[Qemu-devel] [PATCH risu v2 0/3] Miscellaneous fixes for powerpc64

[Qemu-devel] [PATCH risu v2 0/3] Miscellaneous fixes for powerpc64

[Sandipan Das]

The first patch adds the patterns for the missing byte and
doubleword load instructions. The second patch fixes mismatches
encountered when using the load quadword instruction. The last
patch fixes a minor typo in the floating-point register prefix
shown when dumping the register values.

Changelog:
v2 -> Added a cover letter as requested by Peter.

Sandipan Das (3):
  ppc64.risu: Add missing byte and dword loads
  ppc64.risu: Fix pattern for load qword
  risu_reginfo_ppc64.c: Fix register name prefix

 ppc64.risu           | 29 +++++++++++++++++++++++++++--
 risu_reginfo_ppc64.c |  4 ++--
 2 files changed, 29 insertions(+), 4 deletions(-)

-- 
2.14.3

[Qemu-devel] [PATCH risu v2 1/3] ppc64.risu: Add missing byte and dword loads

[Sandipan Das]

The patterns for the following instructions are added:
 * Load Byte and Zero (lbz)
 * Load Byte and Zero with Update (lbzu)
 * Load Byte and Zero Indexed (lbzx)
 * Load Byte and Zero with Update Indexed (lbzux)
 * Load Doubleword (ld)

Signed-off-by: Sandipan Das <sandipan@linux.vnet.ibm.com>
---
 ppc64.risu | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)

diff --git a/ppc64.risu b/ppc64.risu
index e2fd4f6..13b95ac 100644
--- a/ppc64.risu
+++ b/ppc64.risu
@@ -887,6 +887,31 @@ FTSQRT PPC64LE 111111 bf:3 0000000 frb:5 00101000000
 ISEL PPC64LE 011111 rt:5 ra:5 rb:5 bc:5 011110 \
 !constraints { $rt != 1 && $ra != 1 && $rb != 1 && $rt != 13 && $ra != 13 && $rb != 13; }
 
+# format:D book:I page:48 v:P1 lbz Load Byte & Zero
+LBZ PPC64LE 100010 rt:5 ra:5 imm:16 \
+!constraints { $rt != 1 && $ra != 1 && $rt != 13 && $ra != 13 && $ra != 0 && $ra != $rt && $imm <= 32752; } \
+!memory { reg_plus_imm($ra, $imm); }
+
+# format:D book:I page:48 v:P1 lbzu Load Byte & Zero with Update
+LBZU PPC64LE 100011 rt:5 ra:5 imm:16 \
+!constraints { $rt != 1 && $ra != 1 && $rt != 13 && $ra != 13 && $ra != 0 && $ra != $rt && $imm <= 32752; } \
+!memory { reg_plus_imm($ra, $imm); }
+
+# format:X book:I page:49 v:P1 lbzux Load Byte & Zero with Update Indexed
+LBZUX PPC64LE 011111 rt:5 ra:5 rb:5 00011101110 \
+!constraints { $rt != 1 && $ra != 1 && $rb != 1 && $rt != 13 && $ra != 13 && $rb != 13 && $ra != 0 && $ra != $rt && $ra != $rb && $rt != $rb; } \
+!memory { reg_plus_reg($ra, $rb); }
+
+# format:X book:I page:49 v:P1 lbzx Load Byte & Zero Indexed
+LBZX PPC64LE 011111 rt:5 ra:5 rb:5 00010101110 \
+!constraints { $rt != 1 && $ra != 1 && $rb != 1 && $rt != 13 && $ra != 13 && $rb != 13 && $ra != 0 && $ra != $rt && $ra != $rb && $rt != $rb; } \
+!memory { reg_plus_reg($ra, $rb); }
+
+# format:DS book:I page:53 PPC ld Load Dword
+LD PPC64LE 111010 rt:5 ra:5 imm:14 00 \
+!constraints { $rt != 1 && $ra != 1 && $rt != 13 && $ra != 13 && $ra != 0 && $ra != $rt && $imm <= 8176; } \
+!memory { reg_plus_imm($ra, $imm << 2); }
+
 # format:X book:I page:62 v2.06 ldbrx Load Dword Byte-Reverse Indexed
 LDBRX PPC64LE 011111 rt:5 ra:5 rb:5 10000101000 \
 !constraints { $rt != 1 && $ra != 1 && $rb != 1 && $rt != 13 && $ra != 13 && $rb != 13 && $ra != 0 && $ra != $rt && $ra != $rb && $rt != $rb; } \
-- 
2.14.3

[Qemu-devel] [PATCH risu v2 2/3] ppc64.risu: Fix pattern for load qword

[Sandipan Das]

The pattern for the Load Quadword (lq) instruction is fixed.

