[PATCH] [gdb/testsuite] Fix gdb.arch/i386-signal.exp on x86_64

[PATCH] [gdb/testsuite] Fix gdb.arch/i386-signal.exp on x86_64

[Tom de Vries]

On x86_64-linux, with test-case gdb.arch/i386-signal.exp I run into:
...
builtin_spawn -ignore SIGHUP gcc -fno-stack-protector i386-signal.c \
  -fdiagnostics-color=never -fno-pie -g -no-pie -lm -o i386-signal^M
/tmp/cc2xydTG.s: Assembler messages:^M
/tmp/cc2xydTG.s:50: Error: operand size mismatch for `push'^M
compiler exited with status 1
output is:
/tmp/cc2xydTG.s: Assembler messages:^M
/tmp/cc2xydTG.s:50: Error: operand size mismatch for `push'^M

gdb compile failed, /tmp/cc2xydTG.s: Assembler messages:
/tmp/cc2xydTG.s:50: Error: operand size mismatch for `push'
UNTESTED: gdb.arch/i386-signal.exp: failed to compile
...

This is with gas 2.41, it compiles without problems with gas 2.40.  Some more
strict checking was added in commit 5cc007751cd ("x86: further adjust
extend-to-32bit-address conditions").

The offending bit is:
...
    "    push $sigframe\n"
...
which refers to a function:
...
    "    .globl sigframe\n"
    "sigframe:\n"
...

The test-case passes with target board unix/-m32.

Make the test-case work by using pushq instead of push for the
is_amd64_regs_target case.

Tested on x86_64-linux, with target boards:
- unix/-m64 (is_amd64_regs_target == 1), and
- unix/-m32 (is_amd64_regs_target == 0),

PR testsuite/30928
Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30928
---
 gdb/testsuite/gdb.arch/i386-signal.c   | 4 ++++
 gdb/testsuite/gdb.arch/i386-signal.exp | 7 ++++++-
 2 files changed, 10 insertions(+), 1 deletion(-)

diff --git a/gdb/testsuite/gdb.arch/i386-signal.c b/gdb/testsuite/gdb.arch/i386-signal.c
index 19bb1bbaaf8..4bf97e5f159 100644
--- a/gdb/testsuite/gdb.arch/i386-signal.c
+++ b/gdb/testsuite/gdb.arch/i386-signal.c
@@ -45,7 +45,11 @@ asm(".text\n"
     "    .align 8\n"
     "    .globl setup\n"
     "setup:\n"
+#if IS_AMD64_REGS_TARGET
+    "    pushq $sigframe\n"
+#else
     "    push $sigframe\n"
+#endif
     "    jmp func\n"
     "\n"
     "    .cfi_startproc\n"
diff --git a/gdb/testsuite/gdb.arch/i386-signal.exp b/gdb/testsuite/gdb.arch/i386-signal.exp
index f6a88719a40..a69172ed5cb 100644
--- a/gdb/testsuite/gdb.arch/i386-signal.exp
+++ b/gdb/testsuite/gdb.arch/i386-signal.exp
@@ -19,8 +19,13 @@ require {is_any_target "i?86-*-*" "x86_64-*-*"}
 
 standard_testfile
 
+set opts {}
+lappend opts debug
+lappend opts nopie
+lappend opts additional_flags=-DIS_AMD64_REGS_TARGET=[is_amd64_regs_target]
+
 if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" \
-	  executable { debug nopie }] != "" } {
+	  executable $opts] != "" } {
     untested "failed to compile"
     return -1
 }

base-commit: 23b5268986d9b68b965939416a7aa96e8698e403
-- 
2.35.3

Re: [PATCH] [gdb/testsuite] Fix gdb.arch/i386-signal.exp on x86_64

[John Baldwin]

