MMU: Implement a vestigial partition table

This implements a 1-entry partition table, so that instead of getting
the process table base address from the PRTBL SPR, the MMU now reads
the doubleword pointed to by the PTCR register plus 8 to get the
process table base address.  The partition table entry is cached.

Having the PTCR and the vestigial partition table reduces the amount
of software change required in Linux for Microwatt support.

Signed-off-by: Paul Mackerras <paulus@ozlabs.org>

common.vhdl

@@ -51,7 +51,7 @@ package common is
     constant SPR_HSPRG0 : spr_num_t := 304;
     constant SPR_HSPRG1 : spr_num_t := 305;
     constant SPR_PID    : spr_num_t := 48;
-    constant SPR_PRTBL  : spr_num_t := 720;
+    constant SPR_PTCR   : spr_num_t := 464;
     constant SPR_PVR	: spr_num_t := 287;
 
     -- GPR indices in the register file (GPR only)

decode1.vhdl

@@ -609,7 +609,7 @@ begin
                     vi.force_single := '1';
                     -- send MMU-related SPRs to loadstore1
                     case sprn is
-                        when SPR_DAR | SPR_DSISR | SPR_PID | SPR_PRTBL =>
+                        when SPR_DAR | SPR_DSISR | SPR_PID | SPR_PTCR =>
                             vi.override_decode.unit := LDST;
                             vi.override_unit := '1';
                         when others =>

loadstore1.vhdl

@@ -477,7 +477,7 @@ begin
                 v.read_spr := '1';
             when OP_MTSPR =>
                 v.write_spr := '1';
-                v.mmu_op := sprn(9) or sprn(5);
+                v.mmu_op := sprn(8) or sprn(5);
             when OP_FETCH_FAILED =>
                 -- send it to the MMU to do the radix walk
                 v.instr_fault := '1';
@@ -732,7 +732,7 @@ begin
                 write_enable := '1';
                 -- partial decode on SPR number should be adequate given
                 -- the restricted set that get sent down this path
-                if r2.req.sprn(9) = '0' and r2.req.sprn(5) = '0' then
+                if r2.req.sprn(8) = '0' and r2.req.sprn(5) = '0' then
                     if r2.req.sprn(0) = '0' then
                         sprval := x"00000000" & r3.dsisr;
                     else

mmu.vhdl

@@ -29,6 +29,9 @@ architecture behave of mmu is
     type state_t is (IDLE,
                      DO_TLBIE,
                      TLB_WAIT,
+                     PART_TBL_READ,
+                     PART_TBL_WAIT,
+                     PART_TBL_DONE,
                      PROC_TBL_READ,
                      PROC_TBL_WAIT,
                      SEGMENT_CHECK,
@@ -47,12 +50,14 @@ architecture behave of mmu is
         addr      : std_ulogic_vector(63 downto 0);
         inval_all : std_ulogic;
         -- config SPRs
-        prtbl     : std_ulogic_vector(63 downto 0);
+        ptcr      : std_ulogic_vector(63 downto 0);
         pid       : std_ulogic_vector(31 downto 0);
         -- internal state
         state     : state_t;
         done      : std_ulogic;
         err       : std_ulogic;
+        prtbl     : std_ulogic_vector(63 downto 0);
+        ptb_valid : std_ulogic;
         pgtbl0    : std_ulogic_vector(63 downto 0);
         pt0_valid : std_ulogic;
         pgtbl3    : std_ulogic_vector(63 downto 0);
@@ -77,7 +82,7 @@ architecture behave of mmu is
 begin
     -- Multiplex internal SPR values back to loadstore1, selected
     -- by l_in.sprn.
-    l_out.sprval <= r.prtbl when l_in.sprn(9) = '1' else x"00000000" & r.pid;
+    l_out.sprval <= r.ptcr when l_in.sprn(8) = '1' else x"00000000" & r.pid;
 
     mmu_0: process(clk)
     begin
@@ -85,9 +90,10 @@ begin
             if rst = '1' then
                 r.state <= IDLE;
                 r.valid <= '0';
+                r.ptb_valid <= '0';
                 r.pt0_valid <= '0';
                 r.pt3_valid <= '0';
-                r.prtbl <= (others => '0');
+                r.ptcr <= (others => '0');
                 r.pid <= (others => '0');
             else
                 if rin.valid = '1' then
@@ -185,6 +191,7 @@ begin
         variable tlb_load : std_ulogic;
         variable itlb_load : std_ulogic;
         variable tlbie_req : std_ulogic;
+        variable ptbl_rd : std_ulogic;
         variable prtbl_rd : std_ulogic;
         variable pt_valid : std_ulogic;
         variable effpid : std_ulogic_vector(31 downto 0);
@@ -215,6 +222,7 @@ begin
         itlb_load := '0';
         tlbie_req := '0';
         v.inval_all := '0';
+        ptbl_rd := '0';
         prtbl_rd := '0';
 
