/**
* kvm_mips_build_ret_to_host() - Assemble code to return to the host.
* @addr: Address to start writing code.
*
* Assemble the code to handle return from the guest exit handler
* (kvm_mips_handle_exit()) back to the host, i.e. to the caller of the vcpu_run
* function generated by kvm_mips_build_vcpu_run().
*
* Returns: Next address after end of written function.
*/
static void *kvm_mips_build_ret_to_host(void *addr)
{
u32 *p = addr;
unsigned int i;
/* EBASE is already pointing to Linux */
UASM_i_LW(&p, K1, offsetof(struct kvm_vcpu_arch, host_stack), K1);
UASM_i_ADDIU(&p, K1, K1, -(int)sizeof(struct pt_regs));
/*
* r2/v0 is the return code, shift it down by 2 (arithmetic)
* to recover the err code
*/
uasm_i_sra(&p, K0, V0, 2);
uasm_i_move(&p, V0, K0);
/* Load context saved on the host stack */
for (i = 16; i < 31; ++i) {
if (i == 24)
i = 28;
UASM_i_LW(&p, i, offsetof(struct pt_regs, regs[i]), K1);
}
/* Restore RDHWR access */
UASM_i_LA_mostly(&p, K0, (long)&hwrena);
uasm_i_lw(&p, K0, uasm_rel_lo((long)&hwrena), K0);
uasm_i_mtc0(&p, K0, C0_HWRENA);
/* Restore RA, which is the address we will return to */
UASM_i_LW(&p, RA, offsetof(struct pt_regs, regs[RA]), K1);
uasm_i_jr(&p, RA);
uasm_i_nop(&p);
return p;
}
/*
* This places the pte into ENTRYLO0 and writes it with tlbwi
* or tlbwr as appropriate. This is because the index register
* may have the probe fail bit set as a result of a trap on a
* kseg2 access, i.e. without refill. Then it returns.
*/
static void __cpuinit
build_r3000_tlb_reload_write(u32 **p, struct uasm_label **l,
struct uasm_reloc **r, unsigned int pte,
unsigned int tmp)
{
uasm_i_mfc0(p, tmp, C0_INDEX);
uasm_i_mtc0(p, pte, C0_ENTRYLO0); /* cp0 delay */
uasm_il_bltz(p, r, tmp, label_r3000_write_probe_fail); /* cp0 delay */
uasm_i_mfc0(p, tmp, C0_EPC); /* branch delay */
uasm_i_tlbwi(p); /* cp0 delay */
uasm_i_jr(p, tmp);
uasm_i_rfe(p); /* branch delay */
uasm_l_r3000_write_probe_fail(l, *p);
uasm_i_tlbwr(p); /* cp0 delay */
uasm_i_jr(p, tmp);
uasm_i_rfe(p); /* branch delay */
}
/*
* This places the pte into ENTRYLO0 and writes it with tlbwi.
* Then it returns.
*/
static void __cpuinit
build_r3000_pte_reload_tlbwi(u32 **p, unsigned int pte, unsigned int tmp)
{
uasm_i_mtc0(p, pte, C0_ENTRYLO0); /* cp0 delay */
uasm_i_mfc0(p, tmp, C0_EPC); /* cp0 delay */
uasm_i_tlbwi(p);
uasm_i_jr(p, tmp);
uasm_i_rfe(p); /* branch delay */
}
/*
* The R3000 TLB handler is simple.