If rtp is 0 or 12, the instruction will overwrite r0, r1 or
r12, r13 respectively. However, r1 is the stack frame pointer
and r13 is the thread pointer. So, overwriting them can cause
a crash. This is avoided by putting a constraint to prevent
rtp from being 0 or 12.

For a given effective address (ea), this instruction loads
two dwords from ea and ea+8. However, if ea is the start
address of the current stack frame, then the value of the
back chain dword from the previous stack frame, which is at
ea+8, is loaded on to one of the registers. This can cause
a mismatch as the addresses may vary across the master and
the apprentice instances. This is avoided by always adding
8 to the offset used for calculating the ea.

Signed-off-by: Sandipan Das <sandipan@linux.vnet.ibm.com>
---
 ppc64.risu | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/ppc64.risu b/ppc64.risu
index 13b95ac..2018103 100644
--- a/ppc64.risu
+++ b/ppc64.risu
@@ -1050,8 +1050,8 @@ LHZX PPC64LE 011111 rt:5 ra:5 rb:5 01000101110 \
 
 # format:DQ book:I page:59 v2.03 lq Load Qword
 LQ PPC64LE 111000 rtp:5 ra:5 imm:12 0000 \
-!constraints { $rtp % 2 == 0 && $ra != 1 && $ra != 13 && $ra != 0 && $ra != $rtp && $imm <= 2032; } \
-!memory { reg_plus_imm($ra, $imm << 4); }
+!constraints { $rtp % 2 == 0 && $rtp != 0 && $rtp != 12 && $ra != 1 && $ra != 13 && $ra != 0 && $ra != $rtp && $imm <= 2032; } \
+!memory { reg_plus_imm($ra, ($imm << 4) + 8); }
 
 # format:X book:I page:65 v:P1 lswi Load String Word Immediate
 LSWI PPC64LE 011111 rt:5 ra:5 rb:5 10010101010 \
-- 
2.14.3

[Qemu-devel] [PATCH risu v2 3/3] risu_reginfo_ppc64.c: Fix register name prefix

[Sandipan Das]

Use 'f' instead of 'r' as the prefix when dumping the values
of floating-point registers.

Signed-off-by: Sandipan Das <sandipan@linux.vnet.ibm.com>
---
 risu_reginfo_ppc64.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/risu_reginfo_ppc64.c b/risu_reginfo_ppc64.c
index eb9c12b..5f33648 100644
--- a/risu_reginfo_ppc64.c
+++ b/risu_reginfo_ppc64.c
@@ -122,7 +122,7 @@ int reginfo_dump(struct reginfo *ri, FILE * f)
     fprintf(f, "\tdscr   : %16lx\n\n", ri->gregs[44]);
 
     for (i = 0; i < 16; i++) {
-        fprintf(f, "\tf%2d: %.4f\tr%2d: %.4f\n", i, ri->fpregs[i],
+        fprintf(f, "\tf%2d: %.4f\tf%2d: %.4f\n", i, ri->fpregs[i],
                 i + 16, ri->fpregs[i + 16]);
     }
     fprintf(f, "\tfpscr: %f\n\n", ri->fpregs[32]);
@@ -167,7 +167,7 @@ int reginfo_dump_mismatch(struct reginfo *m, struct reginfo *a, FILE *f)
         }
 
         if (m->fpregs[i] != a->fpregs[i]) {
-            fprintf(f, "Mismatch: Register r%d\n", i);
+            fprintf(f, "Mismatch: Register f%d\n", i);
             fprintf(f, "m: [%f] != a: [%f]\n", m->fpregs[i], a->fpregs[i]);
         }
     }
-- 
2.14.3

Re: [Qemu-devel] [PATCH risu v2 0/3] Miscellaneous fixes for powerpc64

[Peter Maydell]

On 6 March 2018 at 12:08, Sandipan Das <sandipan@linux.vnet.ibm.com> wrote:
> The first patch adds the patterns for the missing byte and
> doubleword load instructions. The second patch fixes mismatches
> encountered when using the load quadword instruction. The last
> patch fixes a minor typo in the floating-point register prefix
> shown when dumping the register values.
>
> Changelog:
> v2 -> Added a cover letter as requested by Peter.
>
> Sandipan Das (3):
>   ppc64.risu: Add missing byte and dword loads
>   ppc64.risu: Fix pattern for load qword
>   risu_reginfo_ppc64.c: Fix register name prefix
>
>  ppc64.risu           | 29 +++++++++++++++++++++++++++--
>  risu_reginfo_ppc64.c |  4 ++--
>  2 files changed, 29 insertions(+), 4 deletions(-)

Thanks; I've applied these to risu master.

-- PMM