linux/arch/sparc/kernel/ktlb.S

339 lines
7.3 KiB
ArmAsm
Raw Normal View History

/* arch/sparc64/kernel/ktlb.S: Kernel mapping TLB miss handling.
*
* Copyright (C) 1995, 1997, 2005, 2008 David S. Miller <davem@davemloft.net>
* Copyright (C) 1996 Eddie C. Dost (ecd@brainaid.de)
* Copyright (C) 1996 Miguel de Icaza (miguel@nuclecu.unam.mx)
* Copyright (C) 1996,98,99 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
*/
#include <asm/head.h>
#include <asm/asi.h>
#include <asm/page.h>
#include <asm/pgtable.h>
#include <asm/tsb.h>
.text
.align 32
kvmap_itlb:
/* g6: TAG TARGET */
mov TLB_TAG_ACCESS, %g4
ldxa [%g4] ASI_IMMU, %g4
/* sun4v_itlb_miss branches here with the missing virtual
* address already loaded into %g4
*/
kvmap_itlb_4v:
kvmap_itlb_nonlinear:
/* Catch kernel NULL pointer calls. */
sethi %hi(PAGE_SIZE), %g5
cmp %g4, %g5
bleu,pn %xcc, kvmap_dtlb_longpath
nop
KERN_TSB_LOOKUP_TL1(%g4, %g6, %g5, %g1, %g2, %g3, kvmap_itlb_load)
kvmap_itlb_tsb_miss:
sethi %hi(LOW_OBP_ADDRESS), %g5
cmp %g4, %g5
blu,pn %xcc, kvmap_itlb_vmalloc_addr
mov 0x1, %g5
sllx %g5, 32, %g5
cmp %g4, %g5
blu,pn %xcc, kvmap_itlb_obp
nop
kvmap_itlb_vmalloc_addr:
KERN_PGTABLE_WALK(%g4, %g5, %g2, kvmap_itlb_longpath)
KTSB_LOCK_TAG(%g1, %g2, %g7)
/* Load and check PTE. */
ldxa [%g5] ASI_PHYS_USE_EC, %g5
mov 1, %g7
sllx %g7, TSB_TAG_INVALID_BIT, %g7
brgez,a,pn %g5, kvmap_itlb_longpath
KTSB_STORE(%g1, %g7)
KTSB_WRITE(%g1, %g5, %g6)
/* fallthrough to TLB load */
kvmap_itlb_load:
661: stxa %g5, [%g0] ASI_ITLB_DATA_IN
retry
.section .sun4v_2insn_patch, "ax"
.word 661b
nop
nop
.previous
/* For sun4v the ASI_ITLB_DATA_IN store and the retry
* instruction get nop'd out and we get here to branch
* to the sun4v tlb load code. The registers are setup
* as follows:
*
* %g4: vaddr
* %g5: PTE
* %g6: TAG
*
* The sun4v TLB load wants the PTE in %g3 so we fix that
* up here.
*/
ba,pt %xcc, sun4v_itlb_load
mov %g5, %g3
kvmap_itlb_longpath:
661: rdpr %pstate, %g5
wrpr %g5, PSTATE_AG | PSTATE_MG, %pstate
.section .sun4v_2insn_patch, "ax"
.word 661b
SET_GL(1)
nop
.previous
rdpr %tpc, %g5
ba,pt %xcc, sparc64_realfault_common
mov FAULT_CODE_ITLB, %g4
kvmap_itlb_obp:
OBP_TRANS_LOOKUP(%g4, %g5, %g2, %g3, kvmap_itlb_longpath)
KTSB_LOCK_TAG(%g1, %g2, %g7)
KTSB_WRITE(%g1, %g5, %g6)
ba,pt %xcc, kvmap_itlb_load
nop
kvmap_dtlb_obp:
OBP_TRANS_LOOKUP(%g4, %g5, %g2, %g3, kvmap_dtlb_longpath)
KTSB_LOCK_TAG(%g1, %g2, %g7)
KTSB_WRITE(%g1, %g5, %g6)
ba,pt %xcc, kvmap_dtlb_load
nop
[SPARC64]: Fix boot failures on SunBlade-150 The sequence to move over to the Linux trap tables from the firmware ones needs to be more air tight. It turns out that to be %100 safe we do need to be able to translate OBP mappings in our TLB miss handlers early. In order not to eat up a lot of kernel image memory with static page tables, just use the translations array in the OBP TLB miss handlers. That solves the bulk of the problem. Furthermore, to make sure the OBP TLB miss path will work even before the fixed MMU globals are loaded, explicitly load %g1 to TLB_SFSR at the beginning of the i-TLB and d-TLB miss handlers. To ease the OBP TLB miss walking of the prom_trans[] array, we sort it then delete all of the non-OBP entries in there (for example, there are entries for the kernel image itself which we're not interested in at all). We also save about 32K of kernel image size with this change. Not a bad side effect :-) There are still some reasons why trampoline.S can't use the setup_trap_table() yet. The most noteworthy are: 1) OBP boots secondary processors with non-bias'd stack for some reason. This is easily fixed by using a small bootup stack in the kernel image explicitly for this purpose. 