M7350/kernel/arch/powerpc/mm/hash_native_64.c

711 lines
19 KiB
C
Raw Normal View History

2024-09-09 08:52:07 +00:00
/*
* native hashtable management.
*
* SMP scalability work:
* Copyright (C) 2001 Anton Blanchard <anton@au.ibm.com>, IBM
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*/
#undef DEBUG_LOW
#include <linux/spinlock.h>
#include <linux/bitops.h>
2024-09-09 08:57:42 +00:00
#include <linux/of.h>
2024-09-09 08:52:07 +00:00
#include <linux/threads.h>
#include <linux/smp.h>
#include <asm/machdep.h>
#include <asm/mmu.h>
#include <asm/mmu_context.h>
#include <asm/pgtable.h>
#include <asm/tlbflush.h>
#include <asm/tlb.h>
#include <asm/cputable.h>
#include <asm/udbg.h>
#include <asm/kexec.h>
#include <asm/ppc-opcode.h>
2024-09-09 08:57:42 +00:00
#include <misc/cxl.h>
2024-09-09 08:52:07 +00:00
#ifdef DEBUG_LOW
#define DBG_LOW(fmt...) udbg_printf(fmt)
#else
#define DBG_LOW(fmt...)
#endif
2024-09-09 08:57:42 +00:00
#ifdef __BIG_ENDIAN__
2024-09-09 08:52:07 +00:00
#define HPTE_LOCK_BIT 3
2024-09-09 08:57:42 +00:00
#else
#define HPTE_LOCK_BIT (56+3)
#endif
2024-09-09 08:52:07 +00:00
DEFINE_RAW_SPINLOCK(native_tlbie_lock);
2024-09-09 08:57:42 +00:00
static inline void __tlbie(unsigned long vpn, int psize, int apsize, int ssize)
2024-09-09 08:52:07 +00:00
{
2024-09-09 08:57:42 +00:00
unsigned long va;
2024-09-09 08:52:07 +00:00
unsigned int penc;
2024-09-09 08:57:42 +00:00
unsigned long sllp;
2024-09-09 08:52:07 +00:00
2024-09-09 08:57:42 +00:00
/*
* We need 14 to 65 bits of va for a tlibe of 4K page
* With vpn we ignore the lower VPN_SHIFT bits already.
* And top two bits are already ignored because we can
* only accomadate 76 bits in a 64 bit vpn with a VPN_SHIFT
* of 12.
*/
va = vpn << VPN_SHIFT;
/*
* clear top 16 bits of 64bit va, non SLS segment
* Older versions of the architecture (2.02 and earler) require the
* masking of the top 16 bits.
*/
2024-09-09 08:52:07 +00:00
va &= ~(0xffffULL << 48);
switch (psize) {
case MMU_PAGE_4K:
2024-09-09 08:57:42 +00:00
/* clear out bits after (52) [0....52.....63] */
va &= ~((1ul << (64 - 52)) - 1);
2024-09-09 08:52:07 +00:00
va |= ssize << 8;
2024-09-09 08:57:42 +00:00
sllp = ((mmu_psize_defs[apsize].sllp & SLB_VSID_L) >> 6) |
((mmu_psize_defs[apsize].sllp & SLB_VSID_LP) >> 4);
va |= sllp << 5;
2024-09-09 08:52:07 +00:00
asm volatile(ASM_FTR_IFCLR("tlbie %0,0", PPC_TLBIE(%1,%0), %2)
: : "r" (va), "r"(0), "i" (CPU_FTR_ARCH_206)
: "memory");
break;
default:
2024-09-09 08:57:42 +00:00
/* We need 14 to 14 + i bits of va */
penc = mmu_psize_defs[psize].penc[apsize];
va &= ~((1ul << mmu_psize_defs[apsize].shift) - 1);
2024-09-09 08:52:07 +00:00
va |= penc << 12;
va |= ssize << 8;
2024-09-09 08:57:42 +00:00
/*
* AVAL bits:
* We don't need all the bits, but rest of the bits
* must be ignored by the processor.
