret_trap_continue
 	 ld	[%sp + REGWIN_SZ + PT_PSR], %t_psr
 
@@ -98,7 +99,7 @@ ret_trap_continue:
 	 add	%sp, REGWIN_SZ, %o0
 
 	b	signal_p
-#error	 ld	[%curptr + AOFF_task_sigpending], %g2
+	 ld	[%curptr + TI_FLAGS], %g2
 
 ret_trap_nobufwins:
 	/* Load up the user's out registers so we can pull
@@ -109,7 +110,7 @@ ret_trap_nobufwins:
 	/* If there are already live user windows in the
 	 * set we can return from trap safely.
 	 */
-	ld	[%curptr + AOFF_task_thread + AOFF_thread_uwinmask], %twin_tmp1
+	ld	[%curptr + TI_UWINMASK], %twin_tmp1
 	orcc	%g0, %twin_tmp1, %g0
 	bne	ret_trap_userwins_ok
 	 nop
@@ -168,7 +169,7 @@ ret_trap_unaligned_pc:
 	 nop
 
 	b	signal_p
-#error	 ld	[%curptr + AOFF_task_sigpending], %g2
+	 ld	[%curptr + TI_FLAGS], %g2
 
 ret_trap_kernel:
 		/* Will the rett land us in the invalid window? */
@@ -218,7 +219,8 @@ ret_trap_user_stack_is_bolixed:
 	 add	%sp, REGWIN_SZ, %o0
 
 	b	signal_p
-#error	 ld	[%curptr + AOFF_task_sigpending], %g2
+	 ld	[%curptr + TI_FLAGS], %g2
+
 
 	.globl	C_LABEL(sun4c_rett_stackchk)
 C_LABEL(sun4c_rett_stackchk):

arch/sparc/kernel/sclow.S

@@ -10,6 +10,7 @@
 #include <asm/ptrace.h>
 #include <asm/errno.h>
 #include <asm/winmacro.h>
+#include <asm/thread_info.h>
 #include <asm/psr.h>
 #include <asm/page.h>
 
@@ -40,6 +41,7 @@ LABEL(sunosnop):
 	.globl	LABEL(sunosgetpid)
 LABEL(sunosgetpid):
 	LOAD_CURRENT(l4, l5)
+	ld	[%l4 + TI_TASK], %l4
 	ld	[%l4 + AOFF_task_pid], %i0
 	ld	[%l4 + AOFF_task_p_opptr], %l5
 	ld	[%l5 + AOFF_task_pid], %i1
@@ -50,6 +52,7 @@ LABEL(sunosgetpid):
 	.globl	LABEL(sunosgetuid)
 LABEL(sunosgetuid):
 	LOAD_CURRENT(l4, l5)
+	ld	[%l4 + TI_TASK], %l4
 	lduh	[%l4 + AOFF_task_uid], %i0
 	lduh	[%l4 + AOFF_task_euid], %i1
 	CC_AND_RETT
@@ -59,6 +62,7 @@ LABEL(sunosgetuid):
 	.globl	LABEL(sunosgetgid)
 LABEL(sunosgetgid):
 	LOAD_CURRENT(l4, l5)
+	ld	[%l4 + TI_TASK], %l4
 	lduh	[%l4 + AOFF_task_gid], %i0
 	lduh	[%l4 + AOFF_task_egid], %i1
 	CC_AND_RETT
@@ -77,6 +81,7 @@ LABEL(sunosgdtsize):
 	.globl	LABEL(sunossblock)
 LABEL(sunossblock):
 	LOAD_CURRENT(l4, l5)
+	ld	[%l4 + TI_TASK], %l4
 	set	-65793, %l5
 	and	%i0, %l5, %l5
 	ld	[%l4 + AOFF_task_blocked], %i0
@@ -87,6 +92,7 @@ LABEL(sunossblock):
 	.globl	LABEL(sunossmask)
 LABEL(sunossmask):
 	LOAD_CURRENT(l4, l5)
+	ld	[%l4 + TI_TASK], %l4
 	set	-65793, %l5
 	and	%i0, %l5, %l5
 	ld	[%l4 + AOFF_task_blocked], %i0

arch/sparc/kernel/semaphore.c

@@ -3,6 +3,7 @@
 /* sparc32 semaphore implementation, based on i386 version */
 
 #include <linux/sched.h>
+#include <linux/errno.h>
 
 #include <asm/semaphore.h>
 

arch/sparc/kernel/setup.c

@@ -481,7 +481,7 @@ static int show_cpuinfo(struct seq_file *m, void *__unused)
 		   (short) romvec->pv_printrev,
 		   &cputypval,
 		   linux_num_cpus,
-		   smp_num_cpus
+		   num_online_cpus()
 #ifndef CONFIG_SMP
 		   , loops_per_jiffy/(500000/HZ),
 		   (loops_per_jiffy/(5000/HZ)) % 100

arch/sparc/kernel/signal.c

@@ -19,6 +19,7 @@
 #include <linux/tty.h>
 #include <linux/smp.h>
 #include <linux/smp_lock.h>
+#include <linux/binfmts.h>	/* do_coredum */
 
 #include <asm/uaccess.h>
 #include <asm/bitops.h>
@@ -26,6 +27,7 @@
 #include <asm/svr4.h>
 #include <asm/pgalloc.h>
 #include <asm/pgtable.h>
+#include <asm/cacheflush.h>	/* flush_sig_insns */
 
 #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP)))
 
@@ -34,7 +36,7 @@ extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
 extern void fpload(unsigned long *fpregs, unsigned long *fsr);
 
 asmlinkage int do_signal(sigset_t *oldset, struct pt_regs * regs,
-			 unsigned long orig_o0, int ret_from_syscall);
+			 unsigned long orig_o0, int restart_syscall);
 
 /* This turned off for production... */
 /* #define DEBUG_SIGNALS 1 */
@@ -229,7 +231,7 @@ restore_fpu_state(struct pt_regs *regs, __siginfo_fpu_t *fpu)
 {
 	int err;
 #ifdef CONFIG_SMP
-	if (current->flags & PF_USEDFPU)
+	if (test_tsk_thread_flag(current, TIF_USEDFPU))
 		regs->psr &= ~PSR_EF;
 #else
 	if (current == last_task_used_math) {
@@ -238,7 +240,7 @@ restore_fpu_state(struct pt_regs *regs, __siginfo_fpu_t *fpu)
 	}
 #endif
 	current->used_math = 1;
-	current->flags &= ~PF_USEDFPU;
+	clear_tsk_thread_flag(current, TIF_USEDFPU);
 
 	if (verify_area (VERIFY_READ, fpu, sizeof(*fpu)))
 		return -EFAULT;
@@ -586,12 +588,12 @@ save_fpu_state(struct pt_regs *regs, __siginfo_fpu_t *fpu)
 {
 	int err = 0;
 #ifdef CONFIG_SMP
-	if (current->flags & PF_USEDFPU) {
+	if (test_tsk_thread_flag(current, TIF_USEDFPU)) {
 		put_psr(get_psr() | PSR_EF);
 		fpsave(&current->thread.float_regs[0], &current->thread.fsr,
 		       &current->thread.fpqueue[0], &current->thread.fpqdepth);
 		regs->psr &= ~(PSR_EF);
-		current->flags &= ~(PF_USEDFPU);
+		clear_tsk_thread_flag(current, TIF_USEDFPU);
 	}
 #else
 	if (current == last_task_used_math) {
@@ -1295,7 +1297,7 @@ asmlinkage int do_signal(sigset_t *oldset, struct pt_regs * regs,
 				/* fall through */
 			default:
 				sigaddset(&current->pending.signal, signr);
-				recalc_sigpending(current);
+				recalc_sigpending();
 				current->flags |= PF_SIGNALED;
 				do_exit(exit_code);
 				/* NOT REACHED */

arch/sparc/kernel/sparc-stub.c

@@ -109,6 +109,8 @@
 #include <asm/kgdb.h>
 #include <asm/pgalloc.h>
 #include <asm/pgtable.h>
+#include <asm/cacheflush.h>
+
 /*
  *
  * external low-level support routines

arch/sparc/kernel/sparc_ksyms.c

@@ -142,7 +142,6 @@ EXPORT_SYMBOL(__global_save_flags);
 EXPORT_SYMBOL(__global_restore_flags);
 
 /* Misc SMP information */
-EXPORT_SYMBOL(smp_num_cpus);
 EXPORT_SYMBOL(__cpu_number_map);
 EXPORT_SYMBOL(__cpu_logical_map);
 #endif

arch/sparc/kernel/sun4d_irq.c

@@ -99,9 +99,11 @@ found_it:	seq_printf(p, "%3d: ", i);
 #ifndef CONFIG_SMP
 		seq_printf(p, "%10u ", kstat_irqs(i));
 #else
-		for (x = 0; x < smp_num_cpus; x++)
+		for (x = 0; x < NR_CPUS; x++) {
+			if (cpu_online)
 				seq_printf(p, "%10u ",
 				       kstat.irqs[cpu_logical_map(x)][i]);
+		}
 #endif
 		seq_printf(p, "%c %s",
 			(action->flags & SA_INTERRUPT) ? '+' : ' ',

arch/sparc/kernel/sun4d_smp.c

@@ -43,7 +43,6 @@ extern int linux_num_cpus;
 
 extern void calibrate_delay(void);
 
-extern struct task_struct *current_set[NR_CPUS];
 extern volatile int smp_processors_ready;
 extern unsigned long cpu_present_map;
 extern int smp_num_cpus;

arch/sparc/kernel/sun4m_smp.c

@@ -40,7 +40,6 @@ extern int linux_num_cpus;
 
 extern void calibrate_delay(void);
 
-extern struct task_struct *current_set[NR_CPUS];
 extern volatile int smp_processors_ready;
 extern unsigned long cpu_present_map;
 extern int smp_num_cpus;

arch/sparc/kernel/trampoline.S

@@ -63,8 +63,8 @@ cpu3_startup:
 	and	%g4, 0xc, %g4
 	ld	[%g5 + %g4], %g6
 
-	sethi	%hi(TASK_UNION_SIZE - REGWIN_SZ), %sp
-	or	%sp, %lo(TASK_UNION_SIZE - REGWIN_SZ), %sp
+	sethi	%hi(THREAD_SIZE - REGWIN_SZ), %sp
+	or	%sp, %lo(THREAD_SIZE - REGWIN_SZ), %sp
 	add	%g6, %sp, %sp
 
 	/* Turn on traps (PSR_ET). */
@@ -142,8 +142,8 @@ C_LABEL(sun4d_cpu_startup):
 	srl	%g3, 1, %g4
 	ld	[%g5 + %g4], %g6
 
-	sethi	%hi(TASK_UNION_SIZE - REGWIN_SZ), %sp
-	or	%sp, %lo(TASK_UNION_SIZE - REGWIN_SZ), %sp
+	sethi	%hi(THREAD_SIZE - REGWIN_SZ), %sp
+	or	%sp, %lo(THREAD_SIZE - REGWIN_SZ), %sp
 	add	%g6, %sp, %sp
 
 	/* Turn on traps (PSR_ET). */

arch/sparc/kernel/traps.c

@@ -475,6 +475,23 @@ int thiscpus_mid;
 
 void trap_init(void)
 {
+	extern void thread_info_offsets_are_bolixed_pete(void);
+
+	/* Force linker to barf if mismatched */
+	if (TI_UWINMASK    != offsetof(struct thread_info, uwinmask) ||
+	    TI_TASK        != offsetof(struct thread_info, task) ||
+	    TI_EXECDOMAIN  != offsetof(struct thread_info, exec_domain) ||
+	    TI_FLAGS       != offsetof(struct thread_info, flags) ||
+	    TI_CPU         != offsetof(struct thread_info, cpu) ||
+	    TI_PREEMPT     != offsetof(struct thread_info, preempt_count) ||
+	    TI_SOFTIRQ     != offsetof(struct thread_info, softirq_count) ||
+	    TI_HARDIRQ     != offsetof(struct thread_info, hardirq_count) ||
+	    TI_KSP         != offsetof(struct thread_info, ksp) ||
+	    TI_KPC         != offsetof(struct thread_info, kpc) ||
+	    TI_KPSR        != offsetof(struct thread_info, kpsr) ||
+	    TI_KWIM        != offsetof(struct thread_info, kwim))
+		thread_info_offsets_are_bolixed_pete();
+
 	/* Attach to the address space of init_task. */
 	atomic_inc(&init_mm.mm_count);
 	current->active_mm = &init_mm;

arch/sparc/kernel/windows.c

@@ -32,7 +32,7 @@ void flush_user_windows(void)
 		" restore %%g0, %%g0, %%g0\n"
 	: "=&r" (ctr)
 	: "0" (ctr),
-	  "i" ((const unsigned long)(&(((struct task_struct *)0)->thread.uwinmask)))
+	  "i" ((const unsigned long)TI_UWINMASK)
 	: "g4", "cc");
 }
 

arch/sparc/kernel/wof.S

@@ -13,6 +13,7 @@
 #include <asm/asi.h>
 #include <asm/winmacro.h>
 #include <asm/asmmacro.h>
+#include <asm/thread_info.h>
 
