[PATCH] Fix process crash in 2.4 on attempt to use FPU on MIPS32

[PATCH] Fix process crash in 2.4 on attempt to use FPU on MIPS32

[Sergei Shtylyov]

[-- Attachment #1: Type: text/plain, Size: 403 bytes --]

If there's built-in FPU in a MIPS32 CPU the first time the process tries
to use it, the kernel should crash with "reserved instruction" -- CPU will try
to execute 'dmtc1' which is a MIPS64 only insn. _init_fpu() was apprently 
blindly copied form arch/mips64/... :-)

Since this occured with GXemul recently resending this 1.5 year old patch.

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>



[-- Attachment #2: MIPS32-FPU-init-2.4.patch --]
[-- Type: text/plain, Size: 1217 bytes --]

diff --git a/arch/mips/kernel/r4k_switch.S b/arch/mips/kernel/r4k_switch.S
index 1999483..30b67c8 100644
--- a/arch/mips/kernel/r4k_switch.S
+++ b/arch/mips/kernel/r4k_switch.S
@@ -134,7 +134,6 @@ LEAF(_restore_fp)
 #define FPU_DEFAULT  0x00000000
 
 LEAF(_init_fpu)
-	.set	mips3
 	mfc0	t0, CP0_STATUS
 	li	t1, ST0_CU1
 	or	t0, t1
@@ -146,24 +145,40 @@ LEAF(_init_fpu)
 
 	li	t0, -1
 
-	dmtc1	t0, $f0
-	dmtc1	t0, $f2
-	dmtc1	t0, $f4
-	dmtc1	t0, $f6
-	dmtc1	t0, $f8
-	dmtc1	t0, $f10
-	dmtc1	t0, $f12
-	dmtc1	t0, $f14
-	dmtc1	t0, $f16
-	dmtc1	t0, $f18
-	dmtc1	t0, $f20
-	dmtc1	t0, $f22
-	dmtc1	t0, $f24
-	dmtc1	t0, $f26
-	dmtc1	t0, $f28
+	mtc1	t0, $f0
+	mtc1	t0, $f1
+	mtc1	t0, $f2
+	mtc1	t0, $f3
+	mtc1	t0, $f4
+	mtc1	t0, $f5
+	mtc1	t0, $f6
+	mtc1	t0, $f7
+	mtc1	t0, $f8
+	mtc1	t0, $f9
+	mtc1	t0, $f10
+	mtc1	t0, $f11
+	mtc1	t0, $f12
+	mtc1	t0, $f13
+	mtc1	t0, $f14
+	mtc1	t0, $f15
+	mtc1	t0, $f16
+	mtc1	t0, $f17
+	mtc1	t0, $f18
+	mtc1	t0, $f19
+	mtc1	t0, $f20
+	mtc1	t0, $f21
+	mtc1	t0, $f22
+	mtc1	t0, $f23
+	mtc1	t0, $f24
+	mtc1	t0, $f25
+	mtc1	t0, $f26
+	mtc1	t0, $f27
+	mtc1	t0, $f28
+	mtc1	t0, $f29
+	mtc1	t0, $f30
 	.set	noreorder
 	jr	ra
-	 dmtc1	t0, $f30
 	.set	reorder
+	 mtc1	t0, $f31
 	END(_init_fpu)
 

Re: [PATCH] Fix process crash in 2.4 on attempt to use FPU on MIPS32

[Ralf Baechle]

On Fri, Jul 07, 2006 at 01:25:14AM +0400, Sergei Shtylyov wrote:

> If there's built-in FPU in a MIPS32 CPU the first time the process tries
> to use it, the kernel should crash with "reserved instruction" -- CPU will 
> try
> to execute 'dmtc1' which is a MIPS64 only insn. _init_fpu() was apprently 
> blindly copied form arch/mips64/... :-)
> 
> Since this occured with GXemul recently resending this 1.5 year old patch.

I didn't like the patch back then because it drops the optimizations
for 64-bit processors.  So I just took the 2.6 variant of the code and
bolted it into 2.4.

  Ralf