* vpn cover upto 65 bits of va. (0...65) and we need
* 58..64 bits of va.
*/
va |= (vpn & 0xfe); /* AVAL */
2024-09-09 08:52:07 +00:00
va |= 1; /* L */
asm volatile(ASM_FTR_IFCLR("tlbie %0,1", PPC_TLBIE(%1,%0), %2)
: : "r" (va), "r"(0), "i" (CPU_FTR_ARCH_206)
: "memory");
break;
}
}
2024-09-09 08:57:42 +00:00
static inline void __tlbiel(unsigned long vpn, int psize, int apsize, int ssize)
2024-09-09 08:52:07 +00:00
{
2024-09-09 08:57:42 +00:00
unsigned long va;
2024-09-09 08:52:07 +00:00
unsigned int penc;
2024-09-09 08:57:42 +00:00
unsigned long sllp;
2024-09-09 08:52:07 +00:00
2024-09-09 08:57:42 +00:00
/* VPN_SHIFT can be atmost 12 */
va = vpn << VPN_SHIFT;
/*
* clear top 16 bits of 64 bit va, non SLS segment
* Older versions of the architecture (2.02 and earler) require the
* masking of the top 16 bits.
*/
2024-09-09 08:52:07 +00:00
va &= ~(0xffffULL << 48);
switch (psize) {
case MMU_PAGE_4K:
2024-09-09 08:57:42 +00:00
/* clear out bits after(52) [0....52.....63] */
va &= ~((1ul << (64 - 52)) - 1);
2024-09-09 08:52:07 +00:00
va |= ssize << 8;
2024-09-09 08:57:42 +00:00
sllp = ((mmu_psize_defs[apsize].sllp & SLB_VSID_L) >> 6) |
((mmu_psize_defs[apsize].sllp & SLB_VSID_LP) >> 4);
va |= sllp << 5;
2024-09-09 08:52:07 +00:00
asm volatile(".long 0x7c000224 | (%0 << 11) | (0 << 21)"
: : "r"(va) : "memory");
break;
default:
2024-09-09 08:57:42 +00:00
/* We need 14 to 14 + i bits of va */
penc = mmu_psize_defs[psize].penc[apsize];
va &= ~((1ul << mmu_psize_defs[apsize].shift) - 1);
2024-09-09 08:52:07 +00:00
va |= penc << 12;
va |= ssize << 8;
2024-09-09 08:57:42 +00:00
/*
* AVAL bits:
* We don't need all the bits, but rest of the bits
* must be ignored by the processor.
* vpn cover upto 65 bits of va. (0...65) and we need
* 58..64 bits of va.
*/
va |= (vpn & 0xfe);
2024-09-09 08:52:07 +00:00
va |= 1; /* L */
asm volatile(".long 0x7c000224 | (%0 << 11) | (1 << 21)"
: : "r"(va) : "memory");
break;
}
}
2024-09-09 08:57:42 +00:00
static inline void tlbie(unsigned long vpn, int psize, int apsize,
int ssize, int local)
2024-09-09 08:52:07 +00:00
{
2024-09-09 08:57:42 +00:00
unsigned int use_local;
2024-09-09 08:52:07 +00:00
int lock_tlbie = !mmu_has_feature(MMU_FTR_LOCKLESS_TLBIE);
2024-09-09 08:57:42 +00:00
use_local = local && mmu_has_feature(MMU_FTR_TLBIEL) && !cxl_ctx_in_use();
2024-09-09 08:52:07 +00:00
if (use_local)
use_local = mmu_psize_defs[psize].tlbiel;
if (lock_tlbie && !use_local)
raw_spin_lock(&native_tlbie_lock);
asm volatile("ptesync": : :"memory");
if (use_local) {
2024-09-09 08:57:42 +00:00
__tlbiel(vpn, psize, apsize, ssize);
2024-09-09 08:52:07 +00:00
asm volatile("ptesync": : :"memory");
} else {
2024-09-09 08:57:42 +00:00
__tlbie(vpn, psize, apsize, ssize);
2024-09-09 08:52:07 +00:00
asm volatile("eieio; tlbsync; ptesync": : :"memory");
}