 /* WARNING: This routine is hairy and _very_ complicated, but it
  *          must be as fast as possible as it handles the allocation
@@ -62,7 +63,7 @@ spnwin_patch3_7win:	and	%twin_tmp, 0x7f, %twin_tmp
 	 * andcc %l0, PSR_PS, %g0
 	 */
 
-	/* Datum current->thread.uwinmask contains at all times a bitmask
+	/* Datum current_thread_info->uwinmask contains at all times a bitmask
 	 * where if any user windows are active, at least one bit will
 	 * be set in to mask.  If no user windows are active, the bitmask
 	 * will be all zeroes.
@@ -96,10 +97,10 @@ spnwin_patch2:	and	%glob_tmp, 0xff, %glob_tmp
 	 save	%g0, %g0, %g0				! Go where saving will occur
 	
 	/* See if any user windows are active in the set. */
-	ld	[%curptr + AOFF_task_thread + AOFF_thread_uwinmask], %twin_tmp	! grab win mask
+	ld	[%curptr + TI_UWINMASK], %twin_tmp	! grab win mask
 	orcc	%g0, %twin_tmp, %g0			! check for set bits
 	bne	spwin_exist_uwins			! yep, there are some
-	 andn	%twin_tmp, %glob_tmp, %twin_tmp		! compute new umask
+	 andn	%twin_tmp, %glob_tmp, %twin_tmp		! compute new uwinmask
 
 	/* Save into the window which must be saved and do it.
 	 * Basically if we are here, this means that we trapped
@@ -139,7 +140,7 @@ spwin_exist_uwins:
 	 * But first, store the new user window mask calculated
 	 * above.
 	 */
-	st	%twin_tmp, [%curptr + AOFF_task_thread + AOFF_thread_uwinmask]
+	st	%twin_tmp, [%curptr + TI_UWINMASK]
 	save	%g0, %g0, %g0		! Go to where the saving will occur
 
 spwin_fromuser:
@@ -209,18 +210,20 @@ spwin_user_stack_is_bolixed:
 	bne	spwin_bad_ustack_from_kernel
 	 nop
 
+	ld	[%curptr + TI_TASK], %glob_tmp
+
 	/* Oh well, throw this one window into the per-task window
 	 * buffer, the first one.
 	 */
-	st	%sp, [%curptr + AOFF_task_thread + AOFF_thread_rwbuf_stkptrs]
-	STORE_WINDOW(curptr + AOFF_task_thread + AOFF_thread_reg_window)
+	st	%sp, [%glob_tmp + AOFF_task_thread + AOFF_thread_rwbuf_stkptrs]
+	STORE_WINDOW(glob_tmp + AOFF_task_thread + AOFF_thread_reg_window)
 	restore	%g0, %g0, %g0
 
 	/* LOCATION: Trap Window */
 
 	/* Back in the trap window, update winbuffer save count. */
-	mov	1, %glob_tmp
-	st	%glob_tmp, [%curptr + AOFF_task_thread + AOFF_thread_w_saved]
+	mov	1, %twin_tmp
+	st	%twin_tmp, [%glob_tmp + AOFF_task_thread + AOFF_thread_w_saved]
 
 		/* Compute new user window mask.  What we are basically
 		 * doing is taking two windows, the invalid one at trap
@@ -232,9 +235,9 @@ spwin_user_stack_is_bolixed:
 		or	%twin_tmp, %t_wim, %twin_tmp
 		not	%twin_tmp
 spnwin_patch3:	and	%twin_tmp, 0xff, %twin_tmp	! patched on 7win Sparcs
-		st	%twin_tmp, [%curptr + AOFF_task_thread + AOFF_thread_uwinmask]
+		st	%twin_tmp, [%curptr + TI_UWINMASK]
 
-#define STACK_OFFSET	(TASK_UNION_SIZE - TRACEREG_SZ - REGWIN_SZ)
+#define STACK_OFFSET	(THREAD_SIZE - TRACEREG_SZ - REGWIN_SZ)
 
 	sethi	%hi(STACK_OFFSET), %sp
 	or	%sp, %lo(STACK_OFFSET), %sp
@@ -247,7 +250,7 @@ spnwin_patch3:	and	%twin_tmp, 0xff, %twin_tmp	! patched on 7win Sparcs
 
 	sethi	%hi(STACK_OFFSET), %g6
 	or	%g6, %lo(STACK_OFFSET), %g6
-	sub	%sp, %g6, %g6
+	sub	%sp, %g6, %g6		! curptr
 
 	/* Turn on traps and call c-code to deal with it. */
 	wr	%t_psr, PSR_ET, %psr
@@ -271,7 +274,8 @@ spwin_bad_ustack_from_kernel:
 	 * a per-process window buffer until we can properly handle
 	 * this later on.
 	 */
-	SAVE_BOLIXED_USER_STACK(curptr, glob_tmp)
+	ld	[%curptr + TI_TASK], %glob_tmp	/* Using curptr one last time */
+	SAVE_BOLIXED_USER_STACK(glob_tmp, g6)	/* ...now using g6 as scratch */
 	restore	%g0, %g0, %g0
 
 	/* LOCATION: Trap window */

arch/sparc/kernel/wuf.S

@@ -13,6 +13,7 @@
 #include <asm/asi.h>
 #include <asm/winmacro.h>
 #include <asm/asmmacro.h>
+#include <asm/thread_info.h>
 
 /* Just like the overflow handler we define macros for registers
  * with fixed meanings in this routine.
@@ -40,7 +41,7 @@
 	 * andcc %l0, PSR_PS, %g0
 	 */
 
-	/* Datum current->thread.uwinmask contains at all times a bitmask
+	/* Datum current_thread_info->uwinmask contains at all times a bitmask
 	 * where if any user windows are active, at least one bit will
 	 * be set in to mask.  If no user windows are active, the bitmask
 	 * will be all zeroes.
@@ -138,7 +139,7 @@ fwin_from_user:
 C_LABEL(fwin_mmu_patchme):	b	C_LABEL(sun4c_fwin_stackchk)
 				 andcc	%sp, 0x7, %g0
 
-#define STACK_OFFSET (TASK_UNION_SIZE - TRACEREG_SZ - REGWIN_SZ)
+#define STACK_OFFSET (THREAD_SIZE - TRACEREG_SZ - REGWIN_SZ)
 
 fwin_user_stack_is_bolixed:
 	/* LOCATION: Window 'W' */
@@ -184,8 +185,9 @@ fwin_user_stack_is_bolixed:
 	/* Fix users window mask and buffer save count. */
 	mov	0x1, %g5
 	sll	%g5, %g3, %g5
-	st	%g5, [%curptr + AOFF_task_thread + AOFF_thread_uwinmask]		! one live user window still
-	st	%g0, [%curptr + AOFF_task_thread + AOFF_thread_w_saved]		! no windows in the buffer
+	st	%g5, [%curptr + TI_UWINMASK]		! one live user window still
+	ld	[%curptr + TI_TASK], %g5
+	st	%g0, [%g5 + AOFF_task_thread + AOFF_thread_w_saved]		! no windows in the buffer
 
 	wr	%t_psr, PSR_ET, %psr			! enable traps
 	nop

arch/sparc/mm/btfixup.c

@@ -15,6 +15,7 @@
 #include <asm/pgtable.h>
 #include <asm/oplib.h>
 #include <asm/system.h>
+#include <asm/cacheflush.h>
 
 #define BTFIXUP_OPTIMIZE_NOP
 #define BTFIXUP_OPTIMIZE_OTHER

arch/sparc/mm/fault.c

@@ -138,8 +138,8 @@ static void unhandled_fault(unsigned long address, struct task_struct *tsk,
                      struct pt_regs *regs)
 {
 	if((unsigned long) address < PAGE_SIZE) {
-		printk(KERN_ALERT "Unable to handle kernel NULL "
-		       "pointer dereference");
+		printk(KERN_ALERT
+		    "Unable to handle kernel NULL pointer dereference\n");
 	} else {
 		printk(KERN_ALERT "Unable to handle kernel paging request "
 		       "at virtual address %08lx\n", address);
@@ -401,7 +401,7 @@ asmlinkage void do_sun4c_fault(struct pt_regs *regs, int text_fault, int write,
 {
 	extern void sun4c_update_mmu_cache(struct vm_area_struct *,
 					   unsigned long,pte_t);
-	extern pte_t *sun4c_pte_offset(pmd_t *,unsigned long);
+	extern pte_t *sun4c_pte_offset_kernel(pmd_t *,unsigned long);
 	struct task_struct *tsk = current;
 	struct mm_struct *mm = tsk->mm;
 	pgd_t *pgdp;
@@ -421,7 +421,7 @@ asmlinkage void do_sun4c_fault(struct pt_regs *regs, int text_fault, int write,
 	}
 
 	pgdp = pgd_offset(mm, address);
-	ptep = sun4c_pte_offset((pmd_t *) pgdp, address);
+	ptep = sun4c_pte_offset_kernel((pmd_t *) pgdp, address);
 
 	if (pgd_val(*pgdp)) {
 	    if (write) {

arch/sparc/mm/generic.c

@@ -16,6 +16,7 @@
 
 static inline void forget_pte(pte_t page)
 {
+#if 0 /* old 2.4 code */
 	if (pte_none(page))
 		return;
 	if (pte_present(page)) {
@@ -30,6 +31,12 @@ static inline void forget_pte(pte_t page)
 		return;
 	}
 	swap_free(pte_to_swp_entry(page));
+#else
+	if (!pte_none(page)) {
+		printk("forget_pte: old mapping existed!\n");
+		BUG();
+	}
+#endif
 }
 
 /* Remap IO memory, the same way as remap_page_range(), but use
@@ -69,7 +76,7 @@ static inline int io_remap_pmd_range(pmd_t * pmd, unsigned long address, unsigne
 		end = PGDIR_SIZE;
 	offset -= address;
 	do {
-		pte_t * pte = pte_alloc(current->mm, pmd, address);
+		pte_t * pte = pte_alloc_map(current->mm, pmd, address);
 		if (!pte)
 			return -ENOMEM;
 		io_remap_pte_range(pte, address, end - address, address + offset, prot, space);

arch/sparc/mm/init.c

@@ -32,6 +32,7 @@
 #include <asm/page.h>
 #include <asm/pgtable.h>
 #include <asm/vaddrs.h>
+#include <asm/pgalloc.h>	/* bug in asm-generic/tlb.h: check_pgt_cache */
 #include <asm/tlb.h>
 
 mmu_gather_t mmu_gathers[NR_CPUS];
@@ -60,7 +61,7 @@ pte_t *kmap_pte;
 pgprot_t kmap_prot;
 