*/
static void __init build_r3000_tlb_refill_handler(void)
{
long pgdc = (long)pgd_current;
u32 *p;
int i;
memset(tlb_handler, 0, sizeof(tlb_handler));
p = tlb_handler;
uasm_i_mfc0(&p, K0, C0_BADVADDR);
uasm_i_lui(&p, K1, uasm_rel_hi(pgdc)); /* cp0 delay */
uasm_i_lw(&p, K1, uasm_rel_lo(pgdc), K1);
uasm_i_srl(&p, K0, K0, 22); /* load delay */
uasm_i_sll(&p, K0, K0, 2);
uasm_i_addu(&p, K1, K1, K0);
uasm_i_mfc0(&p, K0, C0_CONTEXT);
uasm_i_lw(&p, K1, 0, K1); /* cp0 delay */
uasm_i_andi(&p, K0, K0, 0xffc); /* load delay */
uasm_i_addu(&p, K1, K1, K0);
uasm_i_lw(&p, K0, 0, K1);
uasm_i_nop(&p); /* load delay */
uasm_i_mtc0(&p, K0, C0_ENTRYLO0);
uasm_i_mfc0(&p, K1, C0_EPC); /* cp0 delay */
uasm_i_tlbwr(&p); /* cp0 delay */
uasm_i_jr(&p, K1);
uasm_i_rfe(&p); /* branch delay */
if (p > tlb_handler + 32)
panic("TLB refill handler space exceeded");
pr_debug("Wrote TLB refill handler (%u instructions).\n",
(unsigned int)(p - tlb_handler));
pr_debug("\t.set push\n");
pr_debug("\t.set noreorder\n");
for (i = 0; i < (p - tlb_handler); i++)
pr_debug("\t.word 0x%08x\n", tlb_handler[i]);
pr_debug("\t.set pop\n");
memcpy((void *)ebase, tlb_handler, 0x80);
}
static void build_bounce_code(unsigned long *spp, unsigned long *gpp)
{
int i;
unsigned long base[32] = {0,};
unsigned int pflag = (unsigned int)KSEG1ADDR(&smp_flag);
unsigned int entry = (unsigned int)__jzsoc_secondary_start;
unsigned int *p;
for(i=0;i<32;i++) {
base[i] = __get_free_pages(GFP_KERNEL, 0);
if(!base[i] || (base[i] & 0xffff))
continue;
smp_bounce.base = base[i];
break;
}
for(i=i-1;i>=0;i--) {
free_pages(base[i], 0);
}
BUG_ON(!smp_bounce.base || (smp_bounce.base & 0xffff));
p = (unsigned int*)smp_bounce.base;
UASM_i_LA(&p, 26, pflag);
UASM_i_LW(&p, 2, 0, 26);
UASM_i_ADDIU(&p, 2, 2, 1);
UASM_i_SW(&p, 2, 0, 26);
/* t7: cpu_start. t8: cpu_ready. t9: cpu_running. */
UASM_i_LA(&p, 15, (unsigned long)cpu_start.bits);
UASM_i_LA(&p, 24, (unsigned long)cpu_ready_e.bits);
UASM_i_LA(&p, 25, (unsigned long)cpu_running.bits);
UASM_i_LA(&p, 29, (unsigned long)spp);
UASM_i_LA(&p, 28, (unsigned long)gpp);
UASM_i_LA(&p, 31, entry);
uasm_i_jr(&p, 31);
uasm_i_nop(&p);
}
static void * __init cps_gen_entry_code(unsigned cpu, enum cps_pm_state state)
{
struct uasm_label *l = labels;
struct uasm_reloc *r = relocs;
u32 *buf, *p;
const unsigned r_online = a0;
const unsigned r_nc_count = a1;
const unsigned r_pcohctl = t7;
const unsigned max_instrs = 256;
unsigned cpc_cmd;
enum {
lbl_incready = 1,
lbl_poll_cont,
lbl_secondary_hang,
lbl_disable_coherence,
lbl_flush_fsb,
lbl_invicache,
lbl_flushdcache,
lbl_hang,
lbl_set_cont,
lbl_secondary_cont,
lbl_decready,
};
/* Allocate a buffer to hold the generated code */
p = buf = kcalloc(max_instrs, sizeof(u32), GFP_KERNEL);
if (!buf)
return NULL;
/* Clear labels & relocs ready for (re)use */
memset(labels, 0, sizeof(labels));
memset(relocs, 0, sizeof(relocs));
if (state == CPS_PM_POWER_GATED) {
/* Power gating relies upon CPS SMP */
if (!mips_cps_smp_in_use())
goto out_err;
/*
* Save CPU state. Note the non-standard calling convention
* with the return address placed in v0 to avoid clobbering
* the ra register before it is saved.
