[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20180227102638.19e8cd53@ThinkPad>
Date: Tue, 27 Feb 2018 10:26:38 +0100
From: Philipp Rudo <prudo@...ux.vnet.ibm.com>
To: AKASHI Takahiro <takahiro.akashi@...aro.org>
Cc: catalin.marinas@....com, will.deacon@....com,
bauerman@...ux.vnet.ibm.com, dhowells@...hat.com,
vgoyal@...hat.com, herbert@...dor.apana.org.au,
davem@...emloft.net, akpm@...ux-foundation.org, mpe@...erman.id.au,
dyoung@...hat.com, bhe@...hat.com, arnd@...db.de,
ard.biesheuvel@...aro.org, julien.thierry@....com,
kexec@...ts.infradead.org, linux-kernel@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH v8 03/13] kexec_file, x86, powerpc: factor out
kexec_file_ops functions
On Tue, 27 Feb 2018 11:03:07 +0900
AKASHI Takahiro <takahiro.akashi@...aro.org> wrote:
> On Mon, Feb 26, 2018 at 12:17:18PM +0100, Philipp Rudo wrote:
> > Hi AKASHI
> >
> > On Thu, 22 Feb 2018 20:17:22 +0900
> > AKASHI Takahiro <takahiro.akashi@...aro.org> wrote:
> >
> > [...]
> >
> > > diff --git a/kernel/kexec_file.c b/kernel/kexec_file.c
> > > index 990adae52151..a6d14a768b3e 100644
> > > --- a/kernel/kexec_file.c
> > > +++ b/kernel/kexec_file.c
> > > @@ -26,34 +26,83 @@
> > > #include <linux/vmalloc.h>
> > > #include "kexec_internal.h"
> > >
> > > +const __weak struct kexec_file_ops * const kexec_file_loaders[] = {NULL};
> > > +
> >
> > Having a weak definition of kexec_file_loaders causes trouble on s390 with
> > gcc 4.8 (newer versions seem to work fine). For me it looks like that in this
> > version gcc doesn't recognize __weak but use the default value for
> > optimization. This leads to _kexec_kernel_image_probe to always return ENOEXEC
> > because the for-loop gets optimized out.
>
> I gave it a try to compile with gcc 4.9 (not 4.8) for arm64
> and didn't see any errors or warnings, but
I talked to our compiler guys, and it looks like its a bug in gcc which was
introduced with gcc 4.8 and removed again with gcc 4.9. So I was just extremely
lucky hitting the sweat spot...
> > The problem can easily be worked around by declaring kexec_file_loaders in
> > include/linux/kexec.h and defining it in arch code. In particular doing this
> >
> > diff --git a/include/linux/kexec.h b/include/linux/kexec.h
> > index 37e9dce518aa..fc0788540d90 100644
> > --- a/include/linux/kexec.h
> > +++ b/include/linux/kexec.h
> > @@ -139,6 +139,8 @@ struct kexec_file_ops {
> > #endif
> > };
> >
> > +extern const struct kexec_file_ops * const kexec_file_loaders[];
> > +
> > /**
> > * struct kexec_buf - parameters for finding a place for a buffer in memory
> > * @image: kexec image in which memory to search.
> > diff --git a/kernel/kexec_file.c b/kernel/kexec_file.c
> > index 17ba407d0e79..4e3d1e4bc7f6 100644
> > --- a/kernel/kexec_file.c
> > +++ b/kernel/kexec_file.c
> > @@ -31,8 +31,6 @@
> > #include <linux/vmalloc.h>
> > #include "kexec_internal.h"
> >
> > -const __weak struct kexec_file_ops * const kexec_file_loaders[] = {NULL};
> > -
> > #ifdef CONFIG_ARCH_HAS_KEXEC_PURGATORY
> > static int kexec_calculate_store_digests(struct kimage *image);
> > #else
>
> Your change is just fine with me, too.
> I will incorporate it in my next version.
Thanks a lot
Philipp
> Thanks,
> -Takahiro AKASHI
>
> > A nice side effect of this solution is, that a developer who forgets to define
> > kexec_file_loaders gets a linker error. So he directly knows what's missing
> > instead of first having to find out where/why an error gets returned.
> >
> > Otherwise the series is fine for me.
> >
> > Thanks
> > Philipp
> >
> > > #ifdef CONFIG_ARCH_HAS_KEXEC_PURGATORY
> > > static int kexec_calculate_store_digests(struct kimage *image);
> > > #else
> > > static int kexec_calculate_store_digests(struct kimage *image) { return 0; };
> > > #endif
> > >
> > > +int _kexec_kernel_image_probe(struct kimage *image, void *buf,
> > > + unsigned long buf_len)
> > > +{
> > > + const struct kexec_file_ops * const *fops;
> > > + int ret = -ENOEXEC;
> > > +
> > > + for (fops = &kexec_file_loaders[0]; *fops && (*fops)->probe; ++fops) {
> > > + ret = (*fops)->probe(buf, buf_len);
> > > + if (!ret) {
> > > + image->fops = *fops;
> > > + return ret;
> > > + }
> > > + }
> > > +
> > > + return ret;
> > > +}
> > > +
> > > /* Architectures can provide this probe function */
> > > int __weak arch_kexec_kernel_image_probe(struct kimage *image, void *buf,
> > > unsigned long buf_len)
> > > {
> > > - return -ENOEXEC;
> > > + return _kexec_kernel_image_probe(image, buf, buf_len);
> > > +}
> > > +
> > > +void *_kexec_kernel_image_load(struct kimage *image)
> > > +{
> > > + if (!image->fops || !image->fops->load)
> > > + return ERR_PTR(-ENOEXEC);
> > > +
> > > + return image->fops->load(image, image->kernel_buf,
> > > + image->kernel_buf_len, image->initrd_buf,
> > > + image->initrd_buf_len, image->cmdline_buf,
> > > + image->cmdline_buf_len);
> > > }
> > >
> > > void * __weak arch_kexec_kernel_image_load(struct kimage *image)
> > > {
> > > - return ERR_PTR(-ENOEXEC);
> > > + return _kexec_kernel_image_load(image);
> > > +}
> > > +
> > > +int _kimage_file_post_load_cleanup(struct kimage *image)
> > > +{
> > > + if (!image->fops || !image->fops->cleanup)
> > > + return 0;
> > > +
> > > + return image->fops->cleanup(image->image_loader_data);
> > > }
> > >
> > > int __weak arch_kimage_file_post_load_cleanup(struct kimage *image)
> > > {
> > > - return -EINVAL;
> > > + return _kimage_file_post_load_cleanup(image);
> > > }
> > >
> > > #ifdef CONFIG_KEXEC_VERIFY_SIG
> > > +int _kexec_kernel_verify_sig(struct kimage *image, void *buf,
> > > + unsigned long buf_len)
> > > +{
> > > + if (!image->fops || !image->fops->verify_sig) {
> > > + pr_debug("kernel loader does not support signature verification.\n");
> > > + return -EKEYREJECTED;
> > > + }
> > > +
> > > + return image->fops->verify_sig(buf, buf_len);
> > > +}
> > > +
> > > int __weak arch_kexec_kernel_verify_sig(struct kimage *image, void *buf,
> > > unsigned long buf_len)
> > > {
> > > - return -EKEYREJECTED;
> > > + return _kexec_kernel_verify_sig(image, buf, buf_len);
> > > }
> > > #endif
> > >
> >
>
Powered by blists - more mailing lists