On 10/1/23 7:14 AM, Tom de Vries via Gdb-patches wrote:
> On x86_64-linux, with test-case gdb.arch/i386-signal.exp I run into:
> ...
> builtin_spawn -ignore SIGHUP gcc -fno-stack-protector i386-signal.c \
>    -fdiagnostics-color=never -fno-pie -g -no-pie -lm -o i386-signal^M
> /tmp/cc2xydTG.s: Assembler messages:^M
> /tmp/cc2xydTG.s:50: Error: operand size mismatch for `push'^M
> compiler exited with status 1
> output is:
> /tmp/cc2xydTG.s: Assembler messages:^M
> /tmp/cc2xydTG.s:50: Error: operand size mismatch for `push'^M
> 
> gdb compile failed, /tmp/cc2xydTG.s: Assembler messages:
> /tmp/cc2xydTG.s:50: Error: operand size mismatch for `push'
> UNTESTED: gdb.arch/i386-signal.exp: failed to compile
> ...
> 
> This is with gas 2.41, it compiles without problems with gas 2.40.  Some more
> strict checking was added in commit 5cc007751cd ("x86: further adjust
> extend-to-32bit-address conditions").
> 
> The offending bit is:
> ...
>      "    push $sigframe\n"
> ...
> which refers to a function:
> ...
>      "    .globl sigframe\n"
>      "sigframe:\n"
> ...
> 
> The test-case passes with target board unix/-m32.
> 
> Make the test-case work by using pushq instead of push for the
> is_amd64_regs_target case.
> 
> Tested on x86_64-linux, with target boards:
> - unix/-m64 (is_amd64_regs_target == 1), and
> - unix/-m32 (is_amd64_regs_target == 0),
> 
> PR testsuite/30928
> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30928
> ---
>   gdb/testsuite/gdb.arch/i386-signal.c   | 4 ++++
>   gdb/testsuite/gdb.arch/i386-signal.exp | 7 ++++++-
>   2 files changed, 10 insertions(+), 1 deletion(-)
> 
> diff --git a/gdb/testsuite/gdb.arch/i386-signal.c b/gdb/testsuite/gdb.arch/i386-signal.c
> index 19bb1bbaaf8..4bf97e5f159 100644
> --- a/gdb/testsuite/gdb.arch/i386-signal.c
> +++ b/gdb/testsuite/gdb.arch/i386-signal.c
> @@ -45,7 +45,11 @@ asm(".text\n"
>       "    .align 8\n"
>       "    .globl setup\n"
>       "setup:\n"
> +#if IS_AMD64_REGS_TARGET
> +    "    pushq $sigframe\n"
> +#else
>       "    push $sigframe\n"
> +#endif
>       "    jmp func\n"
>       "\n"
>       "    .cfi_startproc\n"
> diff --git a/gdb/testsuite/gdb.arch/i386-signal.exp b/gdb/testsuite/gdb.arch/i386-signal.exp
> index f6a88719a40..a69172ed5cb 100644
> --- a/gdb/testsuite/gdb.arch/i386-signal.exp
> +++ b/gdb/testsuite/gdb.arch/i386-signal.exp
> @@ -19,8 +19,13 @@ require {is_any_target "i?86-*-*" "x86_64-*-*"}
>   
>   standard_testfile
>   
> +set opts {}
> +lappend opts debug
> +lappend opts nopie
> +lappend opts additional_flags=-DIS_AMD64_REGS_TARGET=[is_amd64_regs_target]
> +
>   if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" \
> -	  executable { debug nopie }] != "" } {
> +	  executable $opts] != "" } {
>       untested "failed to compile"
>       return -1
>   }
> 
> base-commit: 23b5268986d9b68b965939416a7aa96e8698e403

LGRM

-- 
John Baldwin

Re: [PATCH] [gdb/testsuite] Fix gdb.arch/i386-signal.exp on x86_64

[Tom de Vries]