*/
UASM_i_LA(&p, t0, (long)mips_cps_pm_save);
uasm_i_jalr(&p, v0, t0);
uasm_i_nop(&p);
}
/*
* Load addresses of required CM & CPC registers. This is done early
* because they're needed in both the enable & disable coherence steps
* but in the coupled case the enable step will only run on one VPE.
*/
UASM_i_LA(&p, r_pcohctl, (long)_gcmp_base + GCMPCLCBOFS(COHCTL));
if (coupled_coherence) {
/* Increment ready_count */
uasm_i_sync(&p, stype_ordering);
uasm_build_label(&l, p, lbl_incready);
uasm_i_ll(&p, t1, 0, r_nc_count);
uasm_i_addiu(&p, t2, t1, 1);
uasm_i_sc(&p, t2, 0, r_nc_count);
uasm_il_beqz(&p, &r, t2, lbl_incready);
uasm_i_addiu(&p, t1, t1, 1);
/* Ordering barrier */
uasm_i_sync(&p, stype_ordering);
/*
* If this is the last VPE to become ready for non-coherence
* then it should branch below.
*/
uasm_il_beq(&p, &r, t1, r_online, lbl_disable_coherence);
uasm_i_nop(&p);
if (state < CPS_PM_POWER_GATED) {
/*
* Otherwise this is not the last VPE to become ready
* for non-coherence. It needs to wait until coherence
* has been disabled before proceeding, which it will do
* by polling for the top bit of ready_count being set.
*/
uasm_i_addiu(&p, t1, zero, -1);
uasm_build_label(&l, p, lbl_poll_cont);
uasm_i_lw(&p, t0, 0, r_nc_count);
uasm_il_bltz(&p, &r, t0, lbl_secondary_cont);
uasm_i_ehb(&p);
uasm_i_yield(&p, zero, t1);
uasm_il_b(&p, &r, lbl_poll_cont);
uasm_i_nop(&p);
} else {
/*
* The core will lose power & this VPE will not continue
* so it can simply halt here.
*/
uasm_i_addiu(&p, t0, zero, TCHALT_H);
uasm_i_mtc0(&p, t0, 2, 4);
uasm_build_label(&l, p, lbl_secondary_hang);
uasm_il_b(&p, &r, lbl_secondary_hang);
uasm_i_nop(&p);
}
}
/*
* This is the point of no return - this VPE will now proceed to
* disable coherence. At this point we *must* be sure that no other
* VPE within the core will interfere with the L1 dcache.
*/
uasm_build_label(&l, p, lbl_disable_coherence);
/* Invalidate the L1 icache */
cps_gen_cache_routine(&p, &l, &r, &cpu_data[cpu].icache,
Index_Invalidate_I, lbl_invicache);
/* Writeback & invalidate the L1 dcache */
cps_gen_cache_routine(&p, &l, &r, &cpu_data[cpu].dcache,
Index_Writeback_Inv_D, lbl_flushdcache);
/* Completion barrier */
uasm_i_sync(&p, stype_memory);
uasm_i_ehb(&p);
/*
* Disable all but self interventions. The load from COHCTL is defined
* by the interAptiv & proAptiv SUMs as ensuring that the operation
* resulting from the preceeding store is complete.
