Re: [PATCH bpf-next 0/5] Support local vmtest for riscv64

From: Pu Lehui <pulehui@huaweicloud.com>
To: Eduard Zingerman <eddyz87@gmail.com>,
	Daniel Borkmann <daniel@iogearbox.net>,
	bpf@vger.kernel.org, linux-riscv@lists.infradead.org,
	netdev@vger.kernel.org
Cc: "Björn Töpel" <bjorn@kernel.org>,
	"Alexei Starovoitov" <ast@kernel.org>,
	"Andrii Nakryiko" <andrii@kernel.org>,
	"Martin KaFai Lau" <martin.lau@linux.dev>,
	"Song Liu" <song@kernel.org>, "Yonghong Song" <yhs@fb.com>,
	"John Fastabend" <john.fastabend@gmail.com>,
	"KP Singh" <kpsingh@kernel.org>,
	"Stanislav Fomichev" <sdf@google.com>,
	"Hao Luo" <haoluo@google.com>, "Jiri Olsa" <jolsa@kernel.org>,
	"Mykola Lysenko" <mykolal@fb.com>,
	"Manu Bretelle" <chantr4@gmail.com>,
	"Pu Lehui" <pulehui@huawei.com>
Subject: Re: [PATCH bpf-next 0/5] Support local vmtest for riscv64
Date: Wed, 3 Apr 2024 18:31:59 +0800	[thread overview]
Message-ID: <6feb3403-4214-4143-b0bc-c95daf8eea2b@huaweicloud.com> (raw)
In-Reply-To: <f20d1e2a2f5fa10f29bf1fddbaf99c3f185e8530.camel@gmail.com>

On 2024/4/3 7:40, Eduard Zingerman wrote:
> On Sat, 2024-03-30 at 18:12 +0800, Pu Lehui wrote:
> [...]
> 
>>> Looks like I won't be able to test this patch-set, unless you have
>>> some writeup on how to create a riscv64 dev environment at hand.
>>> Sorry for the noise
>>
>> Yeah, environmental issues are indeed a developer's nightmare. I will
>> try to do something for the newcomers of riscv64 bpf. At present, I have
>> simply built a docker local vmtest environment [0] based on Bjorn's
>> riscv-cross-builder. We can directly run vmtest within this environment.
>> Hopefully it will help.
>>
>> Link: https://github.com/pulehui/riscv-cross-builder/tree/vmtest [0]
> 
> Hi Pu,
> 
> Thank you for sharing the docker file, I've managed to run the tests
> using it. In order to avoid creating files with root permissions I had
> to add the following lines at the end of the Dockerfile:
> 
> + RUN useradd --no-create-home --uid 1000 eddy
> + RUN passwd -d eddy
> + RUN echo 'eddy ALL=(ALL) NOPASSWD:ALL' >> /etc/sudoers
> + # vmtest.sh does 'mount -o loop',
> + # ensure there is a loop device in the container
> + RUN mknod /dev/loop0 b 7 20
> 
> Where 'eddy' is my local user with UID 1000.
> Probably this should be made more generic.
> I used the following command to start the container:
> 
> docker run -ti -u 1000:1000 \
>      --rm -v <path-to-kernel-dir>:/workspace \
>      -v <path-to-rootfs-image-dir>:/rootfs \
>      --privileged ubuntu-vmtest:latest /bin/bash
> 
> Also, I had to add '-d /rootfs/bpf_selftests' option for vmtest.sh in
> order to avoid polluting user directory inside the container.
> Maybe OUTPUT_DIR for vmtest.sh should be mounted as a separate volume.
> 
> I agree with Daniel, it would be great to document all of this

Forgot to reply to this in my last email. It my pleasure to do this.

> somewhere in the repo (or even scripted somehow).
> 
> Using the specified DENYLIST I get the following stats for test_progs:
> 
>    #3/2     arena_htab/arena_htab_asm:FAIL
>    #3       arena_htab:FAIL

Puranjay has submitted to riscv bpf arena and will be merged soon. So I 
didn't add it to DENYLIST.riscv64.

https://lore.kernel.org/bpf/20240326224943.86912-1-puranjay12@gmail.com/

>    #95      get_branch_snapshot:FAIL
>    #172/1   perf_branches/perf_branches_hw:FAIL
>    #172     perf_branches:FAIL

riscv sbi pmu driver not support branch sampling yet.  The following 
patch should be used for better regression.

https://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux.git/commit/?id=ea6873118493

>    #434/3   verifier_arena/basic_alloc3:FAIL
>    #434     verifier_arena:FAIL
>    Summary: 531/3581 PASSED, 64 SKIPPED, 4 FAILED
> 
> Tested-by: Eduard Zingerman <eddyz87@gmail.com>
> 
>> PS: Since the current rootfs of riscv64 is not in the INDEX, I simply
>> modified vmtest.sh to support local rootfs.
> 
> Could you please add this change to the patch-set?

yep, will try to make it more convenient.

> 
> [...]