MIPS: DEC/SNI: O32 wrapper stack switching fixes

Commit 231a35d37293ab88d325a9cb94e5474c156282c0 [[MIPS] RM: Collected
changes] broke DECstation support by introducing an incompatible copy of
arch/mips/dec/prom/call_o32.S in arch/mips/fw/lib/, built unconditionally.
The copy happens to land earlier of the two among the modules used in the
link and is therefore chosen for the DECstation rather than the intended
original.  As a result random kernel data is corrupted because a pointer
to the "%s" formatted output template is used as a temporary stack pointer
rather than being passed down to prom_printf.  This also explains why
prom_printf still works, up to a point -- the next argument is the actual
string to output so it works just fine as the output template until enough
kernel data has been corrupted to cause a crash.

This change adjusts the modified wrapper in arch/mips/fw/lib/call_o32.S to
let callers request no stack switching by passing a null temporary stack
pointer in $a1, reworks the DECstation callers to work with the updated
interface and removes the old copy from arch/mips/dec/prom/call_o32.S.  A
few minor readability adjustments are included as well, most importantly
O32_SZREG is now used throughout where applicable rather than hardcoded
multiplies of 4 and $fp is used to access the argument save area as a more
usual register to operate the stack with rather than $s0.

Finally an update is made to the temporary stack space used by the SNI
platform to guarantee 8-byte alignment as per o32 requirements.

Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/6668/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

2 files changed, 37 insertions, 23 deletions

diff --git a/arch/mips/fw/lib/call_o32.S b/arch/mips/fw/lib/call_o32.S
index b308b2a0613e..4703fe4dbd9a 100644
--- a/arch/mips/fw/lib/call_o32.S
+++ b/arch/mips/fw/lib/call_o32.S
@@ -1,7 +1,7 @@
 /*
  *      O32 interface for the 64 (or N32) ABI.
  *
- *      Copyright (C) 2002  Maciej W. Rozycki
+ *      Copyright (C) 2002, 2014  Maciej W. Rozycki
  *
  *      This program is free software; you can redistribute it and/or
  *      modify it under the terms of the GNU General Public License
@@ -12,28 +12,37 @@
 #include <asm/asm.h>
 #include <asm/regdef.h>
 
+/* O32 register size.  */
+#define O32_SZREG       4
 /* Maximum number of arguments supported.  Must be even!  */
 #define O32_ARGC        32
-/* Number of static registers we save.  */
+/* Number of static registers we save.  */
 #define O32_STATC       11
-/* Frame size for static register  */
-#define O32_FRAMESZ     (SZREG * O32_STATC)
-/* Frame size on new stack */
-#define O32_FRAMESZ_NEW (SZREG + 4 * O32_ARGC)
+/* Argument area frame size.  */
+#define O32_ARGSZ       (O32_SZREG * O32_ARGC)
+/* Static register save area frame size.  */
+#define O32_STATSZ      (SZREG * O32_STATC)
+/* Stack pointer register save area frame size.  */
+#define O32_SPSZ        SZREG
+/* Combined area frame size.  */
+#define O32_FRAMESZ     (O32_ARGSZ + O32_SPSZ + O32_STATSZ)
+/* Switched stack frame size.  */
+#define O32_NFRAMESZ    (O32_ARGSZ + O32_SPSZ)
 
                 .text
 
 /*
  * O32 function call dispatcher, for interfacing 32-bit ROM routines.
  *
- * The standard 64 (N32) calling sequence is supported, with a0
- * holding a function pointer, a1 a new stack pointer, a2-a7 -- its
- * first six arguments and the stack -- remaining ones (up to O32_ARGC,
- * including a2-a7). Static registers, gp and fp are preserved, v0 holds
- * a result. This code relies on the called o32 function for sp and ra
- * restoration and this dispatcher has to be placed in a KSEGx (or KUSEG)
- * address space.  Any pointers passed have to point to addresses within
- * one of these spaces as well.
+ * The standard 64 (N32) calling sequence is supported, with a0 holding
+ * a function pointer, a1 a pointer to the new stack to call the
+ * function with or 0 if no stack switching is requested, a2-a7 -- the
+ * function call's first six arguments, and the stack -- the remaining
+ * arguments (up to O32_ARGC, including a2-a7).  Static registers, gp
+ * and fp are preserved, v0 holds the result.  This code relies on the
+ * called o32 function for sp and ra restoration and this dispatcher has
+ * to be placed in a KSEGx (or KUSEG) address space.  Any pointers
+ * passed have to point to addresses within one of these spaces as well.
  */
 NESTED(call_o32, O32_FRAMESZ, ra)
                 REG_SUBU        sp,O32_FRAMESZ
@@ -51,32 +60,36 @@ NESTED(call_o32, O32_FRAMESZ, ra)
                 REG_S           s0,O32_FRAMESZ-11*SZREG(sp)
 
                 move            jp,a0
-                REG_SUBU        s0,a1,O32_FRAMESZ_NEW
-                REG_S           sp,O32_FRAMESZ_NEW-1*SZREG(s0)
+
+                move            fp,sp
+                beqz            a1,0f
+                REG_SUBU        fp,a1,O32_NFRAMESZ
+0:
+                REG_S           sp,O32_NFRAMESZ-1*SZREG(fp)
 
                 sll             a0,a2,zero
                 sll             a1,a3,zero
                 sll             a2,a4,zero
                 sll             a3,a5,zero
-                sw              a6,0x10(s0)
-                sw              a7,0x14(s0)
+                sw              a6,4*O32_SZREG(fp)
+                sw              a7,5*O32_SZREG(fp)
 
                 PTR_LA          t0,O32_FRAMESZ(sp)
-                PTR_LA          t1,0x18(s0)
+                PTR_LA          t1,6*O32_SZREG(fp)
                 li              t2,O32_ARGC-6
 1:
                 lw              t3,(t0)
                 REG_ADDU        t0,SZREG
                 sw              t3,(t1)
                 REG_SUBU        t2,1
-                REG_ADDU        t1,4
+                REG_ADDU        t1,O32_SZREG
                 bnez            t2,1b
 
-                move            sp,s0
+                move            sp,fp
 
                 jalr            jp
 
-                REG_L           sp,O32_FRAMESZ_NEW-1*SZREG(sp)
+                REG_L           sp,O32_NFRAMESZ-1*SZREG(sp)
 
                 REG_L           s0,O32_FRAMESZ-11*SZREG(sp)
                 REG_L           s1,O32_FRAMESZ-10*SZREG(sp)
diff --git a/arch/mips/fw/sni/sniprom.c b/arch/mips/fw/sni/sniprom.c
index 2c2cb182af4e..6aa264b9856a 100644
--- a/arch/mips/fw/sni/sniprom.c
+++ b/arch/mips/fw/sni/sniprom.c
@@ -40,7 +40,8 @@
 
 #ifdef CONFIG_64BIT
 
-static u8 o32_stk[16384];
+/* O32 stack has to be 8-byte aligned. */
+static u64 o32_stk[4096];
 #define O32_STK   &o32_stk[sizeof(o32_stk)]
 
 #define __PROM_O32(fun, arg) fun arg __asm__(#fun); \