forge/linux-pinenote
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
linux-pinenote/arch/sparc/lib/copy_page.S

251 lines
5.8 KiB
ArmAsm
Raw Normal View History

Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
/* clear_page.S: UltraSparc optimized copy page.
*
* Copyright (C) 1996, 1998, 1999, 2000, 2004 David S. Miller (davem@redhat.com)
* Copyright (C) 1997 Jakub Jelinek (jakub@redhat.com)
*/
#include <asm/visasm.h>
#include <asm/thread_info.h>
#include <asm/page.h>
#include <asm/pgtable.h>
#include <asm/spitfire.h>
#include <asm/head.h>
/* What we used to do was lock a TLB entry into a specific
* TLB slot, clear the page with interrupts disabled, then
* restore the original TLB entry. This was great for
* disturbing the TLB as little as possible, but it meant
* we had to keep interrupts disabled for a long time.
*
* Now, we simply use the normal TLB loading mechanism,
* and this makes the cpu choose a slot all by itself.
* Then we do a normal TLB flush on exit. We need only
* disable preemption during the clear.
*/
#define DCACHE_SIZE (PAGE_SIZE * 2)
sparc64: Remove 4MB and 512K base page size options. Adrian Bunk reported that enabling 4MB page size breaks the build. The problem is that MAX_ORDER combined with the page shift exceeds the SECTION_SIZE_BITS we use in asm-sparc64/sparsemem.h There are several ways I suppose we could work around this. For one we could define a CONFIG_FORCE_MAX_ZONEORDER to decrease MAX_ORDER in these higher page size cases. But I also know that these page size cases are broken wrt. TLB miss handling especially on pre-hypervisor systems, and there isn't an easy way to fix that. These options were meant to be fun experimental hacks anyways, and only 8K and 64K make any sense to support. So remove 512K and 4M base page size support. Of course, we still support these page sizes for huge pages. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 23:43:55 -07:00
#if (PAGE_SHIFT == 13)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
#define PAGE_SIZE_REM 0x80
sparc64: Remove 4MB and 512K base page size options. Adrian Bunk reported that enabling 4MB page size breaks the build. The problem is that MAX_ORDER combined with the page shift exceeds the SECTION_SIZE_BITS we use in asm-sparc64/sparsemem.h There are several ways I suppose we could work around this. For one we could define a CONFIG_FORCE_MAX_ZONEORDER to decrease MAX_ORDER in these higher page size cases. But I also know that these page size cases are broken wrt. TLB miss handling especially on pre-hypervisor systems, and there isn't an easy way to fix that. These options were meant to be fun experimental hacks anyways, and only 8K and 64K make any sense to support. So remove 512K and 4M base page size support. Of course, we still support these page sizes for huge pages. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 23:43:55 -07:00
#elif (PAGE_SHIFT == 16)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
#define PAGE_SIZE_REM 0x100
#else
#error Wrong PAGE_SHIFT specified
#endif
#define TOUCH(reg0, reg1, reg2, reg3, reg4, reg5, reg6, reg7) \
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %reg0, %f48; fsrc2 %reg1, %f50; \
fsrc2 %reg2, %f52; fsrc2 %reg3, %f54; \
fsrc2 %reg4, %f56; fsrc2 %reg5, %f58; \
fsrc2 %reg6, %f60; fsrc2 %reg7, %f62;
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
.text
.align 32
.globl copy_user_page
.type copy_user_page,#function
copy_user_page: /* %o0=dest, %o1=src, %o2=vaddr */
lduw [%g6 + TI_PRE_COUNT], %o4
sparc64: Make PAGE_OFFSET variable. Choose PAGE_OFFSET dynamically based upon cpu type. Original UltraSPARC-I (spitfire) chips only supported a 44-bit virtual address space. Newer chips (T4 and later) support 52-bit virtual addresses and up to 47-bits of physical memory space. Therefore we have to adjust PAGE_SIZE dynamically based upon the capabilities of the chip. Note that this change alone does not allow us to support > 43-bit physical memory, to do that we need to re-arrange our page table support. The current encodings of the pmd_t and pgd_t pointers restricts us to "32 + 11" == 43 bits. This change can waste quite a bit of memory for the various tables. In particular, a future change should work to size and allocate kern_linear_bitmap[] and sparc64_valid_addr_bitmap[] dynamically. This isn't easy as we really cannot take a TLB miss when accessing kern_linear_bitmap[]. We'd have to lock it into the TLB or similar. Signed-off-by: David S. Miller <davem@davemloft.net> Acked-by: Bob Picco <bob.picco@oracle.com>
2013-09-20 21:50:41 -07:00
sethi %hi(PAGE_OFFSET), %g2
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