 #define kmap_get_fixed_pte(vaddr) \
-	pte_offset(pmd_offset(pgd_offset_k(vaddr), (vaddr)), (vaddr))
+	pte_offset_kernel(pmd_offset(pgd_offset_k(vaddr), (vaddr)), (vaddr))
 
 void __init kmap_init(void)
 {
@@ -77,11 +78,13 @@ void show_mem(void)
 	       nr_swap_pages << (PAGE_SHIFT-10));
 	printk("%ld pages of RAM\n", totalram_pages);
 	printk("%d free pages\n", nr_free_pages());
+#if 0 /* undefined pgtable_cache_size, pgd_cache_size */
 	printk("%ld pages in page table cache\n",pgtable_cache_size);
 #ifndef CONFIG_SMP
 	if (sparc_cpu_model == sun4m || sparc_cpu_model == sun4d)
 		printk("%ld entries in page dir cache\n",pgd_cache_size);
 #endif	
+#endif
 }
 
 extern pgprot_t protection_map[16];
@@ -309,6 +312,23 @@ unsigned long __init bootmem_init(unsigned long *pages_avail)
 	return max_pfn;
 }
 
+/*
+ * check_pgt_cache
+ *
+ * This is called at the end of unmapping of VMA (zap_page_range),
+ * to rescan the page cache for architecture specific things,
+ * presumably something like sun4/sun4c PMEGs. Most architectures
+ * define check_pgt_cache empty.
+ *
+ * We simply copy the 2.4 implementation for now.
+ */
+int pgt_cache_water[2] = { 25, 50 };
+
+void check_pgt_cache(void)
+{
+	do_check_pgt_cache(pgt_cache_water[0], pgt_cache_water[1]);
+}
+
 /*
  * paging_init() sets up the page tables: We call the MMU specific
  * init routine based upon the Sun model type on the Sparc.

arch/sparc/mm/io-unit.c

@@ -9,6 +9,9 @@
 #include <linux/init.h>
 #include <linux/slab.h>
 #include <linux/spinlock.h>
+#include <linux/mm.h>
+#include <linux/highmem.h>	/* pte_offset_map => kmap_atomic */
+
 #include <asm/scatterlist.h>
 #include <asm/pgalloc.h>
 #include <asm/pgtable.h>
@@ -17,6 +20,8 @@
 #include <asm/io-unit.h>
 #include <asm/mxcc.h>
 #include <asm/bitops.h>
+#include <asm/cacheflush.h>
+#include <asm/tlbflush.h>
 
 /* #define IOUNIT_DEBUG */
 #ifdef IOUNIT_DEBUG
@@ -188,7 +193,7 @@ static void iounit_map_dma_area(unsigned long va, __u32 addr, int len)
 
 			pgdp = pgd_offset(init_task.mm, addr);
 			pmdp = pmd_offset(pgdp, addr);
-			ptep = pte_offset(pmdp, addr);
+			ptep = pte_offset_map(pmdp, addr);
 
 			set_pte(ptep, pte_val(mk_pte(virt_to_page(page), dvma_prot)));
 			

arch/sparc/mm/iommu.c

-/* $Id: iommu.c,v 1.22 2001/12/17 07:05:09 davem Exp $
+/*
  * iommu.c:  IOMMU specific routines for memory management.
  *
  * Copyright (C) 1995 David S. Miller  (davem@caip.rutgers.edu)
- * Copyright (C) 1995 Pete Zaitcev
+ * Copyright (C) 1995,2002 Pete Zaitcev     (zaitcev@yahoo.com)
  * Copyright (C) 1996 Eddie C. Dost    (ecd@skynet.be)
  * Copyright (C) 1997,1998 Jakub Jelinek    (jj@sunsite.mff.cuni.cz)
  */
@@ -12,6 +12,8 @@
 #include <linux/init.h>
 #include <linux/mm.h>
 #include <linux/slab.h>
+#include <linux/highmem.h>	/* pte_offset_map => kmap_atomic */
+
 #include <asm/scatterlist.h>
 #include <asm/pgalloc.h>
 #include <asm/pgtable.h>
@@ -19,6 +21,8 @@
 #include <asm/io.h>
 #include <asm/mxcc.h>
 #include <asm/mbus.h>
+#include <asm/cacheflush.h>
+#include <asm/tlbflush.h>
 
 /* srmmu.c */
 extern int viking_mxcc_present;
@@ -245,7 +249,8 @@ static void iommu_map_dma_area(unsigned long va, __u32 addr, int len)
 
 			pgdp = pgd_offset(&init_mm, addr);
 			pmdp = pmd_offset(pgdp, addr);
-			ptep = pte_offset(pmdp, addr);
+			ptep = pte_offset_map(pmdp, addr);
+			/* XXX What if we run out of atomic maps above */
 
 			set_pte(ptep, mk_pte(virt_to_page(page), dvma_prot));
 			if (ipte_cache != 0) {

arch/sparc/mm/srmmu.c

-/* $Id: srmmu.c,v 1.234 2001/12/21 04:56:15 davem Exp $
+/*
  * srmmu.c:  SRMMU specific routines for memory management.
  *
  * Copyright (C) 1995 David S. Miller  (davem@caip.rutgers.edu)
- * Copyright (C) 1995 Pete Zaitcev
+ * Copyright (C) 1995,2002 Pete Zaitcev (zaitcev@yahoo.com)
  * Copyright (C) 1996 Eddie C. Dost    (ecd@skynet.be)
  * Copyright (C) 1997,1998 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
  * Copyright (C) 1999,2000 Anton Blanchard (anton@samba.org)
@@ -132,11 +132,32 @@ spinlock_t srmmu_nocache_spinlock;
 static inline unsigned long srmmu_pgd_page(pgd_t pgd)
 { return srmmu_device_memory(pgd_val(pgd))?~0:(unsigned long)__nocache_va((pgd_val(pgd) & SRMMU_PTD_PMASK) << 4); }
 
-static inline unsigned long srmmu_pmd_page(pmd_t pmd)
-{ return srmmu_device_memory(pmd_val(pmd))?~0:(unsigned long)__nocache_va((pmd_val(pmd) & SRMMU_PTD_PMASK) << 4); }
+static inline unsigned long srmmu_pmd_page_kernel(pmd_t pmd)
+{
+	return (unsigned long)
+	    __nocache_va((pmd_val(pmd) & SRMMU_PTD_PMASK) << 4);
+}
+
+static struct page *srmmu_pmd_page(pmd_t pmd)	/* XXX inline later */
+{
+
+	if (srmmu_device_memory(pmd_val(pmd))) {
+		/* XXX Anton obviously had something in mind when he did this.
+		 * But what?
+		 */
+		/* return (struct page *)~0; */
+		BUG();	/* NO WAY */
+	}
+	return virt_to_page(srmmu_pmd_page_kernel(pmd));
+}
 
-static inline struct page *srmmu_pte_page(pte_t pte)
-{ return (mem_map + (unsigned long)(srmmu_device_memory(pte_val(pte))?~0:(((pte_val(pte) & SRMMU_PTE_PMASK) << 4) >> PAGE_SHIFT))); }
+static inline unsigned long srmmu_pte_pfn(pte_t pte)
+{
+	if (srmmu_device_memory(pte_val(pte)))
+		BUG();
+	return (unsigned long)
+	    (((pte_val(pte) & SRMMU_PTE_PMASK) << 4) >> PAGE_SHIFT);
+}
 
 static inline int srmmu_pte_none(pte_t pte)
 { return !(pte_val(pte) & 0xFFFFFFF); }
@@ -219,7 +240,16 @@ static inline void srmmu_pgd_set(pgd_t * pgdp, pmd_t * pmdp)
 { srmmu_set_pte((pte_t *)pgdp, (SRMMU_ET_PTD | (__nocache_pa((unsigned long) pmdp) >> 4))); }
 
 static inline void srmmu_pmd_set(pmd_t * pmdp, pte_t * ptep)
-{ srmmu_set_pte((pte_t *)pmdp, (SRMMU_ET_PTD | (__nocache_pa((unsigned long) ptep) >> 4))); }
+{
+	srmmu_set_pte((pte_t *)pmdp,
+	    (SRMMU_ET_PTD | (__nocache_pa((unsigned long) ptep) >> 4)));
+}
+
+static inline void srmmu_pmd_populate(pmd_t * pmdp, struct page * ptep)
+{
+	srmmu_set_pte((pte_t *)pmdp,
+	    (SRMMU_ET_PTD | (((ptep - mem_map) << PAGE_SHIFT) >> 4)));
+}
 
 static inline pte_t srmmu_pte_modify(pte_t pte, pgprot_t newprot)
 { return __pte((pte_val(pte) & SRMMU_CHG_MASK) | pgprot_val(newprot)); }
@@ -234,7 +264,13 @@ static inline pmd_t *srmmu_pmd_offset(pgd_t * dir, unsigned long address)
 
 /* Find an entry in the third-level page table.. */ 
 static inline pte_t *srmmu_pte_offset(pmd_t * dir, unsigned long address)
-{ return (pte_t *) srmmu_pmd_page(*dir) + ((address >> PAGE_SHIFT) & (SRMMU_PTRS_PER_PTE - 1)); }
+{
+	unsigned long pte;
+
+	pte = srmmu_pmd_page_kernel(*dir);
+	return (pte_t *) pte +
+			((address >> PAGE_SHIFT) & (SRMMU_PTRS_PER_PTE - 1));
+}
 
 unsigned long __srmmu_get_nocache(int size, int align)
 {
@@ -387,14 +423,14 @@ static void srmmu_free_pgd_fast(pgd_t *pgd)
 	srmmu_free_nocache((unsigned long)pgd, SRMMU_PGD_TABLE_SIZE);
 }
 
-static pte_t *srmmu_pte_alloc_one_fast(struct mm_struct *mm, unsigned long address)
+static pte_t *srmmu_pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
 {
 	return (pte_t *)srmmu_get_nocache(SRMMU_PTE_TABLE_SIZE, SRMMU_PTE_TABLE_SIZE);
 }
 
-static pte_t *srmmu_pte_alloc_one(struct mm_struct *mm, unsigned long address)
+static struct page *srmmu_pte_alloc_one(struct mm_struct *mm, unsigned long address)
 {
-	return NULL;
+	return virt_to_page(srmmu_pte_alloc_one_kernel(mm, address));
 }
 
 static void srmmu_free_pte_fast(pte_t *pte)
@@ -402,6 +438,11 @@ static void srmmu_free_pte_fast(pte_t *pte)
 	srmmu_free_nocache((unsigned long)pte, SRMMU_PTE_TABLE_SIZE);
 }
 
+static void srmmu_pte_free(struct page *pte)
+{
+	srmmu_free_nocache((unsigned long)page_address(pte), SRMMU_PTE_TABLE_SIZE);
+}
+
 static pmd_t *srmmu_pmd_alloc_one_fast(struct mm_struct *mm, unsigned long address)
 {
 	return (pmd_t *)srmmu_get_nocache(SRMMU_PMD_TABLE_SIZE, SRMMU_PMD_TABLE_SIZE);
@@ -517,19 +558,15 @@ void srmmu_unmapioaddr(unsigned long virt_addr)
  * mappings on the kernel stack without any special code as we did
  * need on the sun4c.
  */
-struct task_struct *srmmu_alloc_task_struct(void)
-{
-	return (struct task_struct *) __get_free_pages(GFP_KERNEL, 1);
-}
-
-static void srmmu_free_task_struct(struct task_struct *tsk)
+struct thread_info *srmmu_alloc_thread_info(void)
 {
-	free_pages((unsigned long)tsk, 1);
+	return (struct thread_info *)
+	    __get_free_pages(GFP_KERNEL, THREAD_INFO_ORDER);
 }
 
-static void srmmu_get_task_struct(struct task_struct *tsk)
+static void srmmu_free_thread_info(struct thread_info *ti)
 {
-	atomic_inc(&virt_to_page(tsk)->count);
+	free_pages((unsigned long)ti, THREAD_INFO_ORDER);
 }
 