if (lock_tlbie && !use_local)
raw_spin_unlock(&native_tlbie_lock);
}
static inline void native_lock_hpte(struct hash_pte *hptep)
{
2024-09-09 08:57:42 +00:00
unsigned long *word = (unsigned long *)&hptep->v;
2024-09-09 08:52:07 +00:00
while (1) {
if (!test_and_set_bit_lock(HPTE_LOCK_BIT, word))
break;
while(test_bit(HPTE_LOCK_BIT, word))
cpu_relax();
}
}
static inline void native_unlock_hpte(struct hash_pte *hptep)
{
2024-09-09 08:57:42 +00:00
unsigned long *word = (unsigned long *)&hptep->v;
2024-09-09 08:52:07 +00:00
clear_bit_unlock(HPTE_LOCK_BIT, word);
}
2024-09-09 08:57:42 +00:00
static long native_hpte_insert(unsigned long hpte_group, unsigned long vpn,
2024-09-09 08:52:07 +00:00
unsigned long pa, unsigned long rflags,
2024-09-09 08:57:42 +00:00
unsigned long vflags, int psize, int apsize, int ssize)
2024-09-09 08:52:07 +00:00
{
struct hash_pte *hptep = htab_address + hpte_group;
unsigned long hpte_v, hpte_r;
int i;
if (!(vflags & HPTE_V_BOLTED)) {
2024-09-09 08:57:42 +00:00
DBG_LOW(" insert(group=%lx, vpn=%016lx, pa=%016lx,"
2024-09-09 08:52:07 +00:00
" rflags=%lx, vflags=%lx, psize=%d)\n",
2024-09-09 08:57:42 +00:00
hpte_group, vpn, pa, rflags, vflags, psize);
2024-09-09 08:52:07 +00:00
}
for (i = 0; i < HPTES_PER_GROUP; i++) {
2024-09-09 08:57:42 +00:00
if (! (be64_to_cpu(hptep->v) & HPTE_V_VALID)) {
2024-09-09 08:52:07 +00:00
/* retry with lock held */
native_lock_hpte(hptep);
2024-09-09 08:57:42 +00:00
if (! (be64_to_cpu(hptep->v) & HPTE_V_VALID))
2024-09-09 08:52:07 +00:00
break;
native_unlock_hpte(hptep);
}
hptep++;
}
if (i == HPTES_PER_GROUP)
return -1;
2024-09-09 08:57:42 +00:00
hpte_v = hpte_encode_v(vpn, psize, apsize, ssize) | vflags | HPTE_V_VALID;
hpte_r = hpte_encode_r(pa, psize, apsize) | rflags;
2024-09-09 08:52:07 +00:00
if (!(vflags & HPTE_V_BOLTED)) {
DBG_LOW(" i=%x hpte_v=%016lx, hpte_r=%016lx\n",
i, hpte_v, hpte_r);
}
2024-09-09 08:57:42 +00:00
hptep->r = cpu_to_be64(hpte_r);
2024-09-09 08:52:07 +00:00
/* Guarantee the second dword is visible before the valid bit */
eieio();
/*
* Now set the first dword including the valid bit
* NOTE: this also unlocks the hpte
*/
2024-09-09 08:57:42 +00:00
hptep->v = cpu_to_be64(hpte_v);
2024-09-09 08:52:07 +00:00
__asm__ __volatile__ ("ptesync" : : : "memory");
return i | (!!(vflags & HPTE_V_SECONDARY) << 3);
}
static long native_hpte_remove(unsigned long hpte_group)
{
struct hash_pte *hptep;
int i;
int slot_offset;
unsigned long hpte_v;
DBG_LOW(" remove(group=%lx)\n", hpte_group);
/* pick a random entry to start at */
slot_offset = mftb() & 0x7;
for (i = 0; i < HPTES_PER_GROUP; i++) {
hptep = htab_address + hpte_group + slot_offset;
2024-09-09 08:57:42 +00:00
hpte_v = be64_to_cpu(hptep->v);
2024-09-09 08:52:07 +00:00
if ((hpte_v & HPTE_V_VALID) && !(hpte_v & HPTE_V_BOLTED)) {
/* retry with lock held */