sethi %hi(PAGE_SIZE), %o3
sparc64: Make PAGE_OFFSET variable. Choose PAGE_OFFSET dynamically based upon cpu type. Original UltraSPARC-I (spitfire) chips only supported a 44-bit virtual address space. Newer chips (T4 and later) support 52-bit virtual addresses and up to 47-bits of physical memory space. Therefore we have to adjust PAGE_SIZE dynamically based upon the capabilities of the chip. Note that this change alone does not allow us to support > 43-bit physical memory, to do that we need to re-arrange our page table support. The current encodings of the pmd_t and pgd_t pointers restricts us to "32 + 11" == 43 bits. This change can waste quite a bit of memory for the various tables. In particular, a future change should work to size and allocate kern_linear_bitmap[] and sparc64_valid_addr_bitmap[] dynamically. This isn't easy as we really cannot take a TLB miss when accessing kern_linear_bitmap[]. We'd have to lock it into the TLB or similar. Signed-off-by: David S. Miller <davem@davemloft.net> Acked-by: Bob Picco <bob.picco@oracle.com>
2013-09-20 21:50:41 -07:00
ldx [%g2 + %lo(PAGE_OFFSET)], %g2
[SPARC64]: Deal with PTE layout differences in SUN4V. Yes, you heard it right, they changed the PTE layout for SUN4V. Ho hum... This is the simple and inefficient way to support this. It'll get optimized, don't worry. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-11 21:57:54 -08:00
sethi %hi(PAGE_KERNEL_LOCKED), %g3
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
[SPARC64]: Deal with PTE layout differences in SUN4V. Yes, you heard it right, they changed the PTE layout for SUN4V. Ho hum... This is the simple and inefficient way to support this. It'll get optimized, don't worry. Signed-off-by: David S. Miller <davem@davemloft.net>
2006-02-11 21:57:54 -08:00
ldx [%g3 + %lo(PAGE_KERNEL_LOCKED)], %g3
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
sub %o0, %g2, %g1 ! dest paddr
sub %o1, %g2, %g2 ! src paddr
and %o2, %o3, %o0 ! vaddr D-cache alias bit
or %g1, %g3, %g1 ! dest TTE data
or %g2, %g3, %g2 ! src TTE data
sethi %hi(TLBTEMP_BASE), %o3
sethi %hi(DCACHE_SIZE), %o1
add %o0, %o3, %o0 ! dest TTE vaddr
add %o4, 1, %o2
add %o0, %o1, %o1 ! src TTE vaddr
/* Disable preemption. */
mov TLB_TAG_ACCESS, %g3
stw %o2, [%g6 + TI_PRE_COUNT]
/* Load TLB entries. */
rdpr %pstate, %o2
wrpr %o2, PSTATE_IE, %pstate
stxa %o0, [%g3] ASI_DMMU
stxa %g1, [%g0] ASI_DTLB_DATA_IN
membar #Sync
stxa %o1, [%g3] ASI_DMMU
stxa %g2, [%g0] ASI_DTLB_DATA_IN
membar #Sync
wrpr %o2, 0x0, %pstate
[SPARC64]: Kill BRANCH_IF_ANY_CHEETAH() from copy page. Just patch the branch at boot time instead. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 11:26:15 -07:00
cheetah_copy_page_insn:
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ba,pt %xcc, 9f
nop
1:
VISEntryHalf
membar #StoreLoad | #StoreStore | #LoadStore
sethi %hi((PAGE_SIZE/64)-2), %o2
mov %o0, %g1
prefetch [%o1 + 0x000], #one_read
or %o2, %lo((PAGE_SIZE/64)-2), %o2
prefetch [%o1 + 0x040], #one_read
prefetch [%o1 + 0x080], #one_read
prefetch [%o1 + 0x0c0], #one_read
ldd [%o1 + 0x000], %f0
prefetch [%o1 + 0x100], #one_read
ldd [%o1 + 0x008], %f2
prefetch [%o1 + 0x140], #one_read
ldd [%o1 + 0x010], %f4
prefetch [%o1 + 0x180], #one_read
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f0, %f16
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x018], %f6
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f2, %f18
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x020], %f8
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f4, %f20
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x028], %f10
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f6, %f22
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x030], %f12
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f8, %f24
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x038], %f14
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f10, %f26
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x040], %f0
1: ldd [%o1 + 0x048], %f2
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f12, %f28
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x050], %f4
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f14, %f30
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
stda %f16, [%o0] ASI_BLK_P
ldd [%o1 + 0x058], %f6
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f0, %f16
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x060], %f8
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f2, %f18
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x068], %f10
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f4, %f20
