linux-kernel - Re: [PATCH] selftests: breakpoints: Fix wrong argument of ptrace() when single step

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <d560a99f-be45-c03f-e729-56ccdbc9a131@loongson.cn>
Date:   Thu, 4 Feb 2021 10:36:38 +0800
From:   Tiezhu Yang <yangtiezhu@...ngson.cn>
To:     Shuah Khan <skhan@...uxfoundation.org>,
        Shuah Khan <shuah@...nel.org>
Cc:     linux-kselftest@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] selftests: breakpoints: Fix wrong argument of ptrace()
 when single step

On 02/04/2021 02:30 AM, Shuah Khan wrote:
> On 2/3/21 1:39 AM, Tiezhu Yang wrote:
>> According to the error message, the first argument of ptrace() should be
>> PTRACE_SINGLESTEP instead of PTRACE_CONT when ptrace single step.
>>
>> Fixes: f43365ee17f8 ("selftests: arm64: add test for 
>> unaligned/inexact watchpoint handling")
>> Signed-off-by: Tiezhu Yang <yangtiezhu@...ngson.cn>
>> ---
>>   tools/testing/selftests/breakpoints/breakpoint_test_arm64.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git 
>> a/tools/testing/selftests/breakpoints/breakpoint_test_arm64.c 
>> b/tools/testing/selftests/breakpoints/breakpoint_test_arm64.c
>> index ad41ea6..2f4d4d6 100644
>> --- a/tools/testing/selftests/breakpoints/breakpoint_test_arm64.c
>> +++ b/tools/testing/selftests/breakpoints/breakpoint_test_arm64.c
>> @@ -143,7 +143,7 @@ static bool run_test(int wr_size, int wp_size, 
>> int wr, int wp)
>>       if (!set_watchpoint(pid, wp_size, wp))
>>           return false;
>>   -    if (ptrace(PTRACE_CONT, pid, NULL, NULL) < 0) {
>> +    if (ptrace(PTRACE_SINGLESTEP, pid, NULL, NULL) < 0) {
>>           ksft_print_msg(
>>               "ptrace(PTRACE_SINGLESTEP) failed: %s\n",
>>               strerror(errno));
>>
>
> Right before this it does a set_watchpoint(). PTRACE_CONT is what
> makes sense to me. Error might be the one that is incorrect here?

What do you think the following change? If it is OK, I will send v2 soon.

diff --git a/tools/testing/selftests/breakpoints/breakpoint_test_arm64.c 
b/tools/testing/selftests/breakpoints/breakpoint_test_arm64.c
index ad41ea6..e704181 100644
--- a/tools/testing/selftests/breakpoints/breakpoint_test_arm64.c
+++ b/tools/testing/selftests/breakpoints/breakpoint_test_arm64.c
@@ -145,7 +145,7 @@ static bool run_test(int wr_size, int wp_size, int 
wr, int wp)

         if (ptrace(PTRACE_CONT, pid, NULL, NULL) < 0) {
                 ksft_print_msg(
-                       "ptrace(PTRACE_SINGLESTEP) failed: %s\n",
+                       "ptrace(PTRACE_CONT) failed: %s\n",
                         strerror(errno));
                 return false;
         }
@@ -159,7 +159,7 @@ static bool run_test(int wr_size, int wp_size, int 
wr, int wp)
         }
         alarm(0);
         if (WIFEXITED(status)) {
-               ksft_print_msg("child did not single-step\n");
+               ksft_print_msg("child exited prematurely\n");
                 return false;
         }
         if (!WIFSTOPPED(status)) {

Thanks,
Tiezhu

>
> thanks,
> -- Shuah