Re: [linux-next:master] [bpf/tests] daabb2b098: kernel-selftests.net.test_bpf.sh.fail

[Puranjay Mohan <puranjay12@gmail.com>]

Hi Yonghong,

On Wed, Oct 11, 2023 at 4:01 PM Puranjay Mohan <puranjay12@gmail.com> wrote:
>
> On Wed, Oct 11, 2023 at 3:20 PM kernel test robot <oliver.sang@intel.com> wrote:
> >
> >
> > hi, Puranjay Mohan,
> >
> > we reported same issue when this commit is a review patch as:
> > https://lore.kernel.org/all/202309261451.8934f9ad-oliver.sang@intel.com/
> >
> > now we noticed this commit is in linux-next/master and we still observed
> > failure.
> >
> > is there any requirements to run new tests? Thanks
> >
> >
> > Hello,
> >
> > kernel test robot noticed "kernel-selftests.net.test_bpf.sh.fail" on:
> >
> > commit: daabb2b098e04753fa3d1b1feed13e5a61bef61c ("bpf/tests: add tests for cpuv4 instructions")
> > https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
> >
> > [test failed on linux-next/master 719136e5c24768ebdf80b9daa53facebbdd377c3]
> >
> > in testcase: kernel-selftests
> > version: kernel-selftests-x86_64-60acb023-1_20230329
> > with following parameters:
> >
> >         group: net
> >
> >
> >
> > compiler: gcc-12
> > test machine: 36 threads 1 sockets Intel(R) Core(TM) i9-10980XE CPU @ 3.00GHz (Cascade Lake) with 32G memory
> >
> > (please refer to attached dmesg/kmsg for entire log/backtrace)
> >
> >
> >
> > If you fix the issue in a separate patch/commit (i.e. not just a new version of
> > the same patch/commit), kindly add following tags
> > | Reported-by: kernel test robot <oliver.sang@intel.com>
> > | Closes: https://lore.kernel.org/oe-lkp/202310111838.46ff5b6a-oliver.sang@intel.com
> >
> >
> >
> > ....
> > # timeout set to 1500
> > # selftests: net: test_bpf.sh
> > # test_bpf: [FAIL]
> > not ok 13 selftests: net: test_bpf.sh # exit=1
> > ....
>
> The test "ALU_MOVSX | BPF_W" is failing for x86_64. This is a new test
> added by my patch.
> This is how the test looks:

As you have worked more on cpu=v4 instructions.
Is it a requirement that MOVSX32 can only have offset = 8/16 and never 32.
I know that MOVSX32 with offset=32 is equivalent to normal MOV32.

If this is a requirement then the below test is running an invalid
instruction and we can remove it from test_bpf.

> {
>                 "ALU_MOVSX | BPF_W",
>                 .u.insns_int = {
>                         BPF_LD_IMM64(R2, 0x00000000deadbeefLL),
>                         BPF_LD_IMM64(R3, 0xdeadbeefdeadbeefLL),
>                         BPF_MOVSX32_REG(R1, R3, 32),
>                         BPF_JMP_REG(BPF_JEQ, R2, R1, 2),
>                         BPF_MOV32_IMM(R0, 2),
>                         BPF_EXIT_INSN(),
>                         BPF_MOV32_IMM(R0, 1),
>                         BPF_EXIT_INSN(),
>                 },
>                 INTERNAL,
>                 { },
>                 { { 0, 0x1 } },
>  }
>
> I am creating a build for testing on my end.
>
> >
> >
> >
> > The kernel config and materials to reproduce are available at:
> > https://download.01.org/0day-ci/archive/20231011/202310111838.46ff5b6a-oliver.sang@intel.com
> >
> >
> >
> > --
> > 0-DAY CI Kernel Test Service
> > https://github.com/intel/lkp-tests/wiki
> >
>
> Thanks,
> Puranjay