On Thu, Dec 21, 2023 at 04:52:54PM -0800, Alexei Starovoitov wrote:
> On Thu, Dec 21, 2023 at 10:18 AM Daniel Xu <dxu@xxxxxxxxx> wrote:
> >
> > Hi Alexei,
> >
> > On Thu, Dec 21, 2023 at 10:07:33AM -0800, Alexei Starovoitov wrote:
> > > On Thu, Dec 21, 2023 at 9:43 AM Alan Maguire <alan.maguire@xxxxxxxxxx> wrote:
> > > >
> > > > On 21/12/2023 17:05, Alexei Starovoitov wrote:
> > > > > On Thu, Dec 21, 2023 at 12:35 AM Jiri Olsa <olsajiri@xxxxxxxxx> wrote:
> > > > >> you need to pick up only 'BTF_ID(func, ...)' IDs that belongs to SET8 lists,
> > > > >> which are bounded by __BTF_ID__set8__<name> symbols, which also provide size
> > > > >
> > > > > +1
> > > > >
> > > > >>>
> > > > >>> Maybe we need a codemod from:
> > > > >>>
> > > > >>> BTF_ID(func, ...
> > > > >>>
> > > > >>> to:
> > > > >>>
> > > > >>> BTF_ID(kfunc, ...
> > > > >>
> > > > >> I think it's better to keep just 'func' and not to do anything special for
> > > > >> kfuncs in resolve_btfids logic to keep it simple
> > > > >>
> > > > >> also it's going to be already in pahole so if we want to make a fix in future
> > > > >> you need to change pahole, resolve_btfids and possibly also kernel
> > > > >
> > > > > I still don't understand why you guys want to add it to vmlinux BTF.
> > > > > The kernel has no use in this additional data.
> > > > > It already knows about all kfuncs.
> > > > > This extra memory is a waste of space from kernel pov.
> > > > > Unless I am missing something.
> > > > >
> > > > > imo this logic belongs in bpftool only.
> > > > > It can dump vmlinux BTF and emit __ksym protos into vmlinux.h
> > > > >
> > > >
> > > > If the goal is to have bpftool detect all kfuncs, would having a BPF
> > > > kfunc iterator that bpftool could use to iterate over registered kfuncs
> > > > work perhaps?
> > >
> > > The kernel code ? Why ?
> > > bpftool can do the same thing as this patch. Iterate over set8 in vmlinux elf.
> >
> > I think you're right for vmlinux -- bpftool can look at the elf file on
> > a running system. But I'm not sure it works for modules.
> >
> > IIUC, the actual module ELF can go away while the kernel holds onto the
> > memory (as with initramfs). And even if that wasn't the case, in
> > containerized environments you may not be able to always see
> > /lib/modules or similar.
>
> Indeed. Access to .ko files may be difficult even for full root
> without containers.
>
> What is vmlinux BTF before/after for our current set of kfuncs ?
Before:
$ pahole -J --btf_gen_floats -j --lang_exclude=rust --skip_encoding_btf_inconsistent_proto --btf_gen_optimized .tmp_vmlinux.btf
$ ls -l .tmp_vmlinux.btf
-rwxr-xr-x 1 dxu dxu 1159241688 Dec 22 13:44 .tmp_vmlinux.btf*
After:
$ /home/dxu/dev/pahole/build/pahole -J --btf_gen_floats -j --lang_exclude=rust --skip_encoding_btf_inconsistent_proto --btf_gen_optimized .tmp_vmlinux.btf
$ ls -l .tmp_vmlinux.btf
-rwxr-xr-x 1 dxu dxu 1159248104 Dec 22 13:47 .tmp_vmlinux.btf*
1159248104 - 1159241688 = 6416, so ~17B per kfunc (although we are
currently overcounting kfuncs by a bit).
Btw, for some reason the file size is not quite reproducible. I'm seeing
~500B deltas every time I rerun the same command. Maybe I'm doing
something wrong? Not sure.
Thanks,
Daniel
