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

[Heinrich Schuchardt]

On 23.04.20 02:28, AKASHI Takahiro wrote:
> Heinrich,
>
> On Tue, Apr 21, 2020 at 12:23:13PM +0200, Heinrich Schuchardt wrote:
>> On 4/21/20 2:37 AM, 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/crypto/pkcs7_parser.h         | 4 ++++
>>>  lib/crypto/x509_parser.h          | 4 ++++
>>>  lib/efi_loader/efi_image_loader.c | 1 +
>>>  lib/efi_loader/efi_signature.c    | 1 +
>>>  4 files changed, 10 insertions(+)
>>>
>>> diff --git a/lib/crypto/pkcs7_parser.h b/lib/crypto/pkcs7_parser.h
>>> index 6565fdc2d4ca..b8234da45a6a 100644
>>> --- a/lib/crypto/pkcs7_parser.h
>>> +++ b/lib/crypto/pkcs7_parser.h
>>> @@ -5,6 +5,9 @@
>>>   * Written by David Howells (dhowells at redhat.com)
>>>   */
>>>
>>> +#ifndef _PKCS7_PARSER_H
>>> +#define _PKCS7_PARSER_H
>>> +
>>>  #include <linux/oid_registry.h>
>>>  #include <crypto/pkcs7.h>
>>>  #include "x509_parser.h"
>>> @@ -63,3 +66,4 @@ struct pkcs7_message {
>>>  	size_t		data_hdrlen;	/* Length of Data ASN.1 header */
>>>  	const void	*data;		/* Content Data (or 0) */
>>>  };
>>> +#endif /* _PKCS7_PARSER_H */
>>> diff --git a/lib/crypto/x509_parser.h b/lib/crypto/x509_parser.h
>>> index c233f136fb35..4cbdc1d6612d 100644
>>> --- a/lib/crypto/x509_parser.h
>>> +++ b/lib/crypto/x509_parser.h
>>> @@ -5,6 +5,9 @@
>>>   * Written by David Howells (dhowells at redhat.com)
>>>   */
>>>
>>> +#ifndef _X509_PARSER_H
>>> +#define _X509_PARSER_H
>>> +
>>>  #include <linux/time.h>
>>>  #include <crypto/public_key.h>
>>>  #include <keys/asymmetric-type.h>
>>> @@ -55,3 +58,4 @@ extern int x509_decode_time(time64_t *_t,  size_t hdrlen,
>>>   */
>>>  extern int x509_get_sig_params(struct x509_certificate *cert);
>>>  extern int x509_check_for_self_signed(struct x509_certificate *cert);
>>> +#endif /* _X509_PARSER_H */
>>> diff --git a/lib/efi_loader/efi_image_loader.c b/lib/efi_loader/efi_image_loader.c
>>> index 6c270ce94f44..899adf8493d0 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 "../lib/crypto/pkcs7_parser.h"
>>> +#include "../lib/crypto/x509_parser.h"
>>
>> As pkcs7_parser.h includes x509_parser.h there is no reason to include
>> it again.
>
> No, I don't think so.
>
> My style of coding is that, if a symbol is used in a C file and it is
> defined in a header file, the header file should be *directly* included
> in the C file.
> I believe that it is the common way so that we can avoid any "implicit"
> dependencies among header files.

We want U-Boot to build as fast as possible. So we should not include
anything twice. When looking through the rest of the U-Boot code you
will have noticed that we do not include malloc.h and stdio.h everywhere
because it is included via common.h.

As an example of a patch removing other unnecessary includes see for
instance:

cmd: fat: remove unused includes
fd0e30b43b6b2401e68dc32c357869c617d4fdd1

Best regards

Heinrich

>
> Thanks,
> -Takahiro Akashi
>
>
>>>
>>>  const efi_guid_t efi_global_variable_guid = EFI_GLOBAL_VARIABLE_GUID;
>>>  const efi_guid_t efi_guid_device_path = EFI_DEVICE_PATH_PROTOCOL_GUID;
>>> diff --git a/lib/efi_loader/efi_signature.c b/lib/efi_loader/efi_signature.c
>>> index 658e3547da37..6ad09e4acbd7 100644
>>> --- a/lib/efi_loader/efi_signature.c
>>> +++ b/lib/efi_loader/efi_signature.c
>>> @@ -15,6 +15,7 @@
>>>  #include <u-boot/rsa.h>
>>>  #include <u-boot/sha256.h>
>>>  #include "../lib/crypto/pkcs7_parser.h"
>>> +#include "../lib/crypto/x509_parser.h"
>>
>> Same here.
>>
>> The rest is correct and should be merged.
>>
>> Best regards
>>
>> Heinrich
>>
>>>
>>>  const efi_guid_t efi_guid_image_security_database =
>>>  		EFI_IMAGE_SECURITY_DATABASE_GUID;
>>>
>>