Re: [PATCH v10 3/8] selftest/bpf: Fixing powerpc JIT disassembly failure

Hari Bathini <[email protected]>
Newsgroups org.ozlabs.lists.linuxppc-dev,org.kernel.vger.bpf,org.kernel.vger.linux-kselftest,org.kernel.vger.stable
Message-ID <[email protected]>

On 06/08/26 10:49 pm, Hari Bathini wrote:
> 
> 
> On 05/08/26 11:36 am, Saket Kumar Bhaskar wrote:
>> From: Abhishek Dubey <[email protected]>
>>
>> Ensure that the trampoline stubs JITed at the tail of the
>> epilogue do not expose the dummy trampoline address stored
>> in the last 8 bytes(64-bit) and last 4 bytes(32-bit)
>> to the disassembly flow. Prevent the disassembler from
>> ingesting this memory address, as it may occasionally decode
>> into a seemingly valid but incorrect instruction. Fix this
>> issue by truncating the last 8/4 bytes from JITed buffers
>> before supplying them for disassembly.
> 
>>
>> Fixes: b991fc520700 ("selftests/bpf: utility function to get program 
>> disassembly after jit")
> 
> The fixes tag on this patch doesn't really help. Please drop it.

With that..

Reviewed-by: Hari Bathini <[email protected]>

> 
>> Signed-off-by: Abhishek Dubey <[email protected]>
>> Signed-off-by: Saket Kumar Bhaskar <[email protected]>
>> Tested-by: Yeswanth Krishna Tellakula <[email protected]>
>> ---
>>   .../selftests/bpf/jit_disasm_helpers.c        | 25 ++++++++++++++++++-
>>   1 file changed, 24 insertions(+), 1 deletion(-)
>>
>> diff --git a/tools/testing/selftests/bpf/jit_disasm_helpers.c b/tools/ 
>> testing/selftests/bpf/jit_disasm_helpers.c
>> index 3558fe10e28c..c4aa1d69d3bb 100644
>> --- a/tools/testing/selftests/bpf/jit_disasm_helpers.c
>> +++ b/tools/testing/selftests/bpf/jit_disasm_helpers.c
>> @@ -179,9 +179,11 @@ int get_jited_program_text(int fd, char *text, 
>> size_t text_sz)
>>       struct bpf_prog_info info = {};
>>       __u32 info_len = sizeof(info);
>>       __u32 jited_funcs, len, pc;
>> +    __u32 trunc_len = 0, disasm_len;
>>       __u32 *func_lens = NULL;
>>       FILE *text_out = NULL;
>>       uint8_t *image = NULL;
>> +    char *triple = NULL;
>>       int i, err = 0;
>>       if (!llvm_initialized) {
>> @@ -225,9 +227,30 @@ int get_jited_program_text(int fd, char *text, 
>> size_t text_sz)
>>       if (!ASSERT_OK(err, "bpf_prog_get_info_by_fd #2"))
>>           goto out;
>> +    /*
>> +     * last 8 bytes contains dummy_trampoline address in JIT
>> +     * output on 64-bit and last 4 bytes on 32-bit powerpc,
>> +     * which can't disassemble to a valid instruction.
>> +     */
>> +    triple = LLVMGetDefaultTargetTriple();
>> +    if (triple) {
>> +        if (strstr(triple, "powerpc64") || strstr(triple, "ppc64"))
>> +            trunc_len = 8;
>> +        else if (strstr(triple, "powerpc") || strstr(triple, "ppc"))
>> +            trunc_len = 4;
>> +        LLVMDisposeMessage(triple);
>> +    }
>> +
>>       for (pc = 0, i = 0; i < jited_funcs; ++i) {
>> +
>>           fprintf(text_out, "func #%d:\n", i);
>> -        disasm_one_func(text_out, image + pc, func_lens[i]);
>> +        /*
>> +         * Disabled JIT have zero func_lens, hence underflow
>> +         */
>> +        disasm_len = func_lens[i] > trunc_len ?
>> +                    func_lens[i] - trunc_len : 0;
>> +        disasm_one_func(text_out, image + pc, disasm_len);
>> +
>>           fprintf(text_out, "\n");
>>           pc += func_lens[i];
>>       }
> 
> - Hari
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.