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x070], %f12
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f6, %f22
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x078], %f14
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f8, %f24
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x080], %f0
prefetch [%o1 + 0x180], #one_read
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f10, %f26
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
subcc %o2, 1, %o2
add %o0, 0x40, %o0
bne,pt %xcc, 1b
add %o1, 0x40, %o1
ldd [%o1 + 0x048], %f2
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f12, %f28
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x050], %f4
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f14, %f30
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
stda %f16, [%o0] ASI_BLK_P
ldd [%o1 + 0x058], %f6
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f0, %f16
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x060], %f8
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f2, %f18
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x068], %f10
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f4, %f20
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
ldd [%o1 + 0x070], %f12
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f6, %f22
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
add %o0, 0x40, %o0
ldd [%o1 + 0x078], %f14
sparc64: Consistently use fsrc2 rather than fmovd in optimized asm. Because fsrc2, unlike fmovd, does not update the %fsr register. Signed-off-by: David S. Miller <davem@davemloft.net>
2012-06-27 01:25:23 -07:00
fsrc2 %f8, %f24
fsrc2 %f10, %f26
fsrc2 %f12, %f28
fsrc2 %f14, %f30
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
stda %f16, [%o0] ASI_BLK_P
membar #Sync
VISExitHalf
ba,pt %xcc, 5f
nop
9:
VISEntry
ldub [%g6 + TI_FAULT_CODE], %g3
mov %o0, %g1
cmp %g3, 0
rd %asi, %g3
be,a,pt %icc, 1f
wr %g0, ASI_BLK_P, %asi
wr %g0, ASI_BLK_COMMIT_P, %asi
1: ldda [%o1] ASI_BLK_P, %f0
add %o1, 0x40, %o1
ldda [%o1] ASI_BLK_P, %f16
add %o1, 0x40, %o1
sethi %hi(PAGE_SIZE), %o2
1: TOUCH(f0, f2, f4, f6, f8, f10, f12, f14)
ldda [%o1] ASI_BLK_P, %f32
stda %f48, [%o0] %asi
add %o1, 0x40, %o1
sub %o2, 0x40, %o2
add %o0, 0x40, %o0
TOUCH(f16, f18, f20, f22, f24, f26, f28, f30)
ldda [%o1] ASI_BLK_P, %f0
stda %f48, [%o0] %asi
add %o1, 0x40, %o1
sub %o2, 0x40, %o2
add %o0, 0x40, %o0
TOUCH(f32, f34, f36, f38, f40, f42, f44, f46)
ldda [%o1] ASI_BLK_P, %f16
stda %f48, [%o0] %asi
sub %o2, 0x40, %o2
add %o1, 0x40, %o1
cmp %o2, PAGE_SIZE_REM
bne,pt %xcc, 1b
add %o0, 0x40, %o0
sparc64: Remove 4MB and 512K base page size options. Adrian Bunk reported that enabling 4MB page size breaks the build. The problem is that MAX_ORDER combined with the page shift exceeds the SECTION_SIZE_BITS we use in asm-sparc64/sparsemem.h There are several ways I suppose we could work around this. For one we could define a CONFIG_FORCE_MAX_ZONEORDER to decrease MAX_ORDER in these higher page size cases. But I also know that these page size cases are broken wrt. TLB miss handling especially on pre-hypervisor systems, and there isn't an easy way to fix that. These options were meant to be fun experimental hacks anyways, and only 8K and 64K make any sense to support. So remove 512K and 4M base page size support. Of course, we still support these page sizes for huge pages. Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 23:43:55 -07:00
#if (PAGE_SHIFT == 16)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 15:20:36 -07:00
TOUCH(f0, f2, f4, f6, f8, f10, f12, f14)
ldda [%o1] ASI_BLK_P, %f32
stda %f48, [%o0] %asi
add %o1, 0x40, %o1
sub %o2, 0x40, %o2
add %o0, 0x40, %o0
TOUCH(f16, f18, f20, f22, f24, f26, f28, f30)
ldda [%o1] ASI_BLK_P, %f0
stda %f48, [%o0] %asi
add %o1, 0x40, %o1
sub %o2, 0x40, %o2
add %o0, 0x40, %o0
membar #Sync
stda %f32, [%o0] %asi
add %o0, 0x40, %o0
stda %f0, [%o0] %asi
#else
membar #Sync
stda %f0, [%o0] %asi
add %o0, 0x40, %o0
stda %f16, [%o0] %asi
#endif
membar #Sync
wr %g3, 0x0, %asi
VISExit
5:
stxa %g0, [%g1] ASI_DMMU_DEMAP
membar #Sync
sethi %hi(DCACHE_SIZE), %g2
stxa %g0, [%g1 + %g2] ASI_DMMU_DEMAP
membar #Sync
retl
stw %o4, [%g6 + TI_PRE_COUNT]
.size copy_user_page, .-copy_user_page
[SPARC64]: Kill BRANCH_IF_ANY_CHEETAH() from copy page. Just patch the branch at boot time instead. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-08-30 11:26:15 -07:00
.globl cheetah_patch_copy_page
cheetah_patch_copy_page:
sethi %hi(0x01000000), %o1 ! NOP
sethi %hi(cheetah_copy_page_insn), %o0
or %o0, %lo(cheetah_copy_page_insn), %o0
stw %o1, [%o0]
membar #StoreStore
flush %o0
retl
nop
Reference in a new issue Copy permalink