 /* tsunami.S */
@@ -1170,8 +1207,8 @@ void __init srmmu_paging_init(void)
 	srmmu_allocate_ptable_skeleton(PKMAP_BASE, PKMAP_BASE_END);
 
 	pgd = pgd_offset_k(PKMAP_BASE);
-	pmd = pmd_offset(pgd, PKMAP_BASE);
-	pte = pte_offset(pmd, PKMAP_BASE);
+	pmd = srmmu_pmd_offset(pgd, PKMAP_BASE);
+	pte = srmmu_pte_offset(pmd, PKMAP_BASE);
 	pkmap_page_table = pte;
 
 	flush_cache_all();
@@ -1209,6 +1246,10 @@ void __init srmmu_paging_init(void)
 		free_area_init_node(0, NULL, NULL, zones_size,
 				    phys_base, zholes_size);
 	}
+
+/* P3: easy to fix, todo. Current code is utterly broken, though. */
+	if (phys_base != 0)
+		panic("phys_base nonzero");
 }
 
 static void srmmu_mmu_info(struct seq_file *m)
@@ -1282,7 +1323,7 @@ static void __init init_vac_layout(void)
 			if(vac_line_size < min_line_size)
 				min_line_size = vac_line_size;
 			cpu++;
-			if(cpu == smp_num_cpus)
+			if (cpu >= NR_CPUS || !cpu_online(cpu))
 				break;
 #else
 			break;
@@ -1944,9 +1985,8 @@ static void __init get_srmmu_type(void)
 }
 
 /* dont laugh, static pagetables */
-static int srmmu_check_pgt_cache(int low, int high)
+static void srmmu_check_pgt_cache(int low, int high)
 {
-	return 0;
 }
 
 extern unsigned long spwin_mmu_patchme, fwin_mmu_patchme,
@@ -2017,16 +2057,16 @@ void __init ld_mmu_srmmu(void)
 #ifndef CONFIG_SMP	
 	BTFIXUPSET_CALL(___xchg32, ___xchg32_sun4md, BTFIXUPCALL_SWAPG1G2);
 #endif
-	BTFIXUPSET_CALL(do_check_pgt_cache, srmmu_check_pgt_cache, BTFIXUPCALL_NORM);
+	BTFIXUPSET_CALL(do_check_pgt_cache, srmmu_check_pgt_cache, BTFIXUPCALL_NOP);
 
 	BTFIXUPSET_CALL(set_pte, srmmu_set_pte, BTFIXUPCALL_SWAPO0O1);
 	BTFIXUPSET_CALL(switch_mm, srmmu_switch_mm, BTFIXUPCALL_NORM);
 
-	BTFIXUPSET_CALL(pte_page, srmmu_pte_page, BTFIXUPCALL_NORM);
+	BTFIXUPSET_CALL(pte_pfn, srmmu_pte_pfn, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(pmd_page, srmmu_pmd_page, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(pgd_page, srmmu_pgd_page, BTFIXUPCALL_NORM);
 
-	BTFIXUPSET_SETHI(none_mask, 0xF0000000); /* XXX P3: is it used? */
+	BTFIXUPSET_SETHI(none_mask, 0xF0000000);
 
 	BTFIXUPSET_CALL(pte_present, srmmu_pte_present, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(pte_clear, srmmu_pte_clear, BTFIXUPCALL_SWAPO0G0);
@@ -2041,16 +2081,18 @@ void __init ld_mmu_srmmu(void)
 	BTFIXUPSET_CALL(pgd_clear, srmmu_pgd_clear, BTFIXUPCALL_SWAPO0G0);
 
 	BTFIXUPSET_CALL(mk_pte, srmmu_mk_pte, BTFIXUPCALL_NORM);
-	BTFIXUPSET_CALL(pfn_pte, srmmu_pfn_pte, BTFIXUPCALL_NORM);
+	BTFIXUPSET_CALL(mk_pte_phys, srmmu_mk_pte_phys, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(mk_pte_io, srmmu_mk_pte_io, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(pgd_set, srmmu_pgd_set, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(pmd_set, srmmu_pmd_set, BTFIXUPCALL_NORM);
+	BTFIXUPSET_CALL(pmd_populate, srmmu_pmd_populate, BTFIXUPCALL_NORM);
 	
 	BTFIXUPSET_INT(pte_modify_mask, SRMMU_CHG_MASK);
 	BTFIXUPSET_CALL(pmd_offset, srmmu_pmd_offset, BTFIXUPCALL_NORM);
-	BTFIXUPSET_CALL(pte_offset, srmmu_pte_offset, BTFIXUPCALL_NORM);
+	BTFIXUPSET_CALL(pte_offset_kernel, srmmu_pte_offset, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(free_pte_fast, srmmu_free_pte_fast, BTFIXUPCALL_NORM);
-	BTFIXUPSET_CALL(pte_alloc_one_fast, srmmu_pte_alloc_one_fast, BTFIXUPCALL_NORM);
+	BTFIXUPSET_CALL(pte_free, srmmu_pte_free, BTFIXUPCALL_NORM);
+	BTFIXUPSET_CALL(pte_alloc_one_kernel, srmmu_pte_alloc_one_kernel, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(pte_alloc_one, srmmu_pte_alloc_one, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(free_pmd_fast, srmmu_free_pmd_fast, BTFIXUPCALL_NORM);
 	BTFIXUPSET_CALL(pmd_alloc_one_fast, srmmu_pmd_alloc_one_fast, BTFIXUPCALL_NORM);
@@ -2071,10 +2113,8 @@ void __init ld_mmu_srmmu(void)
 	
 	BTFIXUPSET_CALL(mmu_info, srmmu_mmu_info, BTFIXUPCALL_NORM);
 
-	/* Task struct and kernel stack allocating/freeing. */
-	BTFIXUPSET_CALL(alloc_task_struct, srmmu_alloc_task_struct, BTFIXUPCALL_NORM);
-	BTFIXUPSET_CALL(free_task_struct, srmmu_free_task_struct, BTFIXUPCALL_NORM);
-	BTFIXUPSET_CALL(get_task_struct, srmmu_get_task_struct, BTFIXUPCALL_NORM);
+	BTFIXUPSET_CALL(alloc_thread_info, srmmu_alloc_thread_info, BTFIXUPCALL_NORM);
+	BTFIXUPSET_CALL(free_thread_info, srmmu_free_thread_info, BTFIXUPCALL_NORM);
 
 	get_srmmu_type();
 	patch_window_trap_handlers();

drivers/sbus/sbus.c

@@ -16,6 +16,7 @@
 #include <asm/oplib.h>
 #include <asm/bpp.h>
 #include <asm/irq.h>
+#include <asm/pcic.h>		/* pcic_present */
 
 struct sbus_bus *sbus_root = NULL;
 
@@ -334,7 +335,7 @@ static int __init sbus_init(void)
 		   (nd = prom_getchild(iommund)) == 0 ||
 		   (nd = prom_searchsiblings(nd, "sbus")) == 0) {
 #ifdef CONFIG_PCI
-                        if (!pcibios_present()) {       
+                        if (!pcic_present()) {
                                 prom_printf("Neither SBUS nor PCI found.\n");
                                 prom_halt();
                         }

include/asm-sparc/bitops.h

@@ -207,12 +207,74 @@ static __inline__ unsigned long ffz(unsigned long word)
 	return result;
 }
 
+/**
+ * __ffs - find first bit in word.
+ * @word: The word to search
+ *
+ * Undefined if no bit exists, so code should check against 0 first.
+ */
+static __inline__ int __ffs(unsigned long word)
+{
+	int num = 0;
+
+	if ((word & 0xffff) == 0) {
+		num += 16;
+		word >>= 16;
+	}
+	if ((word & 0xff) == 0) {
+		num += 8;
+		word >>= 8;
+	}
+	if ((word & 0xf) == 0) {
+		num += 4;
+		word >>= 4;
+	}
+	if ((word & 0x3) == 0) {
+		num += 2;
+		word >>= 2;
+	}
+	if ((word & 0x1) == 0)
+		num += 1;
+	return num;
+}
+
+/*
+ * Every architecture must define this function. It's the fastest
+ * way of searching a 140-bit bitmap where the first 100 bits are
+ * unlikely to be set. It's guaranteed that at least one of the 140
+ * bits is cleared.
+ */
+static __inline__ int sched_find_first_bit(unsigned long *b)
+{
+
+	if (unlikely(b[0]))
+		return __ffs(b[0]);
+	if (unlikely(b[1]))
+		return __ffs(b[1]) + 32;
+	if (unlikely(b[2]))
+		return __ffs(b[2]) + 64;
+	if (b[3])
+		return __ffs(b[3]) + 96;
+	return __ffs(b[4]) + 128;
+}
+
 /*
  * ffs: find first bit set. This is defined the same way as
  * the libc and compiler builtin ffs routines, therefore
  * differs in spirit from the above ffz (man ffs).
  */
-#define ffs(x) generic_ffs(x)
+static __inline__ int ffs(int x)
+{
+	if (!x)
+		return 0;
+	return __ffs((unsigned long)x);
+}
+
+/*
+ * fls: find last (most-significant) bit set.
+ * Note fls(0) = 0, fls(1) = 1, fls(0x80000000) = 32.
+ */
+#define fls(x) generic_fls(x)
 
 /*
  * hweightN: returns the hamming weight (i.e. the number
@@ -272,6 +334,34 @@ static __inline__ unsigned long find_next_zero_bit(void *addr, unsigned long siz
 #define find_first_zero_bit(addr, size) \
         find_next_zero_bit((addr), (size), 0)
 
+/**
+ * find_next_bit - find the first set bit in a memory region
+ * @addr: The address to base the search on
+ * @offset: The bitnumber to start searching at
+ * @size: The maximum size to search
+ *
+ * Scheduler induced bitop, do not use.
+ */
+static __inline__ int find_next_bit(unsigned long *addr, int size, int offset)
+{
+	unsigned long *p = addr + (offset >> 5);
+	int num = offset & ~0x1f;
+	unsigned long word;
+
+	word = *p++;
+	word &= ~((1 << (offset & 0x1f)) - 1);
+	while (num < size) {
+		if (word != 0) {
+			return __ffs(word) + num;
+		}
+		word = *p++;
+		num += 0x20;
+	}
+	return num;
+}
+
+/*
+ */
 static __inline__ int test_le_bit(int nr, __const__ void * addr)
 {
 	__const__ unsigned char *ADDR = (__const__ unsigned char *) addr;

include/asm-sparc/cacheflush.h

+#ifndef _SPARC_CACHEFLUSH_H
+#define _SPARC_CACHEFLUSH_H
+
+#include <linux/config.h>
+#include <linux/mm.h>		/* Common for other includes */
+// #include <linux/kernel.h> from pgalloc.h
+// #include <linux/sched.h>  from pgalloc.h
+
+// #include <asm/page.h>
+#include <asm/btfixup.h>
+
+/*
+ * Fine grained cache flushing.
+ */
+#ifdef CONFIG_SMP
+
+BTFIXUPDEF_CALL(void, local_flush_cache_all, void)
+BTFIXUPDEF_CALL(void, local_flush_cache_mm, struct mm_struct *)
+BTFIXUPDEF_CALL(void, local_flush_cache_range, struct vm_area_struct *, unsigned long, unsigned long)
+BTFIXUPDEF_CALL(void, local_flush_cache_page, struct vm_area_struct *, unsigned long)
+
+#define local_flush_cache_all() BTFIXUP_CALL(local_flush_cache_all)()
+#define local_flush_cache_mm(mm) BTFIXUP_CALL(local_flush_cache_mm)(mm)
+#define local_flush_cache_range(vma,start,end) BTFIXUP_CALL(local_flush_cache_range)(vma,start,end)
+#define local_flush_cache_page(vma,addr) BTFIXUP_CALL(local_flush_cache_page)(vma,addr)
+
+BTFIXUPDEF_CALL(void, local_flush_page_to_ram, unsigned long)
+BTFIXUPDEF_CALL(void, local_flush_sig_insns, struct mm_struct *, unsigned long)
+
+#define local_flush_page_to_ram(addr) BTFIXUP_CALL(local_flush_page_to_ram)(addr)
+#define local_flush_sig_insns(mm,insn_addr) BTFIXUP_CALL(local_flush_sig_insns)(mm,insn_addr)
+
+extern void smp_flush_cache_all(void);
+extern void smp_flush_cache_mm(struct mm_struct *mm);
+extern void smp_flush_cache_range(struct vm_area_struct *vma,
+				  unsigned long start,
+				  unsigned long end);
+extern void smp_flush_cache_page(struct vm_area_struct *vma, unsigned long page);
+
+extern void smp_flush_page_to_ram(unsigned long page);
+extern void smp_flush_sig_insns(struct mm_struct *mm, unsigned long insn_addr);
+
+#endif /* CONFIG_SMP */
+
+BTFIXUPDEF_CALL(void, flush_cache_all, void)
+BTFIXUPDEF_CALL(void, flush_cache_mm, struct mm_struct *)
+BTFIXUPDEF_CALL(void, flush_cache_range, struct vm_area_struct *, unsigned long, unsigned long)
+BTFIXUPDEF_CALL(void, flush_cache_page, struct vm_area_struct *, unsigned long)
+
+#define flush_cache_all() BTFIXUP_CALL(flush_cache_all)()
+#define flush_cache_mm(mm) BTFIXUP_CALL(flush_cache_mm)(mm)
+#define flush_cache_range(vma,start,end) BTFIXUP_CALL(flush_cache_range)(vma,start,end)
+#define flush_cache_page(vma,addr) BTFIXUP_CALL(flush_cache_page)(vma,addr)
+#define flush_icache_range(start, end)		do { } while (0)
+#define flush_icache_page(vma, pg)		do { } while (0)
+
+#define flush_icache_user_range(vma,pg,adr,len)	do { } while (0)
+
+BTFIXUPDEF_CALL(void, __flush_page_to_ram, unsigned long)
+BTFIXUPDEF_CALL(void, flush_sig_insns, struct mm_struct *, unsigned long)
+
+#define __flush_page_to_ram(addr) BTFIXUP_CALL(__flush_page_to_ram)(addr)
+#define flush_sig_insns(mm,insn_addr) BTFIXUP_CALL(flush_sig_insns)(mm,insn_addr)
+
+extern void flush_page_to_ram(struct page *page);
+
+#define flush_dcache_page(page)			do { } while (0)
+
+#endif /* _SPARC_CACHEFLUSH_H */

