[Bug target/32577] [SH] static inline attribute causes stack overflow
--- Comment #2 from saito at densan dot co dot jp 2007-07-23 01:23 ---
How about following testcase.
I list the value which moves current sp as to some architectures.
In the following case, listed value is 24.
> add #-24,r15 <-- stack grows down
With the result that only gcc-4.2.0 for sh4 accumulates the stack space for
inlined variables. Actually, I need to spread the stack size for sh-linux's
kernel-thread from 0x2000 byte to 0x4000 byte, although I specify optimization
option(-O2). Gcc-4.1.2 does not have this problem.
-testcase--
struct xxx {
int x;
int y;
};
static inline __attribute__((always_inline)) int add(int x, int y)
{
struct xxx xxx;
xxx.x = x;
xxx.y = y;
dummy(&xxx);
return x + y;
}
int x, y;
main(int ac, char** av)
{
x = add(x, y); /* count of add is one */
return x;
}
-end testcase--
compiling options: gcc test.c -O2 -S
case of gcc-4.2.0
---
count of add | 1 2 4 8 16
-
the valuesh4 | 8 16 24 40 68
ppc | 48 64 64 64 64
i686 | 16 16 16 16 16
case of gcc-4.1.2
---
the valuesh4 | 8 8 8 8 8
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32577
[Bug target/32577] New: [SH] static inline attribute causes stack overflow
When I use static inline attribute without any optimization as follows, stack
space is stacked. As a result, a stack overflow is caused. This problem came
from using gcc-4.2.0 for linux kernel which uses many inlines.
-compile options-
sh4-linux-gcc -S test.c
-testcase #1---
static inline __attribute__((always_inline)) int add(int x, int y)
{
return x + y;
}
int x, y;
main(int ac, char** av)
{
x = add(x, y);
return x;
}
-testcase #2---
static inline __attribute__((always_inline)) int add(int x, int y)
{
return x + y;
}
int x, y;
main(int ac, char** av)
{
x = add(x, y);
x = add(x, y);
return x;
}
-assembled code for testcase #1---
.global main
.type main, @function
main:
mov.l r14,@-r15
add #-16,r15
mov r15,r14
mov r14,r1
add #-48,r1
mov.l r4,@(52,r1)
...
-assembled code for testcase #2---
.global main
.type main, @function
main:
mov.l r14,@-r15
add #-24,r15 <-- stack grows down
mov r15,r14
mov r14,r1
add #-40,r1
mov.l r4,@(44,r1)
...
--
Summary: [SH] static inline attribute causes stack overflow
Product: gcc
Version: 4.2.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: saito at densan dot co dot jp
GCC build triplet: sh4-linux
GCC host triplet: sh4-linux
GCC target triplet: sh4-linux
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32577
[Bug target/31022] New: [SH4] internal compiler error with inline
When I compile a testcase that includes float and inline, it asserts internal
compiler error if it is optimized with -O2. If it is optimized with -O, this
error is not asserted.
% gcc -c test.c -O2 -m4
test.c: In function 'fff':
test.c:19: internal compiler error: in reg_overlap_mentioned_p, at
rtlanal.c:1382
Please submit a full bug report,
with preprocessed source if appropriate.
See http://gcc.gnu.org/bugs.html> for instructions.
-test case
typedef union { float f; int i; } fi_type;
static __inline__ float LOG2(float val)
{
fi_type num;
int i;
num.f = val;
i = num.i >> 23;
num.i++;
return num.f + i;
}
float f;
float fff()
{
return LOG2(f);
}
--
Summary: [SH4] internal compiler error with inline
Product: gcc
Version: 4.1.2
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: saito at densan dot co dot jp
GCC build triplet: sh4-linux
GCC host triplet: sh4-linux
GCC target triplet: sh4-linux
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31022
[Bug target/28055] New: [SH] a function attributed to weak is discarded, if '-O' is specified
I tested using gcc-3.4.6 and gcc-4.1.1. The code generated by gcc-4.1.1 is
wrong, because a function attributed to weak is discarded by optimization.
I wander if this problem is gcc4's specification because I can see same problem
with gcc-4.1.1 for PowerPC.
The source code is as follows. The command line is "sh4-linux-gcc -S testcase.c
-O".
void __attribute__ ((weak))
arch_init_clk_ops()
{
}
void clk_init(void)
{
arch_init_clk_ops();
}
The result for gcc-4.1.1:
.weak arch_init_clk_ops
.type arch_init_clk_ops, @function
arch_init_clk_ops:
mov.l r14,@-r15
mov r15,r14
mov r14,r15
mov.l @r15+,r14
rts
nop
.size arch_init_clk_ops, .-arch_init_clk_ops
.align 1
.global clk_init
.type clk_init, @function
clk_init:
mov.l r14,@-r15
mov r15,r14# arch_init_clk_op can not be called
mov r14,r15
mov.l @r15+,r14
rts
nop
The result for gcc-3.4.6:
arch_init_clk_ops:
mov.l r14,@-r15
sts.l pr,@-r15
mov r15,r14
mov r14,r15
lds.l @r15+,pr
rts
mov.l @r15+,r14
.size arch_init_clk_ops, .-arch_init_clk_ops
.align 1
.global clk_init
.type clk_init, @function
clk_init:
mov.l r14,@-r15
sts.l pr,@-r15
mov.l .L3,r1
jsr @r1
mov r15,r14
mov r14,r15
lds.l @r15+,pr
rts
mov.l @r15+,r14
.L4:
.align 2
.L3:
.long arch_init_clk_ops
--
Summary: [SH] a function attributed to weak is discarded, if '-O'
is specified
Product: gcc
Version: 4.1.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: saito at densan dot co dot jp
GCC build triplet: sh4-linux
GCC host triplet: sh4-linux
GCC target triplet: sh4-linux
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28055
[Bug target/28011] [SH] g++ generates wrong code, if '-fno-exceptions' and '-O' options are specified
--- Comment #1 from saito at densan dot co dot jp 2006-06-13 10:43 --- Created an attachment (id=11660) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11660&action=view) testcase This testcase is very large because this file already includes header files and this came from firefox's build. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28011
[Bug target/28011] New: [SH] g++ generates wrong code, if '-fno-exceptions' and '-O' options are specified
g++ generates wrong code, if '-fno-exceptions' and '-O' options are specified.