*/
uasm_i_addiu(&p, t0, zero, 1 << cpu_data[cpu].core);
uasm_i_sw(&p, t0, 0, r_pcohctl);
uasm_i_lw(&p, t0, 0, r_pcohctl);
/* Sync to ensure previous interventions are complete */
uasm_i_sync(&p, stype_intervention);
uasm_i_ehb(&p);
/* Disable coherence */
uasm_i_sw(&p, zero, 0, r_pcohctl);
uasm_i_lw(&p, t0, 0, r_pcohctl);
if (state >= CPS_PM_CLOCK_GATED) {
/* TODO: determine whether required based on CPC version */
cps_gen_flush_fsb(&p, &l, &r, &cpu_data[cpu].dcache,
lbl_flush_fsb);
/* Determine the CPC command to issue */
switch (state) {
case CPS_PM_CLOCK_GATED:
cpc_cmd = CPC_Cx_CMD_CLOCKOFF;
break;
case CPS_PM_POWER_GATED:
cpc_cmd = CPC_Cx_CMD_PWRDOWN;
break;
default:
BUG();
goto out_err;
}
/* Issue the CPC command */
UASM_i_LA(&p, t0, (long)addr_cpc_cl_cmd());
uasm_i_addiu(&p, t1, zero, cpc_cmd);
uasm_i_sw(&p, t1, 0, t0);
if (state == CPS_PM_POWER_GATED) {
/* If anything goes wrong just hang */
uasm_build_label(&l, p, lbl_hang);
uasm_il_b(&p, &r, lbl_hang);
uasm_i_nop(&p);
/*
* There's no point generating more code, the core is
* powered down & if powered back up will run from the
* reset vector not from here.
*/
goto gen_done;
}
/* Completion barrier */
uasm_i_sync(&p, stype_memory);
uasm_i_ehb(&p);
}
if (state == CPS_PM_NC_WAIT) {
/*
* At this point it is safe for all VPEs to proceed with
* execution. This VPE will set the top bit of ready_count
* to indicate to the other VPEs that they may continue.
*/
if (coupled_coherence)
cps_gen_set_top_bit(&p, &l, &r, r_nc_count,
lbl_set_cont);
/*
* VPEs which did not disable coherence will continue
* executing, after coherence has been disabled, from this
* point.
*/
uasm_build_label(&l, p, lbl_secondary_cont);
/* Now perform our wait */
uasm_i_wait(&p, 0);
}
/*
* Re-enable coherence. Note that for CPS_PM_NC_WAIT all coupled VPEs
* will run this. The first will actually re-enable coherence & the
* rest will just be performing a rather unusual nop.
*/
uasm_i_addiu(&p, t0, zero, GCMP_CCB_COHCTL_DOMAIN_MSK);
uasm_i_sw(&p, t0, 0, r_pcohctl);
uasm_i_lw(&p, t0, 0, r_pcohctl);
/* Completion barrier */
uasm_i_sync(&p, stype_memory);
uasm_i_ehb(&p);
if (coupled_coherence && (state == CPS_PM_NC_WAIT)) {
/* Decrement ready_count */
uasm_build_label(&l, p, lbl_decready);
uasm_i_sync(&p, stype_ordering);
uasm_i_ll(&p, t1, 0, r_nc_count);
uasm_i_addiu(&p, t2, t1, -1);
uasm_i_sc(&p, t2, 0, r_nc_count);
uasm_il_beqz(&p, &r, t2, lbl_decready);
uasm_i_andi(&p, v0, t1, (1 << fls(smp_num_siblings)) - 1);
/* Ordering barrier */
uasm_i_sync(&p, stype_ordering);
}
if (coupled_coherence && (state == CPS_PM_CLOCK_GATED)) {
/*
* At this point it is safe for all VPEs to proceed with
* execution. This VPE will set the top bit of ready_count
* to indicate to the other VPEs that they may continue.
*/
cps_gen_set_top_bit(&p, &l, &r, r_nc_count, lbl_set_cont);
/*
* This core will be reliant upon another core sending a
* power-up command to the CPC in order to resume operation.
* Thus an arbitrary VPE can't trigger the core leaving the
* idle state and the one that disables coherence might as well
* be the one to re-enable it. The rest will continue from here
* after that has been done.