native_lock_hpte(hptep);
2024-09-09 08:57:42 +00:00
hpte_v = be64_to_cpu(hptep->v);
2024-09-09 08:52:07 +00:00
if ((hpte_v & HPTE_V_VALID)
&& !(hpte_v & HPTE_V_BOLTED))
break;
native_unlock_hpte(hptep);
}
slot_offset++;
slot_offset &= 0x7;
}
if (i == HPTES_PER_GROUP)
return -1;
/* Invalidate the hpte. NOTE: this also unlocks it */
hptep->v = 0;
return i;
}
static long native_hpte_updatepp(unsigned long slot, unsigned long newpp,
2024-09-09 08:57:42 +00:00
unsigned long vpn, int bpsize,
int apsize, int ssize, int local)
2024-09-09 08:52:07 +00:00
{
struct hash_pte *hptep = htab_address + slot;
unsigned long hpte_v, want_v;
int ret = 0;
2024-09-09 08:57:42 +00:00
want_v = hpte_encode_avpn(vpn, bpsize, ssize);
2024-09-09 08:52:07 +00:00
2024-09-09 08:57:42 +00:00
DBG_LOW(" update(vpn=%016lx, avpnv=%016lx, group=%lx, newpp=%lx)",
vpn, want_v & HPTE_V_AVPN, slot, newpp);
2024-09-09 08:52:07 +00:00
native_lock_hpte(hptep);
2024-09-09 08:57:42 +00:00
hpte_v = be64_to_cpu(hptep->v);
/*
* We need to invalidate the TLB always because hpte_remove doesn't do
* a tlb invalidate. If a hash bucket gets full, we "evict" a more/less
* random entry from it. When we do that we don't invalidate the TLB
* (hpte_remove) because we assume the old translation is still
* technically "valid".
*/
2024-09-09 08:52:07 +00:00
if (!HPTE_V_COMPARE(hpte_v, want_v) || !(hpte_v & HPTE_V_VALID)) {
DBG_LOW(" -> miss\n");
ret = -1;
} else {
DBG_LOW(" -> hit\n");
/* Update the HPTE */
2024-09-09 08:57:42 +00:00
hptep->r = cpu_to_be64((be64_to_cpu(hptep->r) & ~(HPTE_R_PP | HPTE_R_N)) |
(newpp & (HPTE_R_PP | HPTE_R_N | HPTE_R_C)));
2024-09-09 08:52:07 +00:00
}
native_unlock_hpte(hptep);
/* Ensure it is out of the tlb too. */
2024-09-09 08:57:42 +00:00
tlbie(vpn, bpsize, apsize, ssize, local);
2024-09-09 08:52:07 +00:00
return ret;
}
2024-09-09 08:57:42 +00:00
static long native_hpte_find(unsigned long vpn, int psize, int ssize)
2024-09-09 08:52:07 +00:00
{
struct hash_pte *hptep;
unsigned long hash;
unsigned long i;
long slot;
unsigned long want_v, hpte_v;
2024-09-09 08:57:42 +00:00
hash = hpt_hash(vpn, mmu_psize_defs[psize].shift, ssize);
want_v = hpte_encode_avpn(vpn, psize, ssize);
2024-09-09 08:52:07 +00:00
/* Bolted mappings are only ever in the primary group */
slot = (hash & htab_hash_mask) * HPTES_PER_GROUP;
for (i = 0; i < HPTES_PER_GROUP; i++) {
hptep = htab_address + slot;
2024-09-09 08:57:42 +00:00
hpte_v = be64_to_cpu(hptep->v);
2024-09-09 08:52:07 +00:00
if (HPTE_V_COMPARE(hpte_v, want_v) && (hpte_v & HPTE_V_VALID))
/* HPTE matches */
return slot;
++slot;
}
return -1;
}
/*
* Update the page protection bits. Intended to be used to create
* guard pages for kernel data structures on pages which are bolted
* in the HPT. Assumes pages being operated on will not be stolen.