2) Doing a firmware call via the normal C call prom_set_trap_table() goes through the whole OBP enter/exit sequence that saves and restores OBP and Linux kernel state in the MMUs. This path unfortunately does a "flush %g6" while loading up the OBP locked TLB entries for the firmware call. If we setup the %g6 in the trampoline.S code properly, that is in the PAGE_OFFSET linear mapping, but we're not on the kernel trap table yet so those addresses won't translate properly. One idea is to do a by-hand firmware call like we do in the early bootup code and elsewhere here in trampoline.S But this fails as well, as aparently the secondary processors are not booted with OBP's special locked TLB entries loaded. These are necessary for the firwmare to processes TLB misses correctly up until the point where we take over the trap table. This does need to be resolved at some point. Signed-off-by: David S. Miller <davem@davemloft.net>
2005-10-12 19:22:46 +00:00
.align 32
kvmap_dtlb_tsb4m_load:
KTSB_LOCK_TAG(%g1, %g2, %g7)
KTSB_WRITE(%g1, %g5, %g6)
ba,pt %xcc, kvmap_dtlb_load
nop
kvmap_dtlb:
/* %g6: TAG TARGET */
mov TLB_TAG_ACCESS, %g4
ldxa [%g4] ASI_DMMU, %g4
/* sun4v_dtlb_miss branches here with the missing virtual
* address already loaded into %g4
*/
kvmap_dtlb_4v:
brgez,pn %g4, kvmap_dtlb_nonlinear
nop
#ifdef CONFIG_DEBUG_PAGEALLOC
/* Index through the base page size TSB even for linear
* mappings when using page allocation debugging.
*/
KERN_TSB_LOOKUP_TL1(%g4, %g6, %g5, %g1, %g2, %g3, kvmap_dtlb_load)
#else
/* Correct TAG_TARGET is already in %g6, check 4mb TSB. */
KERN_TSB4M_LOOKUP_TL1(%g6, %g5, %g1, %g2, %g3, kvmap_dtlb_load)
#endif
/* TSB entry address left in %g1, lookup linear PTE.
* Must preserve %g1 and %g6 (TAG).
*/
kvmap_dtlb_tsb4m_miss:
sparc64: Validate linear D-TLB misses. When page alloc debugging is not enabled, we essentially accept any virtual address for linear kernel TLB misses. But with kgdb, kernel address probing, and other facilities we can try to access arbitrary crap. So, make sure the address we miss on will translate to physical memory that actually exists. In order to make this work we have to embed the valid address bitmap into the kernel image. And in order to make that less expensive we make an adjustment, in that the max physical memory address is decreased to "1 << 41", even on the chips that support a 42-bit physical address space. We can do this because bit 41 indicates "I/O space" and thus covers non-memory ranges. The result of this is that: 1) kpte_linear_bitmap shrinks from 2K to 1K in size 2) we need 64K more for the valid address bitmap We can't let the valid address bitmap be dynamically allocated once we start using it to validate TLB misses, otherwise we have crazy issues to deal with wrt. recursive TLB misses and such. If we're in a TLB miss it could be the deepest trap level that's legal inside of the cpu. So if we TLB miss referencing the bitmap, the cpu will be out of trap levels and enter RED state. To guard against out-of-range accesses to the bitmap, we have to check to make sure no bits in the physical address above bit 40 are set. We could export and use last_valid_pfn for this check, but that's just an unnecessary extra memory reference. On the plus side of all this, since we load all of these translations into the special 4MB mapping TSB, and we check the TSB first for TLB misses, there should be absolutely no real cost for these new checks in the TLB miss path. Reported-by: heyongli@gmail.com Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-25 23:47:46 +00:00
/* Clear the PAGE_OFFSET top virtual bits, shift
* down to get PFN, and make sure PFN is in range.