         -- Radix tree data structures in memory are big-endian,
@@ -256,11 +264,15 @@ begin
                     if l_in.sprn(3) = '1' then
                         v.pt0_valid := '0';
                         v.pt3_valid := '0';
+                        v.ptb_valid := '0';
                     end if;
                     v.state := DO_TLBIE;
                 else
                     v.valid := '1';
-                    if pt_valid = '0' then
+                    if r.ptb_valid = '0' then
+                        -- need to fetch process table base from partition table
+                        v.state := PART_TBL_READ;
+                    elsif pt_valid = '0' then
                         -- need to fetch process table entry
                         -- set v.shift so we can use finalmask for generating
                         -- the process table entry address
@@ -277,13 +289,14 @@ begin
             end if;
             if l_in.mtspr = '1' then
                 -- Move to PID needs to invalidate L1 TLBs and cached
-                -- pgtbl0 value.  Move to PRTBL does that plus
+                -- pgtbl0 value.  Move to PTCR does that plus
-                -- invalidating the cached pgtbl3 value as well.
+                -- invalidating the cached pgtbl3 and prtbl values as well.
-                if l_in.sprn(9) = '0' then
+                if l_in.sprn(8) = '0' then
                     v.pid := l_in.rs(31 downto 0);
                 else
-                    v.prtbl := l_in.rs;
+                    v.ptcr := l_in.rs;
                     v.pt3_valid := '0';
+                    v.ptb_valid := '0';
                 end if;
                 v.pt0_valid := '0';
                 v.inval_all := '1';
@@ -300,6 +313,22 @@ begin
                 v.state := RADIX_FINISH;
             end if;
 
+        when PART_TBL_READ =>
+            dcreq := '1';
+            ptbl_rd := '1';
+            v.state := PART_TBL_WAIT;
+
+        when PART_TBL_WAIT =>
+            if d_in.done = '1' then
+                v.prtbl := data;
+                v.ptb_valid := '1';
+                v.state := PART_TBL_DONE;
+            end if;
+
+        when PART_TBL_DONE =>
+            v.shift := unsigned('0' & r.prtbl(4 downto 0));
+            v.state := PROC_TBL_READ;
+
         when PROC_TBL_READ =>
             dcreq := '1';
             prtbl_rd := '1';
@@ -449,6 +478,9 @@ begin
         elsif tlb_load = '1' then
             addr := r.addr(63 downto 12) & x"000";
             tlb_data := pte;
+        elsif ptbl_rd = '1' then
+            addr := x"00" & r.ptcr(55 downto 12) & x"008";
+            tlb_data := (others => '0');
         elsif prtbl_rd = '1' then
             addr := prtable_addr;
             tlb_data := (others => '0');

tests/mmu/mmu.c

@@ -24,7 +24,7 @@ static inline void do_tlbie(unsigned long rb, unsigned long rs)
 #define SRR0	26
 #define SRR1	27
 #define PID	48
-#define PRTBL	720
+#define PTCR	464
 
 static inline unsigned long mfspr(int sprnum)
 {
@@ -115,15 +115,18 @@ void zero_memory(void *ptr, unsigned long nbytes)
  */
 unsigned long *pgdir = (unsigned long *) 0x10000;
 unsigned long *proc_tbl = (unsigned long *) 0x12000;
-unsigned long free_ptr = 0x13000;
+unsigned long *part_tbl = (unsigned long *) 0x13000;
+unsigned long free_ptr = 0x14000;
 void *eas_mapped[4];
 int neas_mapped;
 
 void init_mmu(void)
 {
+	/* set up partition table */
+	store_pte(&part_tbl[1], (unsigned long)proc_tbl);
 	/* set up process table */
 	zero_memory(proc_tbl, 512 * sizeof(unsigned long));
-	mtspr(PRTBL, (unsigned long)proc_tbl);
+	mtspr(PTCR, (unsigned long)part_tbl);
 	mtspr(PID, 1);
 	zero_memory(pgdir, 1024 * sizeof(unsigned long));
 	/* RTS = 0 (2GB address space), RPDS = 10 (1024-entry top level) */