*
* No need to lock here because we should be the only user.
*/
static void native_hpte_updateboltedpp(unsigned long newpp, unsigned long ea,
int psize, int ssize)
{
2024-09-09 08:57:42 +00:00
unsigned long vpn;
unsigned long vsid;
2024-09-09 08:52:07 +00:00
long slot;
struct hash_pte *hptep;
vsid = get_kernel_vsid(ea, ssize);
2024-09-09 08:57:42 +00:00
vpn = hpt_vpn(ea, vsid, ssize);
2024-09-09 08:52:07 +00:00
2024-09-09 08:57:42 +00:00
slot = native_hpte_find(vpn, psize, ssize);
2024-09-09 08:52:07 +00:00
if (slot == -1)
panic("could not find page to bolt\n");
hptep = htab_address + slot;
/* Update the HPTE */
2024-09-09 08:57:42 +00:00
hptep->r = cpu_to_be64((be64_to_cpu(hptep->r) &
~(HPTE_R_PP | HPTE_R_N)) |
(newpp & (HPTE_R_PP | HPTE_R_N)));
/*
* Ensure it is out of the tlb too. Bolted entries base and
* actual page size will be same.
*/
tlbie(vpn, psize, psize, ssize, 0);
2024-09-09 08:52:07 +00:00
}
2024-09-09 08:57:42 +00:00
static void native_hpte_invalidate(unsigned long slot, unsigned long vpn,
int bpsize, int apsize, int ssize, int local)
2024-09-09 08:52:07 +00:00
{
struct hash_pte *hptep = htab_address + slot;
unsigned long hpte_v;
unsigned long want_v;
unsigned long flags;
local_irq_save(flags);
2024-09-09 08:57:42 +00:00
DBG_LOW(" invalidate(vpn=%016lx, hash: %lx)\n", vpn, slot);
2024-09-09 08:52:07 +00:00
2024-09-09 08:57:42 +00:00
want_v = hpte_encode_avpn(vpn, bpsize, ssize);
2024-09-09 08:52:07 +00:00
native_lock_hpte(hptep);
2024-09-09 08:57:42 +00:00
hpte_v = be64_to_cpu(hptep->v);
2024-09-09 08:52:07 +00:00
2024-09-09 08:57:42 +00:00
/*
* We need to invalidate the TLB always because hpte_remove doesn't do
* a tlb invalidate. If a hash bucket gets full, we "evict" a more/less
* random entry from it. When we do that we don't invalidate the TLB
* (hpte_remove) because we assume the old translation is still
* technically "valid".
