Re: [PATCH v2 0/4] x86/boot: Remove runtime relocations from compressed kernel

[Sedat Dilek]

On Tue, May 26, 2020 at 5:36 PM Arvind Sankar <nivedita@xxxxxxxxxxxx> wrote:
>
> On Tue, May 26, 2020 at 04:50:38PM +0200, Sedat Dilek wrote:
> > On Tue, May 26, 2020 at 4:47 PM Arvind Sankar <nivedita@xxxxxxxxxxxx> wrote:
> > >
> > > On Tue, May 26, 2020 at 02:44:29PM +0200, Sedat Dilek wrote:
> > > >
> > > > Are those diffs correct when using "x86/boot: Correct relocation
> > > > destination on old linkers"?
> > > >
> > >
> > > It looks ok, but that patch (and even marking the other symbols .hidden)
> > > should be unnecessary after this series.
> >
> > You mean _bss, _ebss and _end?
> >
> > - Sedat -
>
> Yes. Those .hidden markings are there to ensure that when relocations
> are generated (as they are currently), they're generated as
> R_386_RELATIVE (which uses B+A calculation, with A being the link-time
> virtual address of the symbol, and stored in the relocation field)
> rather than R_386_32 (which uses S+A calculation, and so doesn't work
> without runtime processing). After this patchset there aren't any
> relocations, so while the .hidden markings won't hurt, they won't be
> necessary either.
>

So, I am here on Debian/testing AMD64.

How can I test the patchset worked correctly?

- Sedat -