diff mbox series

[v2] crypto: aspeed: fix build module error

Message ID 20220905025433.1610696-1-neal_liu@aspeedtech.com
State Handled Elsewhere, archived
Headers show
Series [v2] crypto: aspeed: fix build module error | expand

Commit Message

Neal Liu Sept. 5, 2022, 2:54 a.m. UTC 
If CONFIG_MODULES=y and CONFIG_CRYPTO_DEV_ASPEED=m,
build modpost would be failed.

Error messages:
  ERROR: modpost: "aspeed_register_hace_hash_algs"
  [drivers/crypto/aspeed/aspeed_crypto.ko] undefined!
  ERROR: modpost: "aspeed_unregister_hace_hash_algs"
  [drivers/crypto/aspeed/aspeed_crypto.ko] undefined!

Change build sequence to fix this.

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Neal Liu <neal_liu@aspeedtech.com>
Tested-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
---
v2: Remove redundant obj- lines.

 drivers/crypto/aspeed/Makefile | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

Comments

Herbert Xu Sept. 5, 2022, 11:04 a.m. UTC | #1 
On Mon, Sep 05, 2022 at 10:54:33AM +0800, Neal Liu wrote:
>
> diff --git a/drivers/crypto/aspeed/Makefile b/drivers/crypto/aspeed/Makefile
> index 421e2ca9c53e..3be78cec0ecb 100644
> --- a/drivers/crypto/aspeed/Makefile
> +++ b/drivers/crypto/aspeed/Makefile
> @@ -1,9 +1,6 @@
> +hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) := aspeed-hace.o aspeed-hace-hash.o
> +hace-crypto-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) := aspeed-hace.o aspeed-hace-crypto.o
> +
>  obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed_crypto.o
> -aspeed_crypto-objs := aspeed-hace.o	\
> -		      $(hace-hash-y)	\
> +aspeed_crypto-objs := $(hace-hash-y)	\
>  		      $(hace-crypto-y)

Does this still build if both HASH and CRYPTO are off?

I think this it's best if you do:

hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) := aspeed-hace-hash.o
hace-crypto-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) := aspeed-hace-crypto.o

obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed_crypto.o
aspeed_crypto-objs := aspeed-hace.o	\
		      $(hace-hash-y)	\
		      $(hace-crypto-y)

Thanks,
Neal Liu Sept. 6, 2022, 2:21 a.m. UTC | #2 
> -----Original Message-----
> From: Herbert Xu <herbert@gondor.apana.org.au>
> Sent: Monday, September 5, 2022 7:04 PM
> To: Neal Liu <neal_liu@aspeedtech.com>
> Cc: David S . Miller <davem@davemloft.net>; Joel Stanley <joel@jms.id.au>;
> Andrew Jeffery <andrew@aj.id.au>; linux-aspeed@lists.ozlabs.org;
> linux-crypto@vger.kernel.org; linux-arm-kernel@lists.infradead.org;
> linux-kernel@vger.kernel.org; BMC-SW <BMC-SW@aspeedtech.com>; kernel
> test robot <lkp@intel.com>; Sudip Mukherjee
> <sudipm.mukherjee@gmail.com>
> Subject: Re: [PATCH v2] crypto: aspeed: fix build module error
> 
> On Mon, Sep 05, 2022 at 10:54:33AM +0800, Neal Liu wrote:
> >
> > diff --git a/drivers/crypto/aspeed/Makefile
> > b/drivers/crypto/aspeed/Makefile index 421e2ca9c53e..3be78cec0ecb
> > 100644
> > --- a/drivers/crypto/aspeed/Makefile
> > +++ b/drivers/crypto/aspeed/Makefile
> > @@ -1,9 +1,6 @@
> > +hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) := aspeed-hace.o
> > +aspeed-hace-hash.o
> > +hace-crypto-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) :=
> aspeed-hace.o
> > +aspeed-hace-crypto.o
> > +
> >  obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed_crypto.o
> > -aspeed_crypto-objs := aspeed-hace.o	\
> > -		      $(hace-hash-y)	\
> > +aspeed_crypto-objs := $(hace-hash-y)	\
> >  		      $(hace-crypto-y)
> 
> Does this still build if both HASH and CRYPTO are off?
> 
> I think this it's best if you do:
> 
> hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) :=
> aspeed-hace-hash.o
> hace-crypto-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) :=
> aspeed-hace-crypto.o
> 
> obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed_crypto.o
> aspeed_crypto-objs := aspeed-hace.o	\
> 		      $(hace-hash-y)	\
> 		      $(hace-crypto-y)
> 

aspeed-hace.o effects only if either hace-hash-y or hace-crypto-y.
If we put aspeed-hace.o in aspeed_crypto-objs, but hace-hash-y and hace-crypto-y are empty, apseed-hace.o is just an useless driver which might still occupy system resources.

For this patch, you're right it would still build if both HASH & CRYPTO are off. But no driver would be run up.
That's why I revise it from your suggestion to this patch.
Thanks.

