[PATCH RFC] ARM: option for loading modules into vmalloc area

From: nicolas.pitre@linaro.org (Nicolas Pitre)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH RFC] ARM: option for loading modules into vmalloc area
Date: Wed, 19 Nov 2014 13:22:46 -0500 (EST)	[thread overview]
Message-ID: <alpine.LFD.2.11.1411191312110.11690@knanqh.ubzr> (raw)
In-Reply-To: <CAKv+Gu-q7Vdd=Buaz2x=8HASCpeY5D-RsOm-120bkOqmgtUeTw@mail.gmail.com>

On Wed, 19 Nov 2014, Ard Biesheuvel wrote:

> On 19 November 2014 18:12, Nicolas Pitre <nicolas.pitre@linaro.org> wrote:
> > On Wed, 19 Nov 2014, Russell King - ARM Linux wrote:
> >
> >> On Wed, Nov 19, 2014 at 11:57:15AM -0500, Nicolas Pitre wrote:
> >> > On Wed, 19 Nov 2014, Russell King - ARM Linux wrote:
> >> > > I don't think I ever did, because its pretty much impossible to do as I
> >> > > explained in a follow up to this thread.
> >> > >
> >> > > We _used_ to do this with the userspace insmod methods, but since we got
> >> > > this kernel-side linker, it's been pretty much impossible to do without
> >> > > rewriting the module code.  That's not going to happen on account of one
> >> > > quirky architecture which Linus doesn't particularly like.
> >> >
> >> > Still... We could try adding a hook in the generic module linker code
> >> > for a pre-relocation pass.  Maybe only ARM would use it, but if the need
> >> > to load big modules is real then I imagine Linus could be amenable to a
> >> > compromise.
> >>
> >> So, how big a table would you allocate for the trampolines, based upon
> >> not knowing anything about the module being loaded?  4K?  8K?  64K?
> >
> > The idea of a pre-relocation pass is to determine that.  That could be
> > something similar to calling apply_relocate() twice: once to determine
> > the number of trampoline entries, and a second time to perform the
> > actual relocation.
> >
> 
> Well, the veneers shouldn't take more than 3 words each, right?
> 
> ldr ip, [pc]
> bx ip
> .long symbol

You could possibly do:

 ldr pc, [pc, #-4]
 .long symbol

Or, as RMK suggested a while ago:

.rep 8
ldr pc, [pc, #(32 - 8)]
.endr
.long sym1, sym2, sym3, sym4, sym5, sym6, sym7, sym8

The later is much nicer on the i and d caches.

> and you would need at most one veneer per unique external symbol
> referenced by one or more R_ARM_CALL relocations. Is there no way to
> just add that to the static mem footprint as padding, and let the
> loader populate it as needed at module relocation time?

That's the actual question: how much padding do you need?  Everything 
converge to that very problem.  We need to determine it without too much 
impact on the generic module loader code.

Nicolas