*/
uasm_build_label(&l, p, lbl_secondary_cont);
/* Ordering barrier */
uasm_i_sync(&p, stype_ordering);
}
/* The core is coherent, time to return to C code */
uasm_i_jr(&p, ra);
uasm_i_nop(&p);
gen_done:
/* Ensure the code didn't exceed the resources allocated for it */
BUG_ON((p - buf) > max_instrs);
BUG_ON((l - labels) > ARRAY_SIZE(labels));
BUG_ON((r - relocs) > ARRAY_SIZE(relocs));
/* Patch branch offsets */
uasm_resolve_relocs(relocs, labels);
/* Flush the icache */
local_flush_icache_range((unsigned long)buf, (unsigned long)p);
return buf;
out_err:
kfree(buf);
return NULL;
}
void build_copy_page(void)
{
int off;
u32 *buf = (u32 *)©_page_array;
struct uasm_label *l = labels;
struct uasm_reloc *r = relocs;
int i;
memset(labels, 0, sizeof(labels));
memset(relocs, 0, sizeof(relocs));
set_prefetch_parameters();
/*
* This algorithm makes the following assumptions:
* - All prefetch biases are multiples of 8 words.
* - The prefetch biases are less than one page.
* - The store prefetch bias isn't greater than the load
* prefetch bias.
*/
BUG_ON(pref_bias_copy_load % (8 * copy_word_size));
BUG_ON(pref_bias_copy_store % (8 * copy_word_size));
BUG_ON(PAGE_SIZE < pref_bias_copy_load);
BUG_ON(pref_bias_copy_store > pref_bias_copy_load);
off = PAGE_SIZE - pref_bias_copy_load;
if (off > 0xffff || !pref_bias_copy_load)
pg_addiu(&buf, A2, A0, off);
else
uasm_i_ori(&buf, A2, A0, off);
if (R4600_V2_HIT_CACHEOP_WAR && cpu_is_r4600_v2_x())
uasm_i_lui(&buf, AT, 0xa000);
off = cache_line_size ? min(8, pref_bias_copy_load / cache_line_size) *
cache_line_size : 0;
while (off) {
build_copy_load_pref(&buf, -off);
off -= cache_line_size;
}
off = cache_line_size ? min(8, pref_bias_copy_store / cache_line_size) *
cache_line_size : 0;
while (off) {
build_copy_store_pref(&buf, -off);
off -= cache_line_size;
}
uasm_l_copy_pref_both(&l, buf);
do {
build_copy_load_pref(&buf, off);
build_copy_load(&buf, T0, off);
build_copy_load_pref(&buf, off + copy_word_size);
build_copy_load(&buf, T1, off + copy_word_size);
build_copy_load_pref(&buf, off + 2 * copy_word_size);
build_copy_load(&buf, T2, off + 2 * copy_word_size);
build_copy_load_pref(&buf, off + 3 * copy_word_size);
build_copy_load(&buf, T3, off + 3 * copy_word_size);
build_copy_store_pref(&buf, off);
build_copy_store(&buf, T0, off);
build_copy_store_pref(&buf, off + copy_word_size);
build_copy_store(&buf, T1, off + copy_word_size);
build_copy_store_pref(&buf, off + 2 * copy_word_size);
build_copy_store(&buf, T2, off + 2 * copy_word_size);
build_copy_store_pref(&buf, off + 3 * copy_word_size);
build_copy_store(&buf, T3, off + 3 * copy_word_size);
off += 4 * copy_word_size;
} while (off < half_copy_loop_size);
pg_addiu(&buf, A1, A1, 2 * off);
pg_addiu(&buf, A0, A0, 2 * off);
off = -off;
do {
build_copy_load_pref(&buf, off);
build_copy_load(&buf, T0, off);
build_copy_load_pref(&buf, off + copy_word_size);
build_copy_load(&buf, T1, off + copy_word_size);
build_copy_load_pref(&buf, off + 2 * copy_word_size);
build_copy_load(&buf, T2, off + 2 * copy_word_size);