tests/modes/modes.c

@@ -33,7 +33,7 @@ static inline void do_tlbie(unsigned long rb, unsigned long rs)
 #define SPRG0	272
 #define SPRG1	273
 #define SPRG3	275
-#define PRTBL	720
+#define PTCR	464
 
 static inline unsigned long mfspr(int sprnum)
 {
@@ -121,15 +121,18 @@ void zero_memory(void *ptr, unsigned long nbytes)
  * Set up an MMU translation tree using memory starting at the 64k point.
  * We use 3 levels, mapping 512GB, with 4kB PGD/PMD/PTE pages.
  */
-unsigned long *proc_tbl = (unsigned long *) 0x10000;
+unsigned long *part_tbl = (unsigned long *) 0x10000;
-unsigned long *pgdir = (unsigned long *) 0x11000;
+unsigned long *proc_tbl = (unsigned long *) 0x11000;
-unsigned long free_ptr = 0x12000;
+unsigned long *pgdir = (unsigned long *) 0x12000;
+unsigned long free_ptr = 0x13000;
 
 void init_mmu(void)
 {
+	/* set up partition table */
+	store_pte(&part_tbl[1], (unsigned long)proc_tbl);
 	/* set up process table */
 	zero_memory(proc_tbl, 512 * sizeof(unsigned long));
-	mtspr(PRTBL, (unsigned long)proc_tbl);
+	mtspr(PTCR, (unsigned long)part_tbl);
 	mtspr(PID, 1);
 	zero_memory(pgdir, 512 * sizeof(unsigned long));
 	/* RTS = 8 (512GB address space), RPDS = 9 (512-entry top level) */

tests/privileged/privileged.c

@@ -14,7 +14,7 @@ extern int call_with_msr(unsigned long arg, int (*fn)(unsigned long), unsigned l
 #define SRR0	26
 #define SRR1	27
 #define PID	48
-#define PRTBL	720
+#define PTCR	464
 #define PVR	287
 
 static inline unsigned long mfspr(int sprnum)
@@ -106,15 +106,18 @@ void zero_memory(void *ptr, unsigned long nbytes)
  */
 unsigned long *pgdir = (unsigned long *) 0x10000;
 unsigned long *proc_tbl = (unsigned long *) 0x12000;
-unsigned long free_ptr = 0x13000;
+unsigned long *part_tbl = (unsigned long *) 0x13000;
+unsigned long free_ptr = 0x14000;
 
 void init_mmu(void)
 {
+	/* set up partition table */
+	store_pte(&part_tbl[1], (unsigned long)proc_tbl);
 	/* set up process table */
 	zero_memory(proc_tbl, 512 * sizeof(unsigned long));
 	/* RTS = 0 (2GB address space), RPDS = 10 (1024-entry top level) */
 	store_pte(&proc_tbl[2 * 1], (unsigned long) pgdir | 10);
-	mtspr(PRTBL, (unsigned long)proc_tbl);
+	mtspr(PTCR, (unsigned long)part_tbl);
 	mtspr(PID, 1);
 	zero_memory(pgdir, 1024 * sizeof(unsigned long));
 }

tests/reservation/reservation.c

@@ -18,7 +18,6 @@ extern unsigned long do_stqcx(unsigned long dst, unsigned long regs);
 #define PID	48
 #define SPRG0	272
 #define SPRG1	273
-#define PRTBL	720
 
 static inline unsigned long mfspr(int sprnum)
 {

tests/spr_read/spr_read.c

@@ -38,7 +38,7 @@ void print_test(char *str)
 #define SPR_HSPRG0	304
 #define SPR_HSPRG1	305
 #define SPR_PID		48
-#define SPR_PRTBL	720
+#define SPR_PTCR	464
 #define SPR_PVR		287
 
 #define __stringify_1(x...)	#x
@@ -83,7 +83,7 @@ int main(void)
 	DO_ONE(SPR_HSPRG0);
 	DO_ONE(SPR_HSPRG1);
 	DO_ONE(SPR_PID);
-	DO_ONE(SPR_PRTBL);
+	DO_ONE(SPR_PTCR);
 	DO_ONE(SPR_PVR);
 
 	puts(PASS);

tests/test_spr_read.console_out

@@ -20,6 +20,6 @@ Test SPR_SPRG3U:PASS
 Test SPR_HSPRG0:PASS
 Test SPR_HSPRG1:PASS
 Test SPR_PID:PASS
-Test SPR_PRTBL:PASS
+Test SPR_PTCR:PASS
 Test SPR_PVR:PASS
 PASS