[PATCH v3 01/13] lib/crypto, efi_loader: avoid multiple inclusions of header files

[AKASHI Takahiro]

Heinrich,

On Wed, Jul 08, 2020 at 12:29:54PM +0200, Heinrich Schuchardt wrote:
> On 08.07.20 07:01, AKASHI Takahiro wrote:
> > By adding extra symbols, we can now avoid including x509_parser and
> > pkcs7_parser.h files multiple times.
> >
> > Signed-off-by: AKASHI Takahiro <takahiro.akashi at linaro.org>
> > Suggested-by: Heinrich Schuchardt <xypron.glpk at gmx.de>
> > ---
> >  lib/efi_loader/efi_image_loader.c | 1 +
> >  lib/efi_loader/efi_signature.c    | 1 +
> >  2 files changed, 2 insertions(+)
> >
> > diff --git a/lib/efi_loader/efi_image_loader.c b/lib/efi_loader/efi_image_loader.c
> > index 06a2ebdb9081..5b00fea2f113 100644
> > --- a/lib/efi_loader/efi_image_loader.c
> > +++ b/lib/efi_loader/efi_image_loader.c
> > @@ -14,6 +14,7 @@
> >  #include <pe.h>
> >  #include <sort.h>
> >  #include <crypto/pkcs7_parser.h>
> > +#include <crypto/x509_parser.h>
> >  #include <linux/err.h>
> >
> >  const efi_guid_t efi_global_variable_guid = EFI_GLOBAL_VARIABLE_GUID;
> > diff --git a/lib/efi_loader/efi_signature.c b/lib/efi_loader/efi_signature.c
> > index e05c471c61ce..da3818ac62e2 100644
> > --- a/lib/efi_loader/efi_signature.c
> > +++ b/lib/efi_loader/efi_signature.c
> > @@ -11,6 +11,7 @@
> >  #include <hexdump.h>
> >  #include <malloc.h>
> >  #include <crypto/pkcs7_parser.h>
> > +#include <crypto/x509_parser.h>
> >  #include <linux/compat.h>
> >  #include <linux/oid_registry.h>
> >  #include <u-boot/rsa.h>
> >

The patch (and the commit message) seems to have been
somehow corrupted during my rework. 

> I would prefer to use
> 
> [PATCH 1/1] lib/crypto: use qualified path for x509_parser.h
> https://lists.denx.de/pipermail/u-boot/2020-July/419214.html

My point is, as I said before, that we should include a header
*directly* if some symbols in that header are used in a given file
even though such a header is *indirectly* inherited by another
header, pkcs7_parser.h in this case.

That way, "#include" explicitly indicates that the C file relies on
another header and associated functions.
I always take this practice.

In this sense, this patch still makes sense.

-Takahiro Akashi

> 
> which has the same effect.
> 
> Best regards
> 
> Heinrich