build_copy_load_pref(&buf, off + 3 * copy_word_size);
build_copy_load(&buf, T3, off + 3 * copy_word_size);
build_copy_store_pref(&buf, off);
build_copy_store(&buf, T0, off);
build_copy_store_pref(&buf, off + copy_word_size);
build_copy_store(&buf, T1, off + copy_word_size);
build_copy_store_pref(&buf, off + 2 * copy_word_size);
build_copy_store(&buf, T2, off + 2 * copy_word_size);
build_copy_store_pref(&buf, off + 3 * copy_word_size);
if (off == -(4 * copy_word_size))
uasm_il_bne(&buf, &r, A2, A0, label_copy_pref_both);
build_copy_store(&buf, T3, off + 3 * copy_word_size);
off += 4 * copy_word_size;
} while (off < 0);
if (pref_bias_copy_load - pref_bias_copy_store) {
pg_addiu(&buf, A2, A0,
pref_bias_copy_load - pref_bias_copy_store);
uasm_l_copy_pref_store(&l, buf);
off = 0;
do {
build_copy_load(&buf, T0, off);
build_copy_load(&buf, T1, off + copy_word_size);
build_copy_load(&buf, T2, off + 2 * copy_word_size);
build_copy_load(&buf, T3, off + 3 * copy_word_size);
build_copy_store_pref(&buf, off);
build_copy_store(&buf, T0, off);
build_copy_store_pref(&buf, off + copy_word_size);
build_copy_store(&buf, T1, off + copy_word_size);
build_copy_store_pref(&buf, off + 2 * copy_word_size);
build_copy_store(&buf, T2, off + 2 * copy_word_size);
build_copy_store_pref(&buf, off + 3 * copy_word_size);
build_copy_store(&buf, T3, off + 3 * copy_word_size);
off += 4 * copy_word_size;
} while (off < half_copy_loop_size);
pg_addiu(&buf, A1, A1, 2 * off);
pg_addiu(&buf, A0, A0, 2 * off);
off = -off;
do {
build_copy_load(&buf, T0, off);
build_copy_load(&buf, T1, off + copy_word_size);
build_copy_load(&buf, T2, off + 2 * copy_word_size);
build_copy_load(&buf, T3, off + 3 * copy_word_size);
build_copy_store_pref(&buf, off);
build_copy_store(&buf, T0, off);
build_copy_store_pref(&buf, off + copy_word_size);
build_copy_store(&buf, T1, off + copy_word_size);
build_copy_store_pref(&buf, off + 2 * copy_word_size);
build_copy_store(&buf, T2, off + 2 * copy_word_size);
build_copy_store_pref(&buf, off + 3 * copy_word_size);
if (off == -(4 * copy_word_size))
uasm_il_bne(&buf, &r, A2, A0,
label_copy_pref_store);
build_copy_store(&buf, T3, off + 3 * copy_word_size);
off += 4 * copy_word_size;
} while (off < 0);
}
if (pref_bias_copy_store) {
pg_addiu(&buf, A2, A0, pref_bias_copy_store);
uasm_l_copy_nopref(&l, buf);
off = 0;
do {
build_copy_load(&buf, T0, off);
build_copy_load(&buf, T1, off + copy_word_size);
build_copy_load(&buf, T2, off + 2 * copy_word_size);
build_copy_load(&buf, T3, off + 3 * copy_word_size);
build_copy_store(&buf, T0, off);
build_copy_store(&buf, T1, off + copy_word_size);
build_copy_store(&buf, T2, off + 2 * copy_word_size);
build_copy_store(&buf, T3, off + 3 * copy_word_size);
off += 4 * copy_word_size;
} while (off < half_copy_loop_size);
pg_addiu(&buf, A1, A1, 2 * off);
pg_addiu(&buf, A0, A0, 2 * off);
off = -off;
do {
build_copy_load(&buf, T0, off);
build_copy_load(&buf, T1, off + copy_word_size);
build_copy_load(&buf, T2, off + 2 * copy_word_size);