*/
sllx %g4, 21, %g5
sparc64: Validate linear D-TLB misses. When page alloc debugging is not enabled, we essentially accept any virtual address for linear kernel TLB misses. But with kgdb, kernel address probing, and other facilities we can try to access arbitrary crap. So, make sure the address we miss on will translate to physical memory that actually exists. In order to make this work we have to embed the valid address bitmap into the kernel image. And in order to make that less expensive we make an adjustment, in that the max physical memory address is decreased to "1 << 41", even on the chips that support a 42-bit physical address space. We can do this because bit 41 indicates "I/O space" and thus covers non-memory ranges. The result of this is that: 1) kpte_linear_bitmap shrinks from 2K to 1K in size 2) we need 64K more for the valid address bitmap We can't let the valid address bitmap be dynamically allocated once we start using it to validate TLB misses, otherwise we have crazy issues to deal with wrt. recursive TLB misses and such. If we're in a TLB miss it could be the deepest trap level that's legal inside of the cpu. So if we TLB miss referencing the bitmap, the cpu will be out of trap levels and enter RED state. To guard against out-of-range accesses to the bitmap, we have to check to make sure no bits in the physical address above bit 40 are set. We could export and use last_valid_pfn for this check, but that's just an unnecessary extra memory reference. On the plus side of all this, since we load all of these translations into the special 4MB mapping TSB, and we check the TSB first for TLB misses, there should be absolutely no real cost for these new checks in the TLB miss path. Reported-by: heyongli@gmail.com Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-25 23:47:46 +00:00
/* Check to see if we know about valid memory at the 4MB
* chunk this physical address will reside within.
*/
sparc64: Validate linear D-TLB misses. When page alloc debugging is not enabled, we essentially accept any virtual address for linear kernel TLB misses. But with kgdb, kernel address probing, and other facilities we can try to access arbitrary crap. So, make sure the address we miss on will translate to physical memory that actually exists. In order to make this work we have to embed the valid address bitmap into the kernel image. And in order to make that less expensive we make an adjustment, in that the max physical memory address is decreased to "1 << 41", even on the chips that support a 42-bit physical address space. We can do this because bit 41 indicates "I/O space" and thus covers non-memory ranges. The result of this is that: 1) kpte_linear_bitmap shrinks from 2K to 1K in size 2) we need 64K more for the valid address bitmap We can't let the valid address bitmap be dynamically allocated once we start using it to validate TLB misses, otherwise we have crazy issues to deal with wrt. recursive TLB misses and such. If we're in a TLB miss it could be the deepest trap level that's legal inside of the cpu. So if we TLB miss referencing the bitmap, the cpu will be out of trap levels and enter RED state. To guard against out-of-range accesses to the bitmap, we have to check to make sure no bits in the physical address above bit 40 are set. We could export and use last_valid_pfn for this check, but that's just an unnecessary extra memory reference. On the plus side of all this, since we load all of these translations into the special 4MB mapping TSB, and we check the TSB first for TLB misses, there should be absolutely no real cost for these new checks in the TLB miss path. Reported-by: heyongli@gmail.com Signed-off-by: David S. Miller <davem@davemloft.net>
2009-08-25 23:47:46 +00:00
srlx %g5, 21 + 41, %g2
brnz,pn %g2, kvmap_dtlb_longpath
nop
/* This unconditional branch and delay-slot nop gets patched
* by the sethi sequence once the bitmap is properly setup.
*/
.globl valid_addr_bitmap_insn
valid_addr_bitmap_insn:
ba,pt %xcc, 2f
nop
.subsection 2
.globl valid_addr_bitmap_patch
valid_addr_bitmap_patch:
sethi %hi(sparc64_valid_addr_bitmap), %g7
or %g7, %lo(sparc64_valid_addr_bitmap), %g7
.previous
srlx %g5, 21 + 22, %g2
srlx %g2, 6, %g5
and %g2, 63, %g2
sllx %g5, 3, %g5
ldx [%g7 + %g5], %g5
mov 1, %g7
sllx %g7, %g2, %g7
andcc %g5, %g7, %g0
be,pn %xcc, kvmap_dtlb_longpath
2: sethi %hi(kpte_linear_bitmap), %g2
or %g2, %lo(kpte_linear_bitmap), %g2
/* Get the 256MB physical address index. */
sllx %g4, 21, %g5
mov 1, %g7
srlx %g5, 21 + 28, %g5
/* Don't try this at home kids... this depends upon srlx
* only taking the low 6 bits of the shift count in %g5.