> Thanks,
> --
> Email: Herbert Xu <herbert@gondor.apana.org.au> Home Page:
> http://gondor.apana.org.au/~herbert/
> PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
Herbert Xu Sept. 6, 2022, 4:53 a.m. UTC | #3 
On Mon, Sep 05, 2022 at 10:54:33AM +0800, Neal Liu wrote:
> If CONFIG_MODULES=y and CONFIG_CRYPTO_DEV_ASPEED=m,
> build modpost would be failed.
> 
> Error messages:
>   ERROR: modpost: "aspeed_register_hace_hash_algs"
>   [drivers/crypto/aspeed/aspeed_crypto.ko] undefined!
>   ERROR: modpost: "aspeed_unregister_hace_hash_algs"
>   [drivers/crypto/aspeed/aspeed_crypto.ko] undefined!
> 
> Change build sequence to fix this.
> 
> Reported-by: kernel test robot <lkp@intel.com>
> Signed-off-by: Neal Liu <neal_liu@aspeedtech.com>
> Tested-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
> ---
> v2: Remove redundant obj- lines.
> 
>  drivers/crypto/aspeed/Makefile | 11 ++++-------
>  1 file changed, 4 insertions(+), 7 deletions(-)

Patch applied.  Thanks.
Herbert Xu Sept. 16, 2022, 10:42 a.m. UTC | #4 
On Tue, Sep 06, 2022 at 02:21:20AM +0000, Neal Liu wrote:
>
> > Does this still build if both HASH and CRYPTO are off?
> > 
> > I think this it's best if you do:
> > 
> > hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) :=
> > aspeed-hace-hash.o
> > hace-crypto-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) :=
> > aspeed-hace-crypto.o
> > 
> > obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed_crypto.o
> > aspeed_crypto-objs := aspeed-hace.o	\
> > 		      $(hace-hash-y)	\
> > 		      $(hace-crypto-y)
> > 
> 
> aspeed-hace.o effects only if either hace-hash-y or hace-crypto-y.
> If we put aspeed-hace.o in aspeed_crypto-objs, but hace-hash-y and hace-crypto-y are empty, apseed-hace.o is just an useless driver which might still occupy system resources.

Apparently it doesn't build after all, at least not on m68k.

So please either adopt my suggestion above, or come up with another
way of preventing the build failure on m68k with both HASH and CRYPTO
disabled.

Thanks,
Dhananjay Phadke Sept. 16, 2022, 5:21 p.m. UTC | #5 
On 9/16/2022 3:42 AM, Herbert Xu wrote:
>>> Does this still build if both HASH and CRYPTO are off?
>>>
>>> I think this it's best if you do:
>>>
>>> hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) :=
>>> aspeed-hace-hash.o
>>> hace-crypto-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) :=
>>> aspeed-hace-crypto.o
>>>
>>> obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed_crypto.o
>>> aspeed_crypto-objs := aspeed-hace.o	\
>>> 		      $(hace-hash-y)	\
>>> 		      $(hace-crypto-y)
>>>
>> aspeed-hace.o effects only if either hace-hash-y or hace-crypto-y.
>> If we put aspeed-hace.o in aspeed_crypto-objs, but hace-hash-y and hace-crypto-y are empty, apseed-hace.o is just an useless driver which might still occupy system resources.
> Apparently it doesn't build after all, at least not on m68k.
> 
> So please either adopt my suggestion above, or come up with another
> way of preventing the build failure on m68k with both HASH and CRYPTO
> disabled.

Curious why compiled on m68k? It's embedded controller in ARM based
Aspeed SoCs. And there's "depends on ARCH_ASPEED" in Kconfig, need
some additional dependencies?

Regards,
Dhananjay
Neal Liu Sept. 19, 2022, 6:36 a.m. UTC | #6 
> >>> Does this still build if both HASH and CRYPTO are off?
> >>>
> >>> I think this it's best if you do:
> >>>
> >>> hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) :=
> >>> aspeed-hace-hash.o
> >>> hace-crypto-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) :=
> >>> aspeed-hace-crypto.o
> >>>
> >>> obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed_crypto.o
> >>> aspeed_crypto-objs := aspeed-hace.o	\
> >>> 		      $(hace-hash-y)	\
> >>> 		      $(hace-crypto-y)
> >>>
> >> aspeed-hace.o effects only if either hace-hash-y or hace-crypto-y.
> >> If we put aspeed-hace.o in aspeed_crypto-objs, but hace-hash-y and
> hace-crypto-y are empty, apseed-hace.o is just an useless driver which might
> still occupy system resources.
> > Apparently it doesn't build after all, at least not on m68k.
> >
> > So please either adopt my suggestion above, or come up with another
> > way of preventing the build failure on m68k with both HASH and CRYPTO
> > disabled.
> 
> Curious why compiled on m68k? It's embedded controller in ARM based
> Aspeed SoCs. And there's "depends on ARCH_ASPEED" in Kconfig, need some
> additional dependencies?

The reason is because the compile test is enabled.
Check this:
https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=31b39755e32568b43c80814c5e13d7b1ab796d73
diff mbox series

Patch

diff --git a/drivers/crypto/aspeed/Makefile b/drivers/crypto/aspeed/Makefile
index 421e2ca9c53e..3be78cec0ecb 100644
--- a/drivers/crypto/aspeed/Makefile
+++ b/drivers/crypto/aspeed/Makefile
@@ -1,9 +1,6 @@ 
+hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) := aspeed-hace.o aspeed-hace-hash.o
+hace-crypto-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) := aspeed-hace.o aspeed-hace-crypto.o
+
 obj-$(CONFIG_CRYPTO_DEV_ASPEED) += aspeed_crypto.o
-aspeed_crypto-objs := aspeed-hace.o	\
-		      $(hace-hash-y)	\
+aspeed_crypto-objs := $(hace-hash-y)	\
 		      $(hace-crypto-y)
-
-obj-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) += aspeed-hace-hash.o
-hace-hash-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_HASH) := aspeed-hace-hash.o
-obj-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) += aspeed-hace-crypto.o
-hace-crypto-$(CONFIG_CRYPTO_DEV_ASPEED_HACE_CRYPTO) := aspeed-hace-crypto.o