linux-kernel - Re: [PATCH] libbpf: Support raw btf placed in the default path

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]

Date:   Mon, 29 Aug 2022 13:48:05 +0800
From:   Tao Chen <chentao.kernel@...ux.alibaba.com>
To:     Andrii Nakryiko <andrii.nakryiko@...il.com>
Cc:     Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Andrii Nakryiko <andrii@...nel.org>,
        Martin KaFai Lau <kafai@...com>,
        Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>,
        John Fastabend <john.fastabend@...il.com>,
        KP Singh <kpsingh@...nel.org>, netdev@...r.kernel.org,
        bpf@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] libbpf: Support raw btf placed in the default path

Hi Nakryiko, thank you for your reply. Yes, i happed to put the raw
BTF made by myself in /boot on kernel4.19, unluckly it reported error.
So is it possible to allow the raw BTF in /boot directly, athough we
can set the specified btf location again to solve the problem with the
bpf_object_open_opts interface.  

As you say, maybe we can remove the locations[i].raw_btf check, just
use the btf__parse. It looks more concise.


在 2022/8/26 上午4:26, Andrii Nakryiko 写道:
> On Wed, Aug 24, 2022 at 7:05 AM chentao.ct
> <chentao.kernel@...ux.alibaba.com> wrote:
>>
>> Now only elf btf can be placed in the default path, raw btf should
>> also can be there.
>>
> 
> It's not clear what you are trying to achieve. Do you want libbpf to
> attempt to load /boot/vmlinux-%1$s as raw BTF as well (so you can sort
> of sneak in pregenerated BTF), or what exactly?
> btf__load_vmlinux_btf() code already supports loading raw BTF, it just
> needs to be explicitly specified in locations table.
> 
> So with your change locations[i].raw_btf check doesn't make sense and
> we need to clean this up.
> 
> But first, let's discuss the use case, instead of your specific solution.
> 
> 
>> Signed-off-by: chentao.ct <chentao.kernel@...ux.alibaba.com>
>> ---
>>   tools/lib/bpf/btf.c | 4 ++--
>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/tools/lib/bpf/btf.c b/tools/lib/bpf/btf.c
>> index bb1e06e..b22b5b3 100644
>> --- a/tools/lib/bpf/btf.c
>> +++ b/tools/lib/bpf/btf.c
>> @@ -4661,7 +4661,7 @@ struct btf *btf__load_vmlinux_btf(void)
>>          } locations[] = {
>>                  /* try canonical vmlinux BTF through sysfs first */
>>                  { "/sys/kernel/btf/vmlinux", true /* raw BTF */ },
>> -               /* fall back to trying to find vmlinux ELF on disk otherwise */
>> +               /* fall back to trying to find vmlinux RAW/ELF on disk otherwise */
>>                  { "/boot/vmlinux-%1$s" },
>>                  { "/lib/modules/%1$s/vmlinux-%1$s" },
>>                  { "/lib/modules/%1$s/build/vmlinux" },
>> @@ -4686,7 +4686,7 @@ struct btf *btf__load_vmlinux_btf(void)
>>                  if (locations[i].raw_btf)
>>                          btf = btf__parse_raw(path);
>>                  else
>> -                       btf = btf__parse_elf(path, NULL);
>> +                       btf = btf__parse(path, NULL);
>>                  err = libbpf_get_error(btf);
>>                  pr_debug("loading kernel BTF '%s': %d\n", path, err);
>>                  if (err)
>> --
>> 2.2.1
>>