build_copy_load(&buf, T3, off + 3 * copy_word_size);
build_copy_store(&buf, T0, off);
build_copy_store(&buf, T1, off + copy_word_size);
build_copy_store(&buf, T2, off + 2 * copy_word_size);
if (off == -(4 * copy_word_size))
uasm_il_bne(&buf, &r, A2, A0,
label_copy_nopref);
build_copy_store(&buf, T3, off + 3 * copy_word_size);
off += 4 * copy_word_size;
} while (off < 0);
}
uasm_i_jr(&buf, RA);
uasm_i_nop(&buf);
BUG_ON(buf > copy_page_array + ARRAY_SIZE(copy_page_array));
uasm_resolve_relocs(relocs, labels);
pr_debug("Synthesized copy page handler (%u instructions).\n",
(u32)(buf - copy_page_array));
pr_debug("\t.set push\n");
pr_debug("\t.set noreorder\n");
for (i = 0; i < (buf - copy_page_array); i++)
pr_debug("\t.word 0x%08x\n", copy_page_array[i]);
pr_debug("\t.set pop\n");
}
void __cpuinit build_clear_page(void)
{
int off;
u32 *buf = (u32 *)&clear_page_array;
struct uasm_label *l = labels;
struct uasm_reloc *r = relocs;
int i;
memset(labels, 0, sizeof(labels));
memset(relocs, 0, sizeof(relocs));
if (current_cpu_data.cputype == CPU_CAVIUM_OCTEON2) {
const unsigned int wb_nudge = 26;
pg_addiu(&buf, T0, A0, PAGE_SIZE);
UASM_i_ADDIU(&buf, A1, A0, 128);
uasm_l_clear_pref(&l, buf);
uasm_i_zcbt(&buf, A0);
uasm_i_pref(&buf, wb_nudge, 0, A0);
UASM_i_ADDIU(&buf, A0, A0, 256);
uasm_i_zcbt(&buf, A1);
uasm_i_pref(&buf, wb_nudge, 0, A1);
UASM_i_ADDIU(&buf, A1, A1, 256);
uasm_i_zcbt(&buf, A0);
uasm_i_pref(&buf, wb_nudge, 0, A0);
UASM_i_ADDIU(&buf, A0, A0, 256);
uasm_i_zcbt(&buf, A1);
uasm_i_pref(&buf, wb_nudge, 0, A1);
UASM_i_ADDIU(&buf, A1, A1, 256);
uasm_i_zcbt(&buf, A0);
uasm_i_pref(&buf, wb_nudge, 0, A0);
UASM_i_ADDIU(&buf, A0, A0, 256);
uasm_i_zcbt(&buf, A1);
uasm_i_pref(&buf, wb_nudge, 0, A1);
UASM_i_ADDIU(&buf, A1, A1, 256);
uasm_i_zcbt(&buf, A0);
uasm_i_pref(&buf, wb_nudge, 0, A0);
UASM_i_ADDIU(&buf, A0, A0, 256);
uasm_i_zcbt(&buf, A1);
uasm_i_pref(&buf, wb_nudge, 0, A1);
uasm_il_bne(&buf, &r, A0, T0, label_clear_pref);
UASM_i_ADDIU(&buf, A1, A1, 256);
} else {
set_prefetch_parameters();
/*
* This algorithm makes the following assumptions:
* - The prefetch bias is a multiple of 2 words.
* - The prefetch bias is less than one page.
*/
BUG_ON(pref_bias_clear_store % (2 * clear_word_size));
BUG_ON(PAGE_SIZE < pref_bias_clear_store);
off = PAGE_SIZE - pref_bias_clear_store;
if (off > 0xffff || !pref_bias_clear_store)
pg_addiu(&buf, A2, A0, off);
else
uasm_i_ori(&buf, A2, A0, off);
if (R4600_V2_HIT_CACHEOP_WAR && cpu_is_r4600_v2_x())
uasm_i_lui(&buf, AT, 0xa000);
off = cache_line_size ? min(8, pref_bias_clear_store / cache_line_size)
* cache_line_size : 0;
while (off) {
build_clear_pref(&buf, -off);
off -= cache_line_size;
}
uasm_l_clear_pref(&l, buf);
do {
build_clear_pref(&buf, off);
build_clear_store(&buf, off);
off += clear_word_size;
} while (off < half_clear_loop_size);
pg_addiu(&buf, A0, A0, 2 * off);
off = -off;
do {
build_clear_pref(&buf, off);
if (off == -clear_word_size)
uasm_il_bne(&buf, &r, A0, A2, label_clear_pref);
build_clear_store(&buf, off);
off += clear_word_size;
} while (off < 0);