*/
2024-09-09 08:52:07 +00:00
if (!HPTE_V_COMPARE(hpte_v, want_v) || !(hpte_v & HPTE_V_VALID))
native_unlock_hpte(hptep);
else
/* Invalidate the hpte. NOTE: this also unlocks it */
hptep->v = 0;
/* Invalidate the TLB */
2024-09-09 08:57:42 +00:00
tlbie(vpn, bpsize, apsize, ssize, local);
2024-09-09 08:52:07 +00:00
local_irq_restore(flags);
}
2024-09-09 08:57:42 +00:00
static void native_hugepage_invalidate(unsigned long vsid,
unsigned long addr,
unsigned char *hpte_slot_array,
int psize, int ssize)
{
int i;
struct hash_pte *hptep;
int actual_psize = MMU_PAGE_16M;
unsigned int max_hpte_count, valid;
unsigned long flags, s_addr = addr;
unsigned long hpte_v, want_v, shift;
unsigned long hidx, vpn = 0, hash, slot;
shift = mmu_psize_defs[psize].shift;
max_hpte_count = 1U << (PMD_SHIFT - shift);
local_irq_save(flags);
for (i = 0; i < max_hpte_count; i++) {
valid = hpte_valid(hpte_slot_array, i);
if (!valid)
continue;
hidx = hpte_hash_index(hpte_slot_array, i);
/* get the vpn */
addr = s_addr + (i * (1ul << shift));
vpn = hpt_vpn(addr, vsid, ssize);
hash = hpt_hash(vpn, shift, ssize);
if (hidx & _PTEIDX_SECONDARY)
hash = ~hash;
slot = (hash & htab_hash_mask) * HPTES_PER_GROUP;
slot += hidx & _PTEIDX_GROUP_IX;
hptep = htab_address + slot;
want_v = hpte_encode_avpn(vpn, psize, ssize);
native_lock_hpte(hptep);
hpte_v = be64_to_cpu(hptep->v);
/* Even if we miss, we need to invalidate the TLB */
if (!HPTE_V_COMPARE(hpte_v, want_v) || !(hpte_v & HPTE_V_VALID))
native_unlock_hpte(hptep);
else
/* Invalidate the hpte. NOTE: this also unlocks it */
hptep->v = 0;
/*
* We need to do tlb invalidate for all the address, tlbie
* instruction compares entry_VA in tlb with the VA specified
* here
*/
tlbie(vpn, psize, actual_psize, ssize, 0);
}
local_irq_restore(flags);
}
static inline int __hpte_actual_psize(unsigned int lp, int psize)
{
int i, shift;
unsigned int mask;
/* start from 1 ignoring MMU_PAGE_4K */
for (i = 1; i < MMU_PAGE_COUNT; i++) {
/* invalid penc */
if (mmu_psize_defs[psize].penc[i] == -1)
continue;
/*
* encoding bits per actual page size
* PTE LP actual page size
* rrrr rrrz >=8KB
* rrrr rrzz >=16KB
* rrrr rzzz >=32KB
* rrrr zzzz >=64KB
* .......
*/
shift = mmu_psize_defs[i].shift - LP_SHIFT;
if (shift > LP_BITS)
shift = LP_BITS;
mask = (1 << shift) - 1;
if ((lp & mask) == mmu_psize_defs[psize].penc[i])
return i;
}
return -1;
}
2024-09-09 08:52:07 +00:00
static void hpte_decode(struct hash_pte *hpte, unsigned long slot,
2024-09-09 08:57:42 +00:00
int *psize, int *apsize, int *ssize, unsigned long *vpn)
2024-09-09 08:52:07 +00:00
{
2024-09-09 08:57:42 +00:00
unsigned long avpn, pteg, vpi;
unsigned long hpte_v = be64_to_cpu(hpte->v);
unsigned long hpte_r = be64_to_cpu(hpte->r);
unsigned long vsid, seg_off;
int size, a_size, shift;
/* Look at the 8 bit LP value */
unsigned int lp = (hpte_r >> LP_SHIFT) & ((1 << LP_BITS) - 1);
if (!(hpte_v & HPTE_V_LARGE)) {
size = MMU_PAGE_4K;