I did following tests. In the case of 1, wrong code was generated.
1. sh4-linux-g++ -S -O -fno-exceptions testcase.cc
2. sh4-linux-g++ -S -O2 -fno-exceptions testcase.cc
Source code is as follows but I will attach the testcase.
PRBool dummy;
asm("POST:");
return target->DispatchEvent(event, aDefaultAction ? aDefaultAction :
&dummy);
The result of case 1.
#APP
POST:
#NO_APP
mov.w .L76,r0
mov.w .L79,r3
add r14,r3
mov.l @(r0,r3),r4
mov.l @r4,r1
mov.l @(20,r1),r2
mov.l @r8,r5
mov.l @(44,r3),r1
tst r1,r1
bf .L27
mov.w .L66,r1# .L66: .short -188
mov.w .L79,r3# .L79: .short 188
add r14,r3
add r1,r3
mov.l r3,@(44,r3) # this code is wrong because 44 is not right value
.L27:# the value should be 232 - (188 - 188) but not 232 - 188
mov.w .L68,r0# .L68: .short 232
mov.l @(r0,r14),r6
jsr @r2
nop
The result of case 2. It seems that this case is right.
#APP
POST:
#NO_APP
mov.w .L62,r3
mov.l @r13,r5
add r14,r3
mov.l @(0,r3),r4
mov.l @(44,r3),r2
mov.l @r4,r1
tst r2,r2
mov.l @(20,r1),r1
bt .L34
.L27:
mov.w .L58,r0
mov.l @(r0,r14),r6
jsr @r1
nop
mov r0,r9
.L24:
mov.l .L63,r1
jsr @r1
mov r8,r4
...
.L34:
mov.w .L67,r2
mov.w .L68,r0
add r14,r2
mov.l r2,@(r0,r14)
bra .L27
nop
--
Summary: [SH] g++ generates wrong code, if '-fno-exceptions' and
'-O' options are specified
Product: gcc
Version: 4.1.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: saito at densan dot co dot jp
GCC build triplet: sh4-linux
GCC host triplet: sh4-linux
GCC target triplet: sh4-linux
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28011
[Bug target/27621] [SH] generated pic code is not right
--- Comment #4 from saito at densan dot co dot jp 2006-05-16 09:13 --- I can not look the problem of the testcase using gcc-4.1.0. I will try to use gcc-4.1.0 on shlinux instead of gcc-3.4.6. Thanks for your support. -- saito at densan dot co dot jp changed: What|Removed |Added Status|UNCONFIRMED |RESOLVED Resolution||FIXED http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27621
[Bug c++/27621] [SH] generated pic code is not right
--- Comment #3 from saito at densan dot co dot jp 2006-05-16 02:53 --- I could not build glibc for sh using 4.0.1 before, then I use 3.4.6 now, but I will test and report the result using gcc-4.1.0. -- saito at densan dot co dot jp changed: What|Removed |Added Component|target |c++ http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27621
[Bug c++/27621] [SH] generated pic code is not right
--- Comment #1 from saito at densan dot co dot jp 2006-05-16 02:28 --- Created an attachment (id=11472) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11472&action=view) testcase -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27621
[Bug c++/27621] New: [SH] generated pic code is not right
This problem depends on generated pic code for sh. For reproducing, needs two files a shared library(musician.so) and a executable file(main). I will attach these files. musician.so is shared library compiled with -fPIC option. musician.so is linked dynamically if main is excuted as follows. g++ -g -o musician.so musician.cc -shared -fPIC g++ -g -o main main.cc musician.so env LD_LIBRARY_PATH=.:$LD_LIBRARY_PATH main actual result: main issues a following error. Segmentation fault expected result: main prints a following message. playing=1 The following code generated by g++ is wrong for the pic code because r12 is not set for this shared library. .weak _ZThn4_N6typist5printEv .type _ZThn4_N6typist5printEv, @function _ZThn4_N6typist5printEv: .LFB46: mov.l .L153,r3 add r12,r3 jmp @r3 add #-4,r4 .L154: .align 2 .L153: .long [EMAIL PROTECTED] -- Summary: [SH] generated pic code is not right Product: gcc Version: 3.4.6 Status: UNCONFIRMED Severity: normal Priority: P3 Component: c++ AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: saito at densan dot co dot jp GCC build triplet: sh4-linux GCC host triplet: sh4-linux GCC target triplet: sh4-linux http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27621