include/asm-sparc/current.h

-#ifndef _SPARC_CURRENT_H
-#define _SPARC_CURRENT_H
+/*
+ *  include/asm-sparc/current.h
+ *
+ * Copyright (C) 1999 IBM Deutschland Entwicklung GmbH, IBM Corporation
+ * Copyright (C) 2002 Pete Zaitcev (zaitcev@yahoo.com)
+ *
+ *  Derived from "include/asm-s390/current.h" by
+ *  Martin Schwidefsky (schwidefsky@de.ibm.com)
+ *  Derived from "include/asm-i386/current.h"
+ */
+#ifndef _ASM_CURRENT_H
+#define _ASM_CURRENT_H
 
-/* Sparc rules... */
-register struct task_struct *current asm("g6");
+/*
+ * At the sparc64 DaveM keeps current_thread_info in %g4.
+ * We might want to consider doing the same to shave a few cycles.
+ */
 
-#endif /* !(_SPARC_CURRENT_H) */
+// Applications do not include kernel headers anymore, period.
+// #ifdef __KERNEL__
+
+#ifndef _ASM_THREAD_INFO_H
+#include <asm/thread_info.h>
+#endif
+
+struct task_struct;
+
+/* Two stage process (inline + #define) for type-checking. */
+/* We also obfuscate get_current() to check if anyone used that by mistake. */
+static inline struct task_struct *__get_current(void)
+{
+	return current_thread_info()->task;
+}
+#define current __get_current()
+
+// #endif /* __KERNEL__ */
+
+#endif /* !(_ASM_CURRENT_H) */

include/asm-sparc/elf.h

@@ -41,7 +41,7 @@ do {	unsigned long *dest = &(__elf_regs[0]);		\
 	dest[34] = src->npc;				\
 	dest[35] = src->y;				\
 	dest[36] = dest[37] = 0; /* XXX */		\
-} while(0)
+} while(0); /* Janitors: Don't touch this colon. */
 
 typedef struct {
 	union {

include/asm-sparc/highmem.h

@@ -25,6 +25,8 @@
 #include <asm/vaddrs.h>
 #include <asm/kmap_types.h>
 #include <asm/pgtable.h>
+#include <asm/cacheflush.h>
+#include <asm/tlbflush.h>
 
 /* undef for production */
 #define HIGHMEM_DEBUG 1

include/asm-sparc/kmap_types.h

@@ -9,6 +9,8 @@ enum km_type {
 	KM_USER1,
 	KM_BIO_SRC_IRQ,
 	KM_BIO_DST_IRQ,
+	KM_PTE0,
+	KM_PTE1,
 	KM_TYPE_NR
 };
 

include/asm-sparc/mmu_context.h

@@ -5,29 +5,6 @@
 
 #ifndef __ASSEMBLY__
 
-/*
- * Every architecture must define this function. It's the fastest
- * way of searching a 168-bit bitmap where the first 128 bits are
- * unlikely to be clear. It's guaranteed that at least one of the 168
- * bits is cleared.
- */
-#if MAX_RT_PRIO != 128 || MAX_PRIO != 168
-# error update this function.
-#endif
-
-static inline int sched_find_first_zero_bit(unsigned long *b)
-{
-	unsigned int rt;
-
-	rt = b[0] & b[1] & b[2] & b[3];
-	if (unlikely(rt != 0xffffffff))
-		return find_first_zero_bit(b, MAX_RT_PRIO);
-
-	if (b[4] != ~0)
-		return ffz(b[4]) + MAX_RT_PRIO;
-	return ffz(b[5]) + 32 + MAX_RT_PRIO;
-}
-
 static inline void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk, unsigned cpu)
 {
 }

include/asm-sparc/page.h

@@ -27,9 +27,6 @@
 #include <asm/head.h>       /* for KERNBASE */
 #include <asm/btfixup.h>
 
-/* This is always 2048*sizeof(long), doesn't change with PAGE_SIZE */
-#define TASK_UNION_SIZE		8192
-
 #ifndef __ASSEMBLY__
 
 /*
@@ -176,8 +173,12 @@ extern __inline__ int get_order(unsigned long size)
 #define PAGE_OFFSET	0xf0000000
 #define __pa(x)                 ((unsigned long)(x) - PAGE_OFFSET)
 #define __va(x)                 ((void *)((unsigned long) (x) + PAGE_OFFSET))
+
+#define pfn_to_page(pfn)        (mem_map + (pfn))
+#define page_to_pfn(page)       ((unsigned long)((page) - mem_map))
 #define virt_to_page(kaddr)	(mem_map + (__pa(kaddr) >> PAGE_SHIFT))
-#define VALID_PAGE(page)	((page - mem_map) < max_mapnr)
+#define pfn_valid(pfn)		((pfn) < max_mapnr)
+#define virt_addr_valid(kaddr)	pfn_valid(__pa(kaddr) >> PAGE_SHIFT)
 
 #define VM_DATA_DEFAULT_FLAGS	(VM_READ | VM_WRITE | VM_EXEC | \
 				 VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)

include/asm-sparc/pgalloc.h

@@ -9,85 +9,7 @@
 #include <asm/page.h>
 #include <asm/btfixup.h>
 
-/* Fine grained cache/tlb flushing. */
-#ifdef CONFIG_SMP
-BTFIXUPDEF_CALL(void, local_flush_cache_all, void)
-BTFIXUPDEF_CALL(void, local_flush_cache_mm, struct mm_struct *)
-BTFIXUPDEF_CALL(void, local_flush_cache_range, struct vm_area_struct *, unsigned long, unsigned long)
-BTFIXUPDEF_CALL(void, local_flush_cache_page, struct vm_area_struct *, unsigned long)
-
-#define local_flush_cache_all() BTFIXUP_CALL(local_flush_cache_all)()
-#define local_flush_cache_mm(mm) BTFIXUP_CALL(local_flush_cache_mm)(mm)
-#define local_flush_cache_range(vma,start,end) BTFIXUP_CALL(local_flush_cache_range)(vma,start,end)
-#define local_flush_cache_page(vma,addr) BTFIXUP_CALL(local_flush_cache_page)(vma,addr)
-
-BTFIXUPDEF_CALL(void, local_flush_tlb_all, void)
-BTFIXUPDEF_CALL(void, local_flush_tlb_mm, struct mm_struct *)
-BTFIXUPDEF_CALL(void, local_flush_tlb_range, struct vm_area_struct *, unsigned long, unsigned long)
-BTFIXUPDEF_CALL(void, local_flush_tlb_page, struct vm_area_struct *, unsigned long)
-
-#define local_flush_tlb_all() BTFIXUP_CALL(local_flush_tlb_all)()
-#define local_flush_tlb_mm(mm) BTFIXUP_CALL(local_flush_tlb_mm)(mm)
-#define local_flush_tlb_range(vma,start,end) BTFIXUP_CALL(local_flush_tlb_range)(vma,start,end)
-#define local_flush_tlb_page(vma,addr) BTFIXUP_CALL(local_flush_tlb_page)(vma,addr)
-
-BTFIXUPDEF_CALL(void, local_flush_page_to_ram, unsigned long)
-BTFIXUPDEF_CALL(void, local_flush_sig_insns, struct mm_struct *, unsigned long)
-
-#define local_flush_page_to_ram(addr) BTFIXUP_CALL(local_flush_page_to_ram)(addr)
-#define local_flush_sig_insns(mm,insn_addr) BTFIXUP_CALL(local_flush_sig_insns)(mm,insn_addr)
-
-extern void smp_flush_cache_all(void);
-extern void smp_flush_cache_mm(struct mm_struct *mm);
-extern void smp_flush_cache_range(struct vm_area_struct *vma,
-				  unsigned long start,
-				  unsigned long end);
-extern void smp_flush_cache_page(struct vm_area_struct *vma, unsigned long page);
-
-extern void smp_flush_tlb_all(void);
-extern void smp_flush_tlb_mm(struct mm_struct *mm);
-extern void smp_flush_tlb_range(struct vm_area_struct *vma,
-				  unsigned long start,
-				  unsigned long end);
-extern void smp_flush_tlb_page(struct vm_area_struct *mm, unsigned long page);
-extern void smp_flush_page_to_ram(unsigned long page);
-extern void smp_flush_sig_insns(struct mm_struct *mm, unsigned long insn_addr);
-#endif
-
-BTFIXUPDEF_CALL(void, flush_cache_all, void)
-BTFIXUPDEF_CALL(void, flush_cache_mm, struct mm_struct *)
-BTFIXUPDEF_CALL(void, flush_cache_range, struct vm_area_struct *, unsigned long, unsigned long)
-BTFIXUPDEF_CALL(void, flush_cache_page, struct vm_area_struct *, unsigned long)
-
-#define flush_cache_all() BTFIXUP_CALL(flush_cache_all)()
-#define flush_cache_mm(mm) BTFIXUP_CALL(flush_cache_mm)(mm)
-#define flush_cache_range(vma,start,end) BTFIXUP_CALL(flush_cache_range)(vma,start,end)
-#define flush_cache_page(vma,addr) BTFIXUP_CALL(flush_cache_page)(vma,addr)
-#define flush_icache_range(start, end)		do { } while (0)
-
-BTFIXUPDEF_CALL(void, flush_tlb_all, void)
-BTFIXUPDEF_CALL(void, flush_tlb_mm, struct mm_struct *)
-BTFIXUPDEF_CALL(void, flush_tlb_range, struct vm_area_struct *, unsigned long, unsigned long)
-BTFIXUPDEF_CALL(void, flush_tlb_page, struct vm_area_struct *, unsigned long)
-
-extern __inline__ void flush_tlb_pgtables(struct mm_struct *mm, unsigned long start, unsigned long end)
-{
-}
-
-#define flush_tlb_all() BTFIXUP_CALL(flush_tlb_all)()
-#define flush_tlb_mm(mm) BTFIXUP_CALL(flush_tlb_mm)(mm)
-#define flush_tlb_range(vma,start,end) BTFIXUP_CALL(flush_tlb_range)(vma,start,end)
-#define flush_tlb_page(vma,addr) BTFIXUP_CALL(flush_tlb_page)(vma,addr)
-
-BTFIXUPDEF_CALL(void, __flush_page_to_ram, unsigned long)
-BTFIXUPDEF_CALL(void, flush_sig_insns, struct mm_struct *, unsigned long)
-
-#define __flush_page_to_ram(addr) BTFIXUP_CALL(__flush_page_to_ram)(addr)
-#define flush_sig_insns(mm,insn_addr) BTFIXUP_CALL(flush_sig_insns)(mm,insn_addr)
-
-extern void flush_page_to_ram(struct page *page);
-
-#define flush_dcache_page(page)			do { } while (0)
+struct page;
 
 extern struct pgtable_cache_struct {
 	unsigned long *pgd_cache;
@@ -101,7 +23,8 @@ extern struct pgtable_cache_struct {
 #define pgtable_cache_size      (pgt_quicklists.pgtable_cache_sz)
 #define pgd_cache_size		(pgt_quicklists.pgd_cache_sz)
 
