Re: [Xen-devel] [PATCH] ARM: xen: unexport HYPERVISOR_platform_op function

[Mark Rutland <mark.rutland@xxxxxxx>]

Hi Julien,

On Sat, Sep 07, 2019 at 11:05:45AM +0100, Julien Grall wrote:
> On 9/6/19 6:20 PM, Andrew Cooper wrote:
> > On 06/09/2019 17:00, Arnd Bergmann wrote:
> > > On Fri, Sep 6, 2019 at 5:55 PM Andrew Cooper <andrew.cooper3@xxxxxxxxxx> 
> > > wrote:
> > > > On 06/09/2019 16:39, Arnd Bergmann wrote:
> > > > > HYPERVISOR_platform_op() is an inline function and should not
> > > > > be exported. Since commit 15bfc2348d54 ("modpost: check for
> > > > > static EXPORT_SYMBOL* functions"), this causes a warning:
> > > > > 
> > > > > WARNING: "HYPERVISOR_platform_op" [vmlinux] is a static 
> > > > > EXPORT_SYMBOL_GPL
> > > > > 
> > > > > Remove the extraneous export.
> > > > > 
> > > > > Fixes: 15bfc2348d54 ("modpost: check for static EXPORT_SYMBOL* 
> > > > > functions")
> > > > > Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
> > > > Something is wonky.  That symbol is (/ really ought to be) in the
> > > > hypercall page and most definitely not inline.
> > > > 
> > > > Which tree is that changeset from?  I can't find the SHA.
> > > This is from linux-next, I think from the kbuild tree.
> > 
> > Thanks.
> > 
> > Julien/Stefano: Why are any of these hypercalls out-of-line?  ARM
> > doesn't use the hypercall page, and there is no argument translation
> > (not even in arm32 as there are no 5-argument hypercalls declared).
> 
> I am not sure how the hypercall page makes things different. You still have
> to store the arguments in the correct register so...
> 
> > 
> > They'd surely be easier to implement with a few static inlines and some
> > common code, than to try and replicate the x86 side hypercall_page
> > interface ?
> 
> ... I don't think they will be easier to implement with a few static
> inlines. The implementation will likely end up to be similar to
> arch/x86/asm/xen/hypercall.h.
> 
> Furthermore, one of the downside of per-arch static inline is it is more
> difficult to ensure the prototype match for all the architectures. Although,
> it might be possible to make them common by only requesting per-arch to
> implement HYPERCALL_N(...).
> 
> So I think the code is better as it is.
> 
> While looking at the code, I also realized that the implementation of
> HYPERCALL_dm_op might be incorrect for Arm32. Similarly do privcmd call, I
> think dm_op call should enable user access as they will be used by
> userspace.
> 
> We don't use dm_op on Arm so far, hence why I think this was unnoticed. I
> will see if I can reproduce it and send a patch.

I'm seeing this when building arm64 defconfig v5.4-rc1:

| [mark@lakrids:~/src/linux]% usekorg 8.1.0  make ARCH=arm64 
CROSS_COMPILE=aarch64-linux- -j56 -s
| arch/arm64/Makefile:62: CROSS_COMPILE_COMPAT not defined or empty, the compat 
vDSO will not be built
| WARNING: "HYPERVISOR_platform_op" [vmlinux] is a static EXPORT_SYMBOL_GPL
| WARNING: "HYPERVISOR_platform_op" [vmlinux] is a static EXPORT_SYMBOL_GPL

I couldn't see a follow-up; do you have a patch for this?

Thanks,
Mark.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel