Re: [PATCH 1/2] perf alsr: Add missing SIMD handling in aslr_tool__process_sample()

["Mi, Dapeng" <dapeng1.mi@linux.intel.com>]

On 9/22/2026 12:21 AM, Ian Rogers wrote:
> On Sun, Sep 20, 2026 at 11:42 PM Mi, Dapeng <dapeng1.mi@linux.intel.com> wrote:
>> Kindly ping ...
>>
>> @Arnaldo, @Namhyung, @Ian, not sure if you have bandwidth to review this
>> small patchset which fixes the Sashiko found minor issues for the "Perf
>> tools: Support eGPRs/SSP/SIMD registers sampling" patchset?
>>
>> If possible, it's better to be queued in this cycle, then we don't need
>> extra backporting. Thanks.
> Reviewed-by: Ian Rogers <irogers@google.com>
>
>> On 9/15/2026 5:11 PM, Dapeng Mi wrote:
>>> Currently the SIMD register sampling is supported, but the corresponding
>>> handling is missed in aslr_tool__process_sample(). This can break parsing
>>> of the sample record.
>>>
>>> Add the missing SIMD register handling in aslr_tool__process_sample().
>>>
>>> Signed-off-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
>>> ---
>>>
>>> These two patches fix Sashiko reported defects for the patchset:
>>> "Perf tools: Support eGPRs/SSP/SIMD registers sampling"
>>> https://lore.kernel.org/all/20260914014049.2061519-1-dapeng1.mi@linux.intel.com/
>>>
>>>  tools/perf/util/aslr.c | 31 +++++++++++++++++++++++++++++++
>>>  1 file changed, 31 insertions(+)
>>>
>>> diff --git a/tools/perf/util/aslr.c b/tools/perf/util/aslr.c
>>> index 027695d96779..a097084201a0 100644
>>> --- a/tools/perf/util/aslr.c
>>> +++ b/tools/perf/util/aslr.c
>>> @@ -19,6 +19,7 @@
>>>  #include <internal/lib.h>  /* page_size */
>>>  #include <linux/compiler.h>
>>>  #include <linux/zalloc.h>
>>> +#include <linux/overflow.h>
>>>  #include <errno.h>
>>>  #include <inttypes.h>
>>>  #include <unistd.h>
>>> @@ -992,6 +993,34 @@ static int aslr_tool__process_sample(const struct perf_tool *tool,
>>>                               COPY_U64();
>>>               }
>>>       }
>>> +
>>> +#define CHECK_SIMD() \
>>> +     do { \
>>> +             if (abi & PERF_SAMPLE_REGS_ABI_SIMD) { \
>>> +                     u64 nr_vector, vec_qwords; \
>>> +                     u64 nr_pred, pred_qwords; \
>>> +                     u64 header_len, vec_len, pred_len, simd_len; \
>>> +                     if (CHECK_BOUNDS(nr + 4, nr + 4)) { \
>>> +                             ret = -EFAULT; \
>>> +                             goto out_put; \
>>> +                     } \
>>> +                     header_len = 4; \
>>> +                     nr_vector = in_array[i + nr]; \
>>> +                     vec_qwords = in_array[i + nr + 1]; \
>>> +                     nr_pred = in_array[i + nr + 2]; \
>>> +                     pred_qwords = in_array[i + nr + 3]; \
>>> +                     if (check_mul_overflow(nr_vector, vec_qwords, &vec_len) || \
>>> +                         check_mul_overflow(nr_pred, pred_qwords, &pred_len) || \
>>> +                         check_add_overflow(vec_len, pred_len, &simd_len) || \
>>> +                         check_add_overflow(header_len, simd_len, &simd_len) || \
>>> +                         check_add_overflow(nr, simd_len, &nr) || \
>>> +                         nr > max_i - i) { \
>>> +                             ret = -EFAULT; \
>>> +                             goto out_put; \
>>> +                     } \
>>> +             } \
>>> +     } while (0)
>>> +
>>>       if (orig_sample_type & PERF_SAMPLE_REGS_USER) {
>>>               u64 abi;
>>>
>>> @@ -1007,6 +1036,7 @@ static int aslr_tool__process_sample(const struct perf_tool *tool,
>>>                               ret = -EFAULT;
>>>                               goto out_put;
>>>                       }
>>> +                     CHECK_SIMD();
>>>                       i += nr;
>>>               }
>>>       }
>>> @@ -1061,6 +1091,7 @@ static int aslr_tool__process_sample(const struct perf_tool *tool,
>>>                               ret = -EFAULT;
>>>                               goto out_put;
>>>                       }
>>> +                     CHECK_SIMD();
> nit: at the end of the function you could #undef CHECK_SIMD to make it
> more explicit that the macro only makes sense within the function.

Sure. Thanks.


>
> Thanks,
> Ian
>
>>>                       i += nr;
>>>               }
>>>       }
>>>
>>> base-commit: 91b0782fc9e9d2f0a40b5256146e014802fdbb36