-BTFIXUPDEF_CALL(int,	 do_check_pgt_cache, int, int)
+extern void check_pgt_cache(void);
+BTFIXUPDEF_CALL(void,	 do_check_pgt_cache, int, int)
 #define do_check_pgt_cache(low,high) BTFIXUP_CALL(do_check_pgt_cache)(low,high)
 
 BTFIXUPDEF_CALL(pgd_t *, get_pgd_fast, void)
@@ -113,6 +36,8 @@ BTFIXUPDEF_CALL(void, free_pgd_fast, pgd_t *)
 #define pgd_free(pgd)	free_pgd_fast(pgd)
 #define pgd_alloc(mm)	get_pgd_fast()
 
+BTFIXUPDEF_CALL(void, pgd_set, pgd_t *, pmd_t *)
+#define pgd_set(pgdp,pmdp) BTFIXUP_CALL(pgd_set)(pgdp,pmdp)
 #define pgd_populate(MM, PGD, PMD)      pgd_set(PGD, PMD)
 
 static __inline__ pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long address)
@@ -127,18 +52,24 @@ BTFIXUPDEF_CALL(void, free_pmd_fast, pmd_t *)
 #define free_pmd_fast(pmd)	BTFIXUP_CALL(free_pmd_fast)(pmd)
 
 #define pmd_free(pmd)           free_pmd_fast(pmd)
+#define pmd_free_tlb(tlb, pmd)	pmd_free(pmd)
 
-#define pmd_populate(MM, PMD, PTE)      pmd_set(PMD, PTE)
+BTFIXUPDEF_CALL(void, pmd_populate, pmd_t *, struct page *)
+#define pmd_populate(MM, PMD, PTE)        BTFIXUP_CALL(pmd_populate)(PMD, PTE)
+BTFIXUPDEF_CALL(void, pmd_set, pmd_t *, pte_t *)
+#define pmd_populate_kernel(MM, PMD, PTE) BTFIXUP_CALL(pmd_set)(PMD, PTE)
 
-BTFIXUPDEF_CALL(pte_t *, pte_alloc_one, struct mm_struct *, unsigned long)
+BTFIXUPDEF_CALL(struct page *, pte_alloc_one, struct mm_struct *, unsigned long)
 #define pte_alloc_one(mm, address)	BTFIXUP_CALL(pte_alloc_one)(mm, address)
-
-BTFIXUPDEF_CALL(pte_t *, pte_alloc_one_fast, struct mm_struct *, unsigned long)
-#define pte_alloc_one_fast(mm, address)	BTFIXUP_CALL(pte_alloc_one_fast)(mm, address)
+BTFIXUPDEF_CALL(pte_t *, pte_alloc_one_kernel, struct mm_struct *, unsigned long)
+#define pte_alloc_one_kernel(mm, addr)	BTFIXUP_CALL(pte_alloc_one_kernel)(mm, addr)
 
 BTFIXUPDEF_CALL(void, free_pte_fast, pte_t *)
 #define free_pte_fast(pte)	BTFIXUP_CALL(free_pte_fast)(pte)
+#define pte_free_kernel(pte)	free_pte_fast(pte)
 
-#define pte_free(pte)		free_pte_fast(pte)
+BTFIXUPDEF_CALL(void, pte_free, struct page *)
+#define pte_free(pte)		BTFIXUP_CALL(pte_free)(pte)
+#define pte_free_tlb(tlb, pte)	pte_free(pte)
 
 #endif /* _SPARC_PGALLOC_H */

include/asm-sparc/pgtable.h

@@ -11,6 +11,8 @@
 
 #include <linux/config.h>
 #include <linux/spinlock.h>
+/* XXX This creates many nasty warnings. */
+/* #include <linux/highmem.h> */	/* kmap_atomic in pte_offset_map */
 #include <asm/asi.h>
 #ifdef CONFIG_SUN4
 #include <asm/pgtsun4.h>
@@ -26,6 +28,8 @@
 
 #ifndef __ASSEMBLY__
 
+struct vm_area_struct;
+
 extern void load_mmu(void);
 extern unsigned long calc_highpages(void);
 			       
@@ -183,7 +187,7 @@ extern unsigned long empty_zero_page;
 
 #define BAD_PAGETABLE __bad_pagetable()
 #define BAD_PAGE __bad_page()
-#define ZERO_PAGE(vaddr) (mem_map + (((unsigned long)&empty_zero_page - PAGE_OFFSET + phys_base) >> PAGE_SHIFT))
+#define ZERO_PAGE(vaddr) (virt_to_page(empty_zero_page))
 
 /* number of bits that fit into a memory pointer */
 #define BITS_PER_PTR      (8*sizeof(unsigned long))
@@ -193,7 +197,7 @@ extern unsigned long empty_zero_page;
 
 #define SIZEOF_PTR_LOG2   2
 
-BTFIXUPDEF_CALL_CONST(unsigned long, pmd_page, pmd_t)
+BTFIXUPDEF_CALL_CONST(struct page *, pmd_page, pmd_t)
 BTFIXUPDEF_CALL_CONST(unsigned long, pgd_page, pgd_t)
 
 #define pmd_page(pmd) BTFIXUP_CALL(pmd_page)(pmd)
@@ -291,13 +295,12 @@ BTFIXUPDEF_CALL_CONST(pte_t, pte_mkyoung, pte_t)
 #define pte_mkyoung(pte) BTFIXUP_CALL(pte_mkyoung)(pte)
 
 #define page_pte_prot(page, prot)	mk_pte(page, prot)
-#define page_pte(page)			page_pte_prot(page, __pgprot(0))
-
-/* Permanent address of a page. */
-#define page_address(page)  ((page)->virtual)
+#define page_pte(page)			mk_pte(page, __pgprot(0))
+#define pfn_pte(pfn, prot)		mk_pte(pfn_to_page(pfn), prot)
 
-BTFIXUPDEF_CALL(struct page *, pte_page, pte_t)
-#define pte_page(pte) BTFIXUP_CALL(pte_page)(pte)
+BTFIXUPDEF_CALL(unsigned long,	 pte_pfn, pte_t)
+#define pte_pfn(pte) BTFIXUP_CALL(pte_pfn)(pte)
+#define pte_page(pte)	pfn_to_page(pte_pfn(pte))
 
 /*
  * Conversion functions: convert a page and protection to a page entry,
@@ -312,12 +315,6 @@ BTFIXUPDEF_CALL_CONST(pte_t, mk_pte_io, unsigned long, pgprot_t, int)
 #define mk_pte_phys(page,pgprot) BTFIXUP_CALL(mk_pte_phys)(page,pgprot)
 #define mk_pte_io(page,pgprot,space) BTFIXUP_CALL(mk_pte_io)(page,pgprot,space)
 
-BTFIXUPDEF_CALL(void, pgd_set, pgd_t *, pmd_t *)
-BTFIXUPDEF_CALL(void, pmd_set, pmd_t *, pte_t *)
-
-#define pgd_set(pgdp,pmdp) BTFIXUP_CALL(pgd_set)(pgdp,pmdp)
-#define pmd_set(pmdp,ptep) BTFIXUP_CALL(pmd_set)(pmdp,ptep)
-
 BTFIXUPDEF_INT(pte_modify_mask)
 
 extern pte_t pte_modify(pte_t pte, pgprot_t newprot) __attribute__((const));
@@ -335,21 +332,32 @@ extern __inline__ pte_t pte_modify(pte_t pte, pgprot_t newprot)
 /* to find an entry in a kernel page-table-directory */
 #define pgd_offset_k(address) pgd_offset(&init_mm, address)
 
-BTFIXUPDEF_CALL(pmd_t *, pmd_offset, pgd_t *, unsigned long)
-BTFIXUPDEF_CALL(pte_t *, pte_offset, pmd_t *, unsigned long)
-
 /* Find an entry in the second-level page table.. */
+BTFIXUPDEF_CALL(pmd_t *, pmd_offset, pgd_t *, unsigned long)
 #define pmd_offset(dir,addr) BTFIXUP_CALL(pmd_offset)(dir,addr)
 
 /* Find an entry in the third-level page table.. */ 
-#define pte_offset(dir,addr) BTFIXUP_CALL(pte_offset)(dir,addr)
+BTFIXUPDEF_CALL(pte_t *, pte_offset_kernel, pmd_t *, unsigned long)
+#define pte_offset_kernel(dir,addr) BTFIXUP_CALL(pte_offset_kernel)(dir,addr)
+
+/* __pte_offset is not BTFIXUP-ed, but PTRS_PER_PTE is, so it's ok. */
+#define __pte_offset(address) \
+	(((address) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1))
+#if 0 /* XXX Should we expose pmd_page_kernel? */
+#define pte_offset_kernel(dir, addr) \
+	((pte_t *) pmd_page_kernel(*(dir)) + __pte_offset(addr))
+#endif
+#define pte_offset_map(dir, addr) \
+	((pte_t *) kmap_atomic(pmd_page(*(dir)), KM_PTE0) + __pte_offset(addr))
+#define pte_offset_map_nested(dir, addr) \
+	((pte_t *) kmap_atomic(pmd_page(*(dir)), KM_PTE1) + __pte_offset(addr))
+
+#define pte_unmap(pte)		kunmap_atomic(pte, KM_PTE0)
+#define pte_unmap_nested(pte)	kunmap_atomic(pte, KM_PTE1)
 
 /* The permissions for pgprot_val to make a page mapped on the obio space */
 extern unsigned int pg_iobits;
 
-#define flush_icache_page(vma, pg)      do { } while(0)
-#define flush_icache_user_range(vma,pg,adr,len)	do { } while (0)
-
 /* Certain architectures need to do special things when pte's
  * within a page table are directly modified.  Thus, the following
  * hook is made available.

include/asm-sparc/pgtsrmmu.h

@@ -8,6 +8,7 @@
 #define _SPARC_PGTSRMMU_H
 
 #include <asm/page.h>
+#include <asm/thread_info.h>	/* TI_UWINMASK for WINDOW_FLUSH */
 
 /* PMD_SHIFT determines the size of the area a second-level page table can map */
 #define SRMMU_PMD_SHIFT         18
@@ -87,7 +88,7 @@
 
 #define WINDOW_FLUSH(tmp1, tmp2)					\
 	mov	0, tmp1;						\
-98:	ld	[%g6 + AOFF_task_thread + AOFF_thread_uwinmask], tmp2;	\
+98:	ld	[%g6 + TI_UWINMASK], tmp2;				\
 	orcc	%g0, tmp2, %g0;						\
 	add	tmp1, 1, tmp1;						\
 	bne	98b;							\

include/asm-sparc/processor.h

@@ -44,6 +44,8 @@
  */
 #define TASK_SIZE	PAGE_OFFSET
 
+struct task_struct;
+
 struct fpq {
 	unsigned long *insn_addr;
 	unsigned long insn;
@@ -55,15 +57,8 @@ typedef struct {
 
 /* The Sparc processor specific thread struct. */
 struct thread_struct {
-	unsigned long uwinmask __attribute__ ((aligned (8)));
 	struct pt_regs *kregs;
 
-	/* Context switch saved kernel state. */
-	unsigned long ksp __attribute__ ((aligned (8)));
-	unsigned long kpc;
-	unsigned long kpsr;
-	unsigned long kwim;
-
 	/* Special child fork kpsr/kwim values. */
 	unsigned long fork_kpsr __attribute__ ((aligned (8)));
 	unsigned long fork_kwim;
@@ -92,8 +87,8 @@ struct thread_struct {
 #define SPARC_FLAG_UNALIGNED    0x2    /* is allowed to do unaligned accesses */
 