*/
sllx %g7, %g5, %g7
/* Divide by 64 to get the offset into the bitmask. */
srlx %g5, 6, %g5
sllx %g5, 3, %g5
/* kern_linear_pte_xor[((mask & bit) ? 1 : 0)] */
ldx [%g2 + %g5], %g2
andcc %g2, %g7, %g0
sethi %hi(kern_linear_pte_xor), %g5
or %g5, %lo(kern_linear_pte_xor), %g5
bne,a,pt %xcc, 1f
add %g5, 8, %g5
1: ldx [%g5], %g2
.globl kvmap_linear_patch
kvmap_linear_patch:
ba,pt %xcc, kvmap_dtlb_tsb4m_load
xor %g2, %g4, %g5
kvmap_dtlb_vmalloc_addr:
KERN_PGTABLE_WALK(%g4, %g5, %g2, kvmap_dtlb_longpath)
KTSB_LOCK_TAG(%g1, %g2, %g7)
/* Load and check PTE. */
ldxa [%g5] ASI_PHYS_USE_EC, %g5
mov 1, %g7
sllx %g7, TSB_TAG_INVALID_BIT, %g7
brgez,a,pn %g5, kvmap_dtlb_longpath
KTSB_STORE(%g1, %g7)
KTSB_WRITE(%g1, %g5, %g6)
/* fallthrough to TLB load */
kvmap_dtlb_load:
661: stxa %g5, [%g0] ASI_DTLB_DATA_IN ! Reload TLB
retry
.section .sun4v_2insn_patch, "ax"
.word 661b
nop
nop
.previous
/* For sun4v the ASI_DTLB_DATA_IN store and the retry
* instruction get nop'd out and we get here to branch
* to the sun4v tlb load code. The registers are setup
* as follows:
*
* %g4: vaddr
* %g5: PTE
* %g6: TAG
*
* The sun4v TLB load wants the PTE in %g3 so we fix that
* up here.
*/
ba,pt %xcc, sun4v_dtlb_load
mov %g5, %g3
#ifdef CONFIG_SPARSEMEM_VMEMMAP
kvmap_vmemmap:
sub %g4, %g5, %g5
srlx %g5, 22, %g5
sethi %hi(vmemmap_table), %g1
sllx %g5, 3, %g5
or %g1, %lo(vmemmap_table), %g1
ba,pt %xcc, kvmap_dtlb_load
ldx [%g1 + %g5], %g5
#endif
kvmap_dtlb_nonlinear:
/* Catch kernel NULL pointer derefs. */
sethi %hi(PAGE_SIZE), %g5
cmp %g4, %g5
bleu,pn %xcc, kvmap_dtlb_longpath
nop
#ifdef CONFIG_SPARSEMEM_VMEMMAP
/* Do not use the TSB for vmemmap. */
mov (VMEMMAP_BASE >> 40), %g5
sllx %g5, 40, %g5
cmp %g4,%g5
bgeu,pn %xcc, kvmap_vmemmap
nop
#endif
KERN_TSB_LOOKUP_TL1(%g4, %g6, %g5, %g1, %g2, %g3, kvmap_dtlb_load)
kvmap_dtlb_tsbmiss:
sethi %hi(MODULES_VADDR), %g5
cmp %g4, %g5
blu,pn %xcc, kvmap_dtlb_longpath
mov (VMALLOC_END >> 40), %g5
sllx %g5, 40, %g5
cmp %g4, %g5
bgeu,pn %xcc, kvmap_dtlb_longpath
nop
kvmap_check_obp:
sethi %hi(LOW_OBP_ADDRESS), %g5
cmp %g4, %g5
blu,pn %xcc, kvmap_dtlb_vmalloc_addr
mov 0x1, %g5
sllx %g5, 32, %g5
cmp %g4, %g5
blu,pn %xcc, kvmap_dtlb_obp
nop
ba,pt %xcc, kvmap_dtlb_vmalloc_addr
nop
kvmap_dtlb_longpath:
661: rdpr %pstate, %g5
wrpr %g5, PSTATE_AG | PSTATE_MG, %pstate
.section .sun4v_2insn_patch, "ax"
.word 661b
SET_GL(1)
ldxa [%g0] ASI_SCRATCHPAD, %g5
.previous
rdpr %tl, %g3
cmp %g3, 1
661: mov TLB_TAG_ACCESS, %g4
ldxa [%g4] ASI_DMMU, %g5
.section .sun4v_2insn_patch, "ax"
.word 661b
ldx [%g5 + HV_FAULT_D_ADDR_OFFSET], %g5
nop
.previous
be,pt %xcc, sparc64_realfault_common
mov FAULT_CODE_DTLB, %g4
ba,pt %xcc, winfix_trampoline
nop