[Buildroot] [PATCH] package/libopenssl does not support riscv32

[Buildroot] [PATCH] package/libopenssl does not support riscv32

[Yann E. MORIN]

riscv32 is (surprise!) a 32-bit architecture. But it has been Y2038-safe
from its inception. As such, there are no legacy binaries that may use
the 32-bit time syscalls, and thus they are not available on riscv32.

Code that directly calls to the syscalls without using the C libraries
wrappers thus need to handle this case by themselves. That's what
upstream tried to do with:
    https://github.com/openssl/openssl/commit/5b5e2985f355c8e99c196d9ce5d02c15bebadfbc

We initially carried that patch with 2bb26c1a1d24 (package/libopenssl:
fix build on riscv32).

However, as Arnd Bergmann puts it [0]:

    The patch looks wrong to me: __NR_io_pgetevents_time64 must be used
    whenever time_t is 64-bit wide on a 32-bit architecture, while
    __NR_io_getevents/__NR_io_pgetevents must be used when time_t is the
    same width as 'long'.

    Checking whether __NR_io_getevents is defined is wrong for all
    architectures other than riscv

And Arnd agrees that patch should be reverted [1] [2] (there are further
comments in that stream, that are worth reading).

As such, we've reverted 2bb26c1a1d24 with 6cfb4ad7f76a.

This means we have no working solution to enable openssl on riscv32 for
now. So, rather than fail the build, or backport a dysfunctional patch,
let's just forbid openssl on riscv32.

Drop the default from the choice selection; it was anyway superfluous:
the default of a choice, if left unspecified, is the first entry of the
choice.

While at it, drop a useless redundant comment in ibm-sw-tpm2: if we
select FORCE_LIBOPENSSL, it is obvious that's because libressl is not
supported... Besides none of the other users of FORCE_LIBOPENSSL have
such a comment.

Fixes:
    http://autobuild.buildroot.org/results/eb9/eb9a64d4ffae8569b5225083f282cf87ffa7c681/
    ...
    http://autobuild.buildroot.org/results/07e/07e413b24ba8adc9558c80267ce16dda339bf032/

[0] https://github.com/openssl/openssl/commit/5b5e2985f355c8e99c196d9ce5d02c15bebadfbc#commitcomment-44782859
[1] https://github.com/openssl/openssl/commit/5b5e2985f355c8e99c196d9ce5d02c15bebadfbc#commitcomment-47826509
[2] https://github.com/openssl/openssl/commit/5b5e2985f355c8e99c196d9ce5d02c15bebadfbc#commitcomment-47830530

Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Matthew Weber <matthew.weber@rockwellcollins.com>
Cc: Mark Corbin <mark@dibsco.co.uk>
---
 package/hostapd/Config.in        | 1 +
 package/ibm-sw-tpm2/Config.in    | 3 ++-
 package/openssl/Config.in        | 8 +++++++-
 package/softether/Config.in      | 2 ++
 package/tpm2-tss/Config.in       | 2 ++
 package/wpa_supplicant/Config.in | 2 ++
 6 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/package/hostapd/Config.in b/package/hostapd/Config.in
index 54d2086c09..38cead90ca 100644
--- a/package/hostapd/Config.in
+++ b/package/hostapd/Config.in
@@ -76,6 +76,7 @@ config BR2_PACKAGE_HOSTAPD_WPS
 
 config BR2_PACKAGE_HOSTAPD_WPA3
 	bool "Enable WPA3 support"
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	select BR2_PACKAGE_OPENSSL
 	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
 	help
diff --git a/package/ibm-sw-tpm2/Config.in b/package/ibm-sw-tpm2/Config.in
index b402ba7150..57781a54bb 100644
--- a/package/ibm-sw-tpm2/Config.in
+++ b/package/ibm-sw-tpm2/Config.in
@@ -2,8 +2,8 @@ config BR2_PACKAGE_IBM_SW_TPM2
 	bool "ibm-sw-tpm2"
 	depends on BR2_TOOLCHAIN_HAS_THREADS
 	depends on !BR2_STATIC_LIBS # ibm-sw-tpm2
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	select BR2_PACKAGE_OPENSSL
-	# libressl is not supported
 	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
 	help
 	  This project is an implementation of the TCG TPM 2.0
@@ -14,4 +14,5 @@ config BR2_PACKAGE_IBM_SW_TPM2
 	  https://sourceforge.net/projects/ibmswtpm2/
 
 comment "ibm-sw-tpm2 needs a toolchain w/ threads, dynamic library"
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	depends on !BR2_TOOLCHAIN_HAS_THREADS || BR2_STATIC_LIBS
diff --git a/package/openssl/Config.in b/package/openssl/Config.in
index 5ce91cb80f..dd47091300 100644
--- a/package/openssl/Config.in
+++ b/package/openssl/Config.in
@@ -1,5 +1,11 @@
+config BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
+	bool
+	default y
+	depends on !BR2_RISCV_32
+
 config BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
 	bool
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 
 config BR2_PACKAGE_OPENSSL
 	bool "openssl support"
@@ -11,12 +17,12 @@ if BR2_PACKAGE_OPENSSL
 
 choice
 	prompt "ssl library"
-	default BR2_PACKAGE_LIBOPENSSL
 	help
 	  Select OpenSSL or LibreSSL.
 