if (pref_bias_clear_store) {
pg_addiu(&buf, A2, A0, pref_bias_clear_store);
uasm_l_clear_nopref(&l, buf);
off = 0;
do {
build_clear_store(&buf, off);
off += clear_word_size;
} while (off < half_clear_loop_size);
pg_addiu(&buf, A0, A0, 2 * off);
off = -off;
do {
if (off == -clear_word_size)
uasm_il_bne(&buf, &r, A0, A2,
label_clear_nopref);
build_clear_store(&buf, off);
off += clear_word_size;
} while (off < 0);
}
}
uasm_i_jr(&buf, RA);
uasm_i_nop(&buf);
BUG_ON(buf > clear_page_array + ARRAY_SIZE(clear_page_array));
uasm_resolve_relocs(relocs, labels);
pr_debug("Synthesized clear page handler (%u instructions).\n",
(u32)(buf - clear_page_array));
pr_debug("\t.set push\n");
pr_debug("\t.set noreorder\n");
for (i = 0; i < (buf - clear_page_array); i++)
pr_debug("\t.word 0x%08x\n", clear_page_array[i]);
pr_debug("\t.set pop\n");
}
void build_clear_page(void)
{
int off;
u32 *buf = &__clear_page_start;
struct uasm_label *l = labels;
struct uasm_reloc *r = relocs;
int i;
static atomic_t run_once = ATOMIC_INIT(0);
if (atomic_xchg(&run_once, 1)) {
return;
}
memset(labels, 0, sizeof(labels));
memset(relocs, 0, sizeof(relocs));
set_prefetch_parameters();
/*
* This algorithm makes the following assumptions:
* - The prefetch bias is a multiple of 2 words.
* - The prefetch bias is less than one page.
*/
BUG_ON(pref_bias_clear_store % (2 * clear_word_size));
BUG_ON(PAGE_SIZE < pref_bias_clear_store);
off = PAGE_SIZE - pref_bias_clear_store;
if (off > 0xffff || !pref_bias_clear_store)
pg_addiu(&buf, A2, A0, off);
else
uasm_i_ori(&buf, A2, A0, off);
if (R4600_V2_HIT_CACHEOP_WAR && cpu_is_r4600_v2_x())
uasm_i_lui(&buf, AT, uasm_rel_hi(0xa0000000));
off = cache_line_size ? min(8, pref_bias_clear_store / cache_line_size)
* cache_line_size : 0;
while (off) {
build_clear_pref(&buf, -off);
off -= cache_line_size;
}
uasm_l_clear_pref(&l, buf);
do {
build_clear_pref(&buf, off);
build_clear_store(&buf, off);
off += clear_word_size;
} while (off < half_clear_loop_size);
pg_addiu(&buf, A0, A0, 2 * off);
off = -off;
do {
build_clear_pref(&buf, off);
if (off == -clear_word_size)
uasm_il_bne(&buf, &r, A0, A2, label_clear_pref);
build_clear_store(&buf, off);
off += clear_word_size;
} while (off < 0);
if (pref_bias_clear_store) {
pg_addiu(&buf, A2, A0, pref_bias_clear_store);
uasm_l_clear_nopref(&l, buf);
off = 0;
do {
build_clear_store(&buf, off);
off += clear_word_size;
} while (off < half_clear_loop_size);
pg_addiu(&buf, A0, A0, 2 * off);
off = -off;
do {
if (off == -clear_word_size)
uasm_il_bne(&buf, &r, A0, A2,
label_clear_nopref);
build_clear_store(&buf, off);
off += clear_word_size;
} while (off < 0);
}
uasm_i_jr(&buf, RA);
uasm_i_nop(&buf);
BUG_ON(buf > &__clear_page_end);
uasm_resolve_relocs(relocs, labels);
pr_debug("Synthesized clear page handler (%u instructions).\n",
(u32)(buf - &__clear_page_start));
pr_debug("\t.set push\n");
pr_debug("\t.set noreorder\n");
for (i = 0; i < (buf - &__clear_page_start); i++)
pr_debug("\t.word 0x%08x\n", (&__clear_page_start)[i]);
pr_debug("\t.set pop\n");
}