 #define INIT_THREAD  { \
-/* uwinmask, kregs, ksp, kpc, kpsr, kwim */ \
-   0,        0,     0,   0,   0,    0, \
+/* kregs, */ \
+   0,  \
 /* fork_kpsr, fork_kwim */ \
    0,         0, \
 /* reg_window */  \
@@ -115,10 +110,7 @@ struct thread_struct {
 }
 
 /* Return saved PC of a blocked thread. */
-extern __inline__ unsigned long thread_saved_pc(struct thread_struct *t)
-{
-	return t->kpc;
-}
+extern unsigned long thread_saved_pc(struct task_struct *t);
 
 /* Do necessary setup to start up a newly executed thread. */
 extern __inline__ void start_thread(struct pt_regs * regs, unsigned long pc,
@@ -163,7 +155,7 @@ extern pid_t kernel_thread(int (*fn)(void *), void * arg, unsigned long flags);
 	if (!(__TSK) || (__TSK) == current || \
             (__TSK)->state == TASK_RUNNING) \
 		goto __out; \
-	fp = (__TSK)->thread.ksp + bias; \
+	fp = (__TSK)->thread_info->ksp + bias; \
 	do { \
 		/* Bogus frame pointer? */ \
 		if (fp < (task_base + sizeof(struct task_struct)) || \
@@ -185,22 +177,9 @@ __out:	__ret; \
 #define KSTK_ESP(tsk)  ((tsk)->thread.kregs->u_regs[UREG_FP])
 
 #ifdef __KERNEL__
-#define THREAD_SIZE (2*PAGE_SIZE)
 
 extern struct task_struct *last_task_used_math;
 
-/* Allocation and freeing of basic task resources. */
-BTFIXUPDEF_CALL(struct task_struct *, alloc_task_struct, void)
-BTFIXUPDEF_CALL(void, free_task_struct, struct task_struct *)
-BTFIXUPDEF_CALL(void, get_task_struct, struct task_struct *)
-
-#define alloc_task_struct() BTFIXUP_CALL(alloc_task_struct)()
-#define free_task_struct(tsk) BTFIXUP_CALL(free_task_struct)(tsk)
-#define get_task_struct(tsk) BTFIXUP_CALL(get_task_struct)(tsk)
-
-#define init_task	(init_task_union.task)
-#define init_stack	(init_task_union.stack)
-
 #define cpu_relax()	do { } while (0)
 
 #endif

include/asm-sparc/spinlock.h

@@ -12,15 +12,7 @@
 
 #include <asm/psr.h>
 
-/*
- * Define this to use the verbose/debugging versions in
- * arch/sparc/lib/debuglocks.c
- *
- * Be sure to make dep whenever changing this option.
- */
-#define SPIN_LOCK_DEBUG
-
-#ifdef SPIN_LOCK_DEBUG
+#ifdef CONFIG_DEBUG_SPINLOCK
 struct _spinlock_debug {
 	unsigned char lock;
 	unsigned long owner_pc;
@@ -36,9 +28,9 @@ extern void _do_spin_lock(spinlock_t *lock, char *str);
 extern int _spin_trylock(spinlock_t *lock);
 extern void _do_spin_unlock(spinlock_t *lock);
 
-#define spin_trylock(lp)	_spin_trylock(lp)
-#define spin_lock(lock)		_do_spin_lock(lock, "spin_lock")
-#define spin_unlock(lock)	_do_spin_unlock(lock)
+#define _raw_spin_trylock(lp)	_spin_trylock(lp)
+#define _raw_spin_lock(lock)	_do_spin_lock(lock, "spin_lock")
+#define _raw_spin_unlock(lock)	_do_spin_unlock(lock)
 
 struct _rwlock_debug {
 	volatile unsigned int lock;
@@ -56,35 +48,35 @@ extern void _do_read_unlock(rwlock_t *rw, char *str);
 extern void _do_write_lock(rwlock_t *rw, char *str);
 extern void _do_write_unlock(rwlock_t *rw);
 
-#define read_lock(lock)	\
+#define _raw_read_lock(lock)	\
 do {	unsigned long flags; \
 	__save_and_cli(flags); \
 	_do_read_lock(lock, "read_lock"); \
 	__restore_flags(flags); \
 } while(0)
 
-#define read_unlock(lock) \
+#define _raw_read_unlock(lock) \
 do {	unsigned long flags; \
 	__save_and_cli(flags); \
 	_do_read_unlock(lock, "read_unlock"); \
 	__restore_flags(flags); \
 } while(0)
 
-#define write_lock(lock) \
+#define _raw_write_lock(lock) \
 do {	unsigned long flags; \
 	__save_and_cli(flags); \
 	_do_write_lock(lock, "write_lock"); \
 	__restore_flags(flags); \
 } while(0)
 
-#define write_unlock(lock) \
+#define _raw_write_unlock(lock) \
 do {	unsigned long flags; \
 	__save_and_cli(flags); \
 	_do_write_unlock(lock); \
 	__restore_flags(flags); \
 } while(0)
 
-#else /* !SPIN_LOCK_DEBUG */
+#else /* !CONFIG_DEBUG_SPINLOCK */
 
 typedef unsigned char spinlock_t;
 #define SPIN_LOCK_UNLOCKED	0
@@ -97,7 +89,7 @@ do { \
 	barrier(); \
 } while(*((volatile unsigned char *)lock))
 
-extern __inline__ void spin_lock(spinlock_t *lock)
+extern __inline__ void _raw_spin_lock(spinlock_t *lock)
 {
 	__asm__ __volatile__(
 	"\n1:\n\t"
@@ -117,7 +109,7 @@ extern __inline__ void spin_lock(spinlock_t *lock)
 	: "g2", "memory", "cc");
 }
 
-extern __inline__ int spin_trylock(spinlock_t *lock)
+extern __inline__ int _raw_spin_trylock(spinlock_t *lock)
 {
 	unsigned int result;
 	__asm__ __volatile__("ldstub [%1], %0"
@@ -127,7 +119,7 @@ extern __inline__ int spin_trylock(spinlock_t *lock)
 	return (result == 0);
 }
 
-extern __inline__ void spin_unlock(spinlock_t *lock)
+extern __inline__ void _raw_spin_unlock(spinlock_t *lock)
 {
 	__asm__ __volatile__("stb %%g0, [%0]" : : "r" (lock) : "memory");
 }
@@ -178,7 +170,7 @@ extern __inline__ void _read_lock(rwlock_t *rw)
 	: "g2", "g4", "memory", "cc");
 }
 
-#define read_lock(lock) \
+#define _raw_read_lock(lock) \
 do {	unsigned long flags; \
 	__save_and_cli(flags); \
 	_read_lock(lock); \
@@ -198,14 +190,14 @@ extern __inline__ void _read_unlock(rwlock_t *rw)
 	: "g2", "g4", "memory", "cc");
 }
 
-#define read_unlock(lock) \
+#define _raw_read_unlock(lock) \
 do {	unsigned long flags; \
 	__save_and_cli(flags); \
 	_read_unlock(lock); \
 	__restore_flags(flags); \
 } while(0)
 
-extern __inline__ void write_lock(rwlock_t *rw)
+extern __inline__ void _raw_write_lock(rwlock_t *rw)
 {
 	register rwlock_t *lp asm("g1");
 	lp = rw;
@@ -218,9 +210,9 @@ extern __inline__ void write_lock(rwlock_t *rw)
 	: "g2", "g4", "memory", "cc");
 }
 
-#define write_unlock(rw)	do { (rw)->lock = 0; } while(0)
+#define _raw_write_unlock(rw)	do { (rw)->lock = 0; } while(0)
 
-#endif /* SPIN_LOCK_DEBUG */
+#endif /* CONFIG_DEBUG_SPINLOCK */
 
 #endif /* !(__ASSEMBLY__) */
 

include/asm-sparc/system.h

@@ -4,19 +4,18 @@
 #ifndef __SPARC_SYSTEM_H
 #define __SPARC_SYSTEM_H
 
+#include <linux/config.h>
 #include <linux/kernel.h>
+#include <linux/threads.h>	/* NR_CPUS */
 
 #include <asm/segment.h>
-
-#ifdef __KERNEL__
+#include <asm/thread_info.h>
 #include <asm/page.h>
-#include <asm/oplib.h>
+#include <asm/openprom.h>	/* romvec. XXX will be dealt later. Promise. */
 #include <asm/psr.h>
 #include <asm/ptrace.h>
 #include <asm/btfixup.h>
 
-#endif /* __KERNEL__ */
-
 #ifndef __ASSEMBLY__
 
 /*
@@ -48,6 +47,8 @@ extern enum sparc_cpu sparc_cpu_model;
 
 #define SUN4M_NCPUS            4              /* Architectural limit of sun4m. */
 
+extern struct thread_info *current_set[NR_CPUS];
+
 extern unsigned long empty_bad_page;
 extern unsigned long empty_bad_page_table;
 extern unsigned long empty_zero_page;
@@ -67,28 +68,43 @@ extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
 		   void *fpqueue, unsigned long *fpqdepth);
 
 #ifdef CONFIG_SMP
-#define SWITCH_ENTER \
-	if(prev->flags & PF_USEDFPU) { \
+#define SWITCH_ENTER(prv) \
+	do {			\
+	if (test_tsk_thread_flag(prv, TIF_USEDFPU) { \
 		put_psr(get_psr() | PSR_EF); \
-		fpsave(&prev->thread.float_regs[0], &prev->thread.fsr, \
-		       &prev->thread.fpqueue[0], &prev->thread.fpqdepth); \
-		prev->flags &= ~PF_USEDFPU; \
-		prev->thread.kregs->psr &= ~PSR_EF; \
-	}
-
-#define SWITCH_DO_LAZY_FPU
+		fpsave(&(prv)->thread.float_regs[0], &(prv)->thread.fsr, \
+		       &(prv)->thread.fpqueue[0], &(prv)->thread.fpqdepth); \
+		clear_tsk_thread_flag(prv, TIF_USEDFPU); \
+		(prv)->thread.kregs->psr &= ~PSR_EF; \
+	} \
+	} while(0)
+
+#define SWITCH_DO_LAZY_FPU(next)	/* */
 #else
-#define SWITCH_ENTER
-#define SWITCH_DO_LAZY_FPU if(last_task_used_math != next) next->thread.kregs->psr&=~PSR_EF;
+#define SWITCH_ENTER(prv)		/* */
+#define SWITCH_DO_LAZY_FPU(nxt)	\
+	do {			\
+	if (last_task_used_math != (nxt))		\
+		(nxt)->thread.kregs->psr&=~PSR_EF;	\
+	} while(0)
 #endif
 
+// #define prepare_arch_schedule(prev)	task_lock(prev)
+// #define finish_arch_schedule(prev)	task_unlock(prev)
+#define prepare_arch_schedule(prev)	do{ }while(0)
+#define finish_arch_schedule(prev)	do{ }while(0)
+
 /*
  * Flush windows so that the VM switch which follows
  * would not pull the stack from under us.
  *
  * SWITCH_ENTER and SWITH_DO_LAZY_FPU do not work yet (e.g. SMP does not work)
+ * XXX WTF is the above comment? Found in late teen 2.4.x.
+ *
+ * XXX prepare_arch_switch() is much smarter than this in sparc64, are we sure?
+ * XXX Cosider if doing it the flush_user_windows way is faster (by uwinmask).
  */
-#define prepare_to_switch() do { \
+#define prepare_arch_switch(rq) do { \
 	__asm__ __volatile__( \
 	".globl\tflush_patch_switch\nflush_patch_switch:\n\t" \
 	"save %sp, -0x40, %sp; save %sp, -0x40, %sp; save %sp, -0x40, %sp\n\t" \
@@ -96,6 +112,7 @@ extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
 	"save %sp, -0x40, %sp\n\t" \
 	"restore; restore; restore; restore; restore; restore; restore"); \
 } while(0)
+#define finish_arch_switch(rq)	do{ }while(0)
 