a_size = MMU_PAGE_4K;
} else {
2024-09-09 08:52:07 +00:00
for (size = 0; size < MMU_PAGE_COUNT; size++) {
/* valid entries have a shift value */
if (!mmu_psize_defs[size].shift)
continue;
2024-09-09 08:57:42 +00:00
a_size = __hpte_actual_psize(lp, size);
if (a_size != -1)
2024-09-09 08:52:07 +00:00
break;
}
}
/* This works for all page sizes, and for 256M and 1T segments */
2024-09-09 08:57:42 +00:00
*ssize = hpte_v >> HPTE_V_SSIZE_SHIFT;
2024-09-09 08:52:07 +00:00
shift = mmu_psize_defs[size].shift;
2024-09-09 08:57:42 +00:00
avpn = (HPTE_V_AVPN_VAL(hpte_v) & ~mmu_psize_defs[size].avpnm);
pteg = slot / HPTES_PER_GROUP;
if (hpte_v & HPTE_V_SECONDARY)
pteg = ~pteg;
switch (*ssize) {
case MMU_SEGSIZE_256M:
/* We only have 28 - 23 bits of seg_off in avpn */
seg_off = (avpn & 0x1f) << 23;
vsid = avpn >> 5;
/* We can find more bits from the pteg value */
if (shift < 23) {
vpi = (vsid ^ pteg) & htab_hash_mask;
seg_off |= vpi << shift;
}
*vpn = vsid << (SID_SHIFT - VPN_SHIFT) | seg_off >> VPN_SHIFT;
break;
case MMU_SEGSIZE_1T:
/* We only have 40 - 23 bits of seg_off in avpn */
seg_off = (avpn & 0x1ffff) << 23;
vsid = avpn >> 17;
if (shift < 23) {
2024-09-09 08:52:07 +00:00
vpi = (vsid ^ (vsid << 25) ^ pteg) & htab_hash_mask;
2024-09-09 08:57:42 +00:00
seg_off |= vpi << shift;
2024-09-09 08:52:07 +00:00
}
2024-09-09 08:57:42 +00:00
*vpn = vsid << (SID_SHIFT_1T - VPN_SHIFT) | seg_off >> VPN_SHIFT;
break;
default:
*vpn = size = 0;
2024-09-09 08:52:07 +00:00
}
2024-09-09 08:57:42 +00:00
*psize = size;
*apsize = a_size;
2024-09-09 08:52:07 +00:00
}
/*
* clear all mappings on kexec. All cpus are in real mode (or they will
* be when they isi), and we are the only one left. We rely on our kernel
* mapping being 0xC0's and the hardware ignoring those two real bits.
*
* TODO: add batching support when enabled. remember, no dynamic memory here,
* athough there is the control page available...
*/
static void native_hpte_clear(void)
{
2024-09-09 08:57:42 +00:00
unsigned long vpn = 0;
2024-09-09 08:52:07 +00:00
unsigned long slot, slots, flags;
struct hash_pte *hptep = htab_address;
2024-09-09 08:57:42 +00:00
unsigned long hpte_v;
2024-09-09 08:52:07 +00:00
unsigned long pteg_count;
2024-09-09 08:57:42 +00:00
int psize, apsize, ssize;
2024-09-09 08:52:07 +00:00
pteg_count = htab_hash_mask + 1;
local_irq_save(flags);
/* we take the tlbie lock and hold it. Some hardware will
* deadlock if we try to tlbie from two processors at once.
*/
raw_spin_lock(&native_tlbie_lock);
slots = pteg_count * HPTES_PER_GROUP;
for (slot = 0; slot < slots; slot++, hptep++) {
/*
* we could lock the pte here, but we are the only cpu
* running, right? and for crash dump, we probably
* don't want to wait for a maybe bad cpu.
*/
2024-09-09 08:57:42 +00:00
hpte_v = be64_to_cpu(hptep->v);
2024-09-09 08:52:07 +00:00
/*
* Call __tlbie() here rather than tlbie() since we
* already hold the native_tlbie_lock.