On 10/7/23 00:06, John Baldwin wrote:
> On 10/1/23 7:14 AM, Tom de Vries via Gdb-patches wrote:
>> On x86_64-linux, with test-case gdb.arch/i386-signal.exp I run into:
>> ...
>> builtin_spawn -ignore SIGHUP gcc -fno-stack-protector i386-signal.c \
>>    -fdiagnostics-color=never -fno-pie -g -no-pie -lm -o i386-signal^M
>> /tmp/cc2xydTG.s: Assembler messages:^M
>> /tmp/cc2xydTG.s:50: Error: operand size mismatch for `push'^M
>> compiler exited with status 1
>> output is:
>> /tmp/cc2xydTG.s: Assembler messages:^M
>> /tmp/cc2xydTG.s:50: Error: operand size mismatch for `push'^M
>>
>> gdb compile failed, /tmp/cc2xydTG.s: Assembler messages:
>> /tmp/cc2xydTG.s:50: Error: operand size mismatch for `push'
>> UNTESTED: gdb.arch/i386-signal.exp: failed to compile
>> ...
>>
>> This is with gas 2.41, it compiles without problems with gas 2.40.  
>> Some more
>> strict checking was added in commit 5cc007751cd ("x86: further adjust
>> extend-to-32bit-address conditions").
>>
>> The offending bit is:
>> ...
>>      "    push $sigframe\n"
>> ...
>> which refers to a function:
>> ...
>>      "    .globl sigframe\n"
>>      "sigframe:\n"
>> ...
>>
>> The test-case passes with target board unix/-m32.
>>
>> Make the test-case work by using pushq instead of push for the
>> is_amd64_regs_target case.
>>
>> Tested on x86_64-linux, with target boards:
>> - unix/-m64 (is_amd64_regs_target == 1), and
>> - unix/-m32 (is_amd64_regs_target == 0),
>>
>> PR testsuite/30928
>> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=30928
>> ---
>>   gdb/testsuite/gdb.arch/i386-signal.c   | 4 ++++
>>   gdb/testsuite/gdb.arch/i386-signal.exp | 7 ++++++-
>>   2 files changed, 10 insertions(+), 1 deletion(-)
>>
>> diff --git a/gdb/testsuite/gdb.arch/i386-signal.c 
>> b/gdb/testsuite/gdb.arch/i386-signal.c
>> index 19bb1bbaaf8..4bf97e5f159 100644
>> --- a/gdb/testsuite/gdb.arch/i386-signal.c
>> +++ b/gdb/testsuite/gdb.arch/i386-signal.c
>> @@ -45,7 +45,11 @@ asm(".text\n"
>>       "    .align 8\n"
>>       "    .globl setup\n"
>>       "setup:\n"
>> +#if IS_AMD64_REGS_TARGET
>> +    "    pushq $sigframe\n"
>> +#else
>>       "    push $sigframe\n"
>> +#endif
>>       "    jmp func\n"
>>       "\n"
>>       "    .cfi_startproc\n"
>> diff --git a/gdb/testsuite/gdb.arch/i386-signal.exp 
>> b/gdb/testsuite/gdb.arch/i386-signal.exp
>> index f6a88719a40..a69172ed5cb 100644
>> --- a/gdb/testsuite/gdb.arch/i386-signal.exp
>> +++ b/gdb/testsuite/gdb.arch/i386-signal.exp
>> @@ -19,8 +19,13 @@ require {is_any_target "i?86-*-*" "x86_64-*-*"}
>>   standard_testfile
>> +set opts {}
>> +lappend opts debug
>> +lappend opts nopie
>> +lappend opts 
>> additional_flags=-DIS_AMD64_REGS_TARGET=[is_amd64_regs_target]
>> +
>>   if { [gdb_compile "${srcdir}/${subdir}/${srcfile}" "${binfile}" \
>> -      executable { debug nopie }] != "" } {
>> +      executable $opts] != "" } {
>>       untested "failed to compile"
>>       return -1
>>   }
>>
>> base-commit: 23b5268986d9b68b965939416a7aa96e8698e403
> 
> LGRM
> 

Thanks for the review.

I've committed after adding a note about 
https://sourceware.org/pipermail/binutils/2023-October/129818.html to 
the commit log.

Thanks,
- Tom

Re: [PATCH] [gdb/testsuite] Fix gdb.arch/i386-signal.exp on x86_64

[Jan Beulich]

On 07.10.2023 10:37, Tom de Vries wrote:
> I've committed after adding a note about 
> https://sourceware.org/pipermail/binutils/2023-October/129818.html to 
> the commit log.

Perhaps the better course of action would have been to backport
fb1c10585ead to the respective gdb branch(es)?

Jan

Re: [PATCH] [gdb/testsuite] Fix gdb.arch/i386-signal.exp on x86_64

[Tom de Vries]

On 10/16/23 09:53, Jan Beulich wrote:
> On 07.10.2023 10:37, Tom de Vries wrote:
>> I've committed after adding a note about
>> https://sourceware.org/pipermail/binutils/2023-October/129818.html to
>> the commit log.
> 
> Perhaps the better course of action would have been to backport
> fb1c10585ead to the respective gdb branch(es)?

The committed patch makes the gdb test-case robust against a problem in 
a released gas version.  So this is the right fix.

Backporting the gas fix to a gdb release branch has no effect for the 
case that gas is build from binutils releases and branches.

Thanks,
- Tom