 	/* Much care has gone into this code, do not touch it.
 	 *
@@ -111,9 +128,8 @@ extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
 #define switch_to(prev, next, last) do {						\
 	__label__ here;									\
 	register unsigned long task_pc asm("o7");					\
-	extern struct task_struct *current_set[NR_CPUS];				\
-	SWITCH_ENTER									\
-	SWITCH_DO_LAZY_FPU								\
+	SWITCH_ENTER(prev);								\
+	SWITCH_DO_LAZY_FPU(next);							\
 	next->active_mm->cpu_vm_mask |= (1 << smp_processor_id());			\
 	task_pc = ((unsigned long) &&here) - 0x8;					\
 	__asm__ __volatile__(								\
@@ -140,11 +156,13 @@ extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
 	"nop\n\t"									\
 	"nop\n\t"									\
 	"jmpl	%%o7 + 0x8, %%g0\n\t"							\
-	" mov	%%g3, %0\n\t"								\
+	" ld	[%%g3 + %5], %0\n\t"								\
         : "=&r" (last)									\
-        : "r" (&(current_set[hard_smp_processor_id()])), "r" (next),			\
-	  "i" ((const unsigned long)(&((struct task_struct *)0)->thread.kpsr)),		\
-	  "i" ((const unsigned long)(&((struct task_struct *)0)->thread.ksp)),		\
+        : "r" (&(current_set[hard_smp_processor_id()])),	\
+	  "r" ((next)->thread_info),				\
+	  "i" (TI_KPSR),					\
+	  "i" (TI_KSP),						\
+	  "i" (TI_TASK),					\
 	  "r" (task_pc)									\
 	: "g1", "g2", "g3", "g4", "g5", "g7", "l0", "l1",				\
 	"l4", "l5", "l6", "l7", "i0", "i1", "i2", "i3", "i4", "i5", "o0", "o1", "o2",	\

include/asm-sparc/thread_info.h

+/*
+ * thread_info.h: sparc low-level thread information
+ * adapted from the ppc version by Pete Zaitcev, which was
+ * adapted from the i386 version by Paul Mackerras
+ *
+ * Copyright (C) 2002  David Howells (dhowells@redhat.com)
+ * Copyright (c) 2002  Pete Zaitcev (zaitcev@yahoo.com)
+ * - Incorporating suggestions made by Linus Torvalds and Dave Miller
+ */
+
+#ifndef _ASM_THREAD_INFO_H
+#define _ASM_THREAD_INFO_H
+
+// XXX todo: comment thread_info components and see what breaks.
+
+#ifdef __KERNEL__
+
+#ifndef __ASSEMBLY__
+
+#include <asm/btfixup.h>
+
+/*
+ * Low level task data.
+ *
+ * If you change this, change the TI_* offsets below to match. XXX check_asm.
+ *
+ * The uwinmask is a first class citizen among w_saved and friends.
+ * XXX Is this a good idea? wof.S/wuf.S have to use w_saved anyway,
+ *     so they waste a register on current, and an ld on fetching it.
+ */
+struct thread_info {
+	unsigned long		uwinmask;
+	struct task_struct	*task;		/* main task structure */
+	struct exec_domain	*exec_domain;	/* execution domain */
+	unsigned long		flags;		/* low level flags */
+
+	int			cpu;		/* cpu we're on */
+	int			preempt_count;
+	int			softirq_count;
+	int			hardirq_count;
+
+	/* Context switch saved kernel state. */
+	unsigned long ksp;	/* ... ksp __attribute__ ((aligned (8))); */
+	unsigned long kpc;
+	unsigned long kpsr;
+	unsigned long kwim;
+};
+
+/*
+ * macros/functions for gaining access to the thread information structure
+ */
+#define INIT_THREAD_INFO(tsk)			\
+{						\
+	uwinmask:	0,			\
+	task:		&tsk,			\
+	exec_domain:	&default_exec_domain,	\
+	flags:		0,			\
+	cpu:		0,			\
+}
+
+#define init_thread_info	(init_thread_union.thread_info)
+#define init_stack		(init_thread_union.stack)
+
+/* how to get the thread information struct from C */
+register struct thread_info *current_thread_info_reg asm("g6");
+#define current_thread_info()   (current_thread_info_reg)
+
+/*
+ * thread information allocation
+ */
+#ifdef CONFIG_SUN4
+#define THREAD_INFO_ORDER  0
+#else
+#define THREAD_INFO_ORDER  1
+#endif
+
+BTFIXUPDEF_CALL(struct thread_info *, alloc_thread_info, void)
+#define alloc_thread_info() BTFIXUP_CALL(alloc_thread_info)()
+
+BTFIXUPDEF_CALL(void, free_thread_info, struct thread_info *)
+#define free_thread_info(ti) BTFIXUP_CALL(free_thread_info)(ti)
+
+#define get_thread_info(ti)	get_task_struct((ti)->task)
+#define put_thread_info(ti)	put_task_struct((ti)->task)
+
+#endif /* __ASSEMBLY__ */
+
+/*
+ * Size of kernel stack for each process.
+ * Observe the order of get_free_pages() in alloc_thread_info().
+ * The sun4 has 8K stack too, because it's short on memory, and 16K is a waste.
+ *
+ * XXX Watch how INIT_THREAD_SIZE evolves in linux/sched.h and elsewhere.
+ *     On 2.5.24 it happens to match 8192 magically.
+ */
+#define THREAD_SIZE		8192
+
+/*
+ * Offsets in thread_info structure, used in assembly code
+ */
+#define TI_UWINMASK	0x00	/* uwinmask */
+#define TI_TASK		0x04
+#define TI_EXECDOMAIN	0x08	/* exec_domain */
+#define TI_FLAGS	0x0c
+#define TI_CPU		0x10
+#define TI_PREEMPT	0x14	/* preempt_count */
+#define TI_SOFTIRQ	0x18	/* softirq_count */
+#define TI_HARDIRQ	0x1c	/* hardirq_count */
+#define TI_KSP		0x20	/* ksp */
+#define TI_KPC		0x24	/* kpc (ldd'ed with kpc) */
+#define TI_KPSR		0x28	/* kpsr */
+#define TI_KWIM		0x2c	/* kwim (ldd'ed with kpsr) */
+
+#define PREEMPT_ACTIVE		0x4000000
+
+/*
+ * thread information flag bit numbers
+ */
+#define TIF_SYSCALL_TRACE	0	/* syscall trace active */
+#define TIF_NOTIFY_RESUME	1	/* resumption notification requested */
+#define TIF_SIGPENDING		2	/* signal pending */
+#define TIF_NEED_RESCHED	3	/* rescheduling necessary */
+#define TIF_USEDFPU		8	/* FPU was used by this task
+					 * this quantum (SMP) */
+#define TIF_POLLING_NRFLAG	9	/* true if poll_idle() is polling
+					 * TIF_NEED_RESCHED */
+
+/* as above, but as bit values */
+#define _TIF_SYSCALL_TRACE	(1<<TIF_SYSCALL_TRACE)
+#define _TIF_NOTIFY_RESUME	(1<<TIF_NOTIFY_RESUME)
+#define _TIF_SIGPENDING		(1<<TIF_SIGPENDING)
+#define _TIF_NEED_RESCHED	(1<<TIF_NEED_RESCHED)
+#define _TIF_USEDFPU		(1<<TIF_USEDFPU)
+#define _TIF_POLLING_NRFLAG	(1<<TIF_POLLING_NRFLAG)
+
+#endif /* __KERNEL__ */
+
+#endif /* _ASM_THREAD_INFO_H */

include/asm-sparc/tlb.h

+#ifndef _SPARC_TLB_H
+#define _SPARC_TLB_H
+
+#define tlb_start_vma(tlb, vma) \
+do {								\
+	flush_cache_range(vma, vma->vm_start, vma->vm_end);	\
+} while (0)
+
+#define tlb_end_vma(tlb, vma) \
+do {								\
+	flush_tlb_range(vma, vma->vm_start, vma->vm_end);	\
+} while (0)
+
+#define tlb_remove_tlb_entry(tlb, pte, address) \
+	do { } while (0)
+
+#define tlb_flush(tlb) \
+do {								\
+	flush_tlb_mm((tlb)->mm);				\
+} while (0)
+
 #include <asm-generic/tlb.h>
+
+#endif /* _SPARC_TLB_H */

include/asm-sparc/tlbflush.h

+#ifndef _SPARC_TLBFLUSH_H
+#define _SPARC_TLBFLUSH_H
+
+#include <linux/config.h>
+#include <linux/mm.h>
+// #include <asm/processor.h>
+
+/*
+ * TLB flushing:
+ *
+ *  - flush_tlb() flushes the current mm struct TLBs	XXX Exists?
+ *  - flush_tlb_all() flushes all processes TLBs 
+ *  - flush_tlb_mm(mm) flushes the specified mm context TLB's
+ *  - flush_tlb_page(vma, vmaddr) flushes one page
+ *  - flush_tlb_range(vma, start, end) flushes a range of pages
+ *  - flush_tlb_kernel_range(start, end) flushes a range of kernel pages
+ *  - flush_tlb_pgtables(mm, start, end) flushes a range of page tables
+ */
+
+#ifdef CONFIG_SMP
+
+BTFIXUPDEF_CALL(void, local_flush_tlb_all, void)
+BTFIXUPDEF_CALL(void, local_flush_tlb_mm, struct mm_struct *)
+BTFIXUPDEF_CALL(void, local_flush_tlb_range, struct vm_area_struct *, unsigned long, unsigned long)
+BTFIXUPDEF_CALL(void, local_flush_tlb_page, struct vm_area_struct *, unsigned long)
+
+#define local_flush_tlb_all() BTFIXUP_CALL(local_flush_tlb_all)()
+#define local_flush_tlb_mm(mm) BTFIXUP_CALL(local_flush_tlb_mm)(mm)
+#define local_flush_tlb_range(vma,start,end) BTFIXUP_CALL(local_flush_tlb_range)(vma,start,end)
+#define local_flush_tlb_page(vma,addr) BTFIXUP_CALL(local_flush_tlb_page)(vma,addr)
+
+extern void smp_flush_tlb_all(void);
+extern void smp_flush_tlb_mm(struct mm_struct *mm);
+extern void smp_flush_tlb_range(struct vm_area_struct *vma,
+				  unsigned long start,
+				  unsigned long end);
+extern void smp_flush_tlb_page(struct vm_area_struct *mm, unsigned long page);
+
+#endif /* CONFIG_SMP */
+
+BTFIXUPDEF_CALL(void, flush_tlb_all, void)
+BTFIXUPDEF_CALL(void, flush_tlb_mm, struct mm_struct *)
+BTFIXUPDEF_CALL(void, flush_tlb_range, struct vm_area_struct *, unsigned long, unsigned long)
+BTFIXUPDEF_CALL(void, flush_tlb_page, struct vm_area_struct *, unsigned long)
+
+// Thanks to Anton Blanchard, our pagetables became uncached in 2.4. Wee!
+// extern void flush_tlb_pgtables(struct mm_struct *mm,
+//     unsigned long start, unsigned long end);
+#define flush_tlb_pgtables(mm, start, end)	do{ }while(0)
+
+#define flush_tlb_all() BTFIXUP_CALL(flush_tlb_all)()
+#define flush_tlb_mm(mm) BTFIXUP_CALL(flush_tlb_mm)(mm)
+#define flush_tlb_range(vma,start,end) BTFIXUP_CALL(flush_tlb_range)(vma,start,end)
+#define flush_tlb_page(vma,addr) BTFIXUP_CALL(flush_tlb_page)(vma,addr)
+
+// #define flush_tlb() flush_tlb_mm(current->active_mm)	/* XXX Sure? */
+
+/*
+ * This is a kludge, until I know better. --zaitcev XXX
+ */
+#define flush_tlb_kernel_range(start, end) flush_tlb_all()
+
+#endif /* _SPARC_TLBFLUSH_H */

include/asm-sparc/uaccess.h

@@ -10,6 +10,7 @@
 #ifdef __KERNEL__
 #include <linux/sched.h>
 #include <linux/string.h>
+#include <linux/errno.h>
 #include <asm/vac-ops.h>
 #include <asm/a.out.h>
 #endif