*/
if (hpte_v & HPTE_V_VALID) {
2024-09-09 08:57:42 +00:00
hpte_decode(hptep, slot, &psize, &apsize, &ssize, &vpn);
2024-09-09 08:52:07 +00:00
hptep->v = 0;
2024-09-09 08:57:42 +00:00
__tlbie(vpn, psize, apsize, ssize);
2024-09-09 08:52:07 +00:00
}
}
asm volatile("eieio; tlbsync; ptesync":::"memory");
raw_spin_unlock(&native_tlbie_lock);
local_irq_restore(flags);
}
/*
* Batched hash table flush, we batch the tlbie's to avoid taking/releasing
* the lock all the time
*/
static void native_flush_hash_range(unsigned long number, int local)
{
2024-09-09 08:57:42 +00:00
unsigned long vpn;
unsigned long hash, index, hidx, shift, slot;
2024-09-09 08:52:07 +00:00
struct hash_pte *hptep;
unsigned long hpte_v;
unsigned long want_v;
unsigned long flags;
real_pte_t pte;
struct ppc64_tlb_batch *batch = &__get_cpu_var(ppc64_tlb_batch);
unsigned long psize = batch->psize;
int ssize = batch->ssize;
int i;
local_irq_save(flags);
for (i = 0; i < number; i++) {
2024-09-09 08:57:42 +00:00
vpn = batch->vpn[i];
2024-09-09 08:52:07 +00:00
pte = batch->pte[i];
2024-09-09 08:57:42 +00:00
pte_iterate_hashed_subpages(pte, psize, vpn, index, shift) {
hash = hpt_hash(vpn, shift, ssize);
2024-09-09 08:52:07 +00:00
hidx = __rpte_to_hidx(pte, index);
if (hidx & _PTEIDX_SECONDARY)
hash = ~hash;
slot = (hash & htab_hash_mask) * HPTES_PER_GROUP;
slot += hidx & _PTEIDX_GROUP_IX;
hptep = htab_address + slot;
2024-09-09 08:57:42 +00:00
want_v = hpte_encode_avpn(vpn, psize, ssize);
2024-09-09 08:52:07 +00:00
native_lock_hpte(hptep);
2024-09-09 08:57:42 +00:00
hpte_v = be64_to_cpu(hptep->v);
2024-09-09 08:52:07 +00:00
if (!HPTE_V_COMPARE(hpte_v, want_v) ||
!(hpte_v & HPTE_V_VALID))
native_unlock_hpte(hptep);
else
hptep->v = 0;
} pte_iterate_hashed_end();
}
if (mmu_has_feature(MMU_FTR_TLBIEL) &&
mmu_psize_defs[psize].tlbiel && local) {
asm volatile("ptesync":::"memory");
for (i = 0; i < number; i++) {
2024-09-09 08:57:42 +00:00
vpn = batch->vpn[i];
2024-09-09 08:52:07 +00:00
pte = batch->pte[i];
2024-09-09 08:57:42 +00:00
pte_iterate_hashed_subpages(pte, psize,
vpn, index, shift) {
__tlbiel(vpn, psize, psize, ssize);
2024-09-09 08:52:07 +00:00
} pte_iterate_hashed_end();
}
asm volatile("ptesync":::"memory");
} else {
int lock_tlbie = !mmu_has_feature(MMU_FTR_LOCKLESS_TLBIE);
if (lock_tlbie)
raw_spin_lock(&native_tlbie_lock);
asm volatile("ptesync":::"memory");
for (i = 0; i < number; i++) {
2024-09-09 08:57:42 +00:00
vpn = batch->vpn[i];
2024-09-09 08:52:07 +00:00
pte = batch->pte[i];
2024-09-09 08:57:42 +00:00
pte_iterate_hashed_subpages(pte, psize,
vpn, index, shift) {
__tlbie(vpn, psize, psize, ssize);
2024-09-09 08:52:07 +00:00
} pte_iterate_hashed_end();
}
asm volatile("eieio; tlbsync; ptesync":::"memory");
if (lock_tlbie)
raw_spin_unlock(&native_tlbie_lock);
}
local_irq_restore(flags);
}
void __init hpte_init_native(void)
{
ppc_md.hpte_invalidate = native_hpte_invalidate;
ppc_md.hpte_updatepp = native_hpte_updatepp;
ppc_md.hpte_updateboltedpp = native_hpte_updateboltedpp;
ppc_md.hpte_insert = native_hpte_insert;
ppc_md.hpte_remove = native_hpte_remove;
ppc_md.hpte_clear_all = native_hpte_clear;
2024-09-09 08:57:42 +00:00
ppc_md.flush_hash_range = native_flush_hash_range;
ppc_md.hugepage_invalidate = native_hugepage_invalidate;
2024-09-09 08:52:07 +00:00
}