 config BR2_PACKAGE_LIBOPENSSL
 	bool "openssl"
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	select BR2_PACKAGE_ZLIB
 	help
 	  A collaborative effort to develop a robust, commercial-grade,
diff --git a/package/softether/Config.in b/package/softether/Config.in
index bfb45b0315..01104b5d39 100644
--- a/package/softether/Config.in
+++ b/package/softether/Config.in
@@ -3,6 +3,7 @@ config BR2_PACKAGE_SOFTETHER
 	depends on BR2_TOOLCHAIN_HAS_THREADS
 	depends on BR2_USE_MMU # fork()
 	depends on BR2_USE_WCHAR
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	select BR2_PACKAGE_LIBICONV if !BR2_ENABLE_LOCALE
 	select BR2_PACKAGE_OPENSSL
 	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
@@ -27,4 +28,5 @@ config BR2_PACKAGE_SOFTETHER
 
 comment "softether needs a toolchain w/ wchar, threads"
 	depends on BR2_USE_MMU
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	depends on !(BR2_USE_WCHAR && BR2_TOOLCHAIN_HAS_THREADS)
diff --git a/package/tpm2-tss/Config.in b/package/tpm2-tss/Config.in
index 2fc56dbd76..b1ab00dd38 100644
--- a/package/tpm2-tss/Config.in
+++ b/package/tpm2-tss/Config.in
@@ -1,6 +1,7 @@
 config BR2_PACKAGE_TPM2_TSS
 	bool "tpm2-tss"
 	depends on !BR2_STATIC_LIBS # dlfcn.h
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	select BR2_PACKAGE_LIBURIPARSER
 	select BR2_PACKAGE_OPENSSL
 	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
@@ -55,4 +56,5 @@ config BR2_PACKAGE_TPM2_TSS_FAPI
 endif
 
 comment "tpm2-tss needs a toolchain w/ dynamic library"
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	depends on BR2_STATIC_LIBS
diff --git a/package/wpa_supplicant/Config.in b/package/wpa_supplicant/Config.in
index cba7fa07e1..c6e4de1d48 100644
--- a/package/wpa_supplicant/Config.in
+++ b/package/wpa_supplicant/Config.in
@@ -49,6 +49,7 @@ config BR2_PACKAGE_WPA_SUPPLICANT_WIFI_DISPLAY
 config BR2_PACKAGE_WPA_SUPPLICANT_MESH_NETWORKING
 	bool "Enable mesh networking"
 	depends on BR2_PACKAGE_WPA_SUPPLICANT_AP_SUPPORT
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	select BR2_PACKAGE_OPENSSL
 	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
 	help
@@ -83,6 +84,7 @@ config BR2_PACKAGE_WPA_SUPPLICANT_WPS
 
 config BR2_PACKAGE_WPA_SUPPLICANT_WPA3
 	bool "Enable WPA3 support"
+	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
 	select BR2_PACKAGE_OPENSSL
 	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
 	help
-- 
2.25.1

[Buildroot] [PATCH] package/libopenssl does not support riscv32

[Arnout Vandecappelle]

 Hi Yann,

On 04/03/2021 12:14, Yann E. MORIN wrote:

> Drop the default from the choice selection; it was anyway superfluous:
> the default of a choice, if left unspecified, is the first entry of the
> choice.

 Bikeshedding time!

 IMHO, it is still good to have an explicit default. Its effect remains exactly
the same (if the default cannot be selected due to missing dependencies, the
first available options will be chosen), but it makes it clear for the reader
what is the default, and it makes sure that inserting a new option or reordering
the options doesn't affect the default.

[snip]
>  config BR2_PACKAGE_HOSTAPD_WPA3
>  	bool "Enable WPA3 support"
> +	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS

 We normally add a comment of depends on. But I guess it's pretty obvious here.

>  	select BR2_PACKAGE_OPENSSL
>  	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
>  	help


[snip]
> diff --git a/package/tpm2-tss/Config.in b/package/tpm2-tss/Config.in
> index 2fc56dbd76..b1ab00dd38 100644
> --- a/package/tpm2-tss/Config.in
> +++ b/package/tpm2-tss/Config.in
> @@ -1,6 +1,7 @@
>  config BR2_PACKAGE_TPM2_TSS

 Unfortunately, TPM2_TSS is selected by ima-evm-utils, tpm2-abrmd, tpm2-pkcs11,
tpm2-tools, tpm2-totp. Fortunately, it doesn't recurse any further.

[snip]
>  config BR2_PACKAGE_WPA_SUPPLICANT_MESH_NETWORKING

 WPA_SUPPLICANT is selected by connman but fortunately no mesh or WPA3. We got
lucky there!


 With the propagation of the TPM2_TSS dependency (and for those, a comment
*should* be added), you can add my

 Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

and apply directly to master.

 Regards,
 Arnout

>  	bool "Enable mesh networking"
>  	depends on BR2_PACKAGE_WPA_SUPPLICANT_AP_SUPPORT
> +	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
>  	select BR2_PACKAGE_OPENSSL
>  	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
>  	help
> @@ -83,6 +84,7 @@ config BR2_PACKAGE_WPA_SUPPLICANT_WPS
>  
>  config BR2_PACKAGE_WPA_SUPPLICANT_WPA3
>  	bool "Enable WPA3 support"
> +	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
>  	select BR2_PACKAGE_OPENSSL
>  	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
>  	help
> 

[Buildroot] [PATCH] package/libopenssl does not support riscv32

[Yann E. MORIN]

Arnout, All,

On 2021-03-04 20:36 +0100, Arnout Vandecappelle spake thusly:
> On 04/03/2021 12:14, Yann E. MORIN wrote:
> 
> > Drop the default from the choice selection; it was anyway superfluous:
> > the default of a choice, if left unspecified, is the first entry of the
> > choice.
>  Bikeshedding time!

Aha! ;-)

>  IMHO, it is still good to have an explicit default. Its effect remains exactly
> the same (if the default cannot be selected due to missing dependencies, the
> first available options will be chosen), but it makes it clear for the reader
> what is the default, and it makes sure that inserting a new option or reordering
> the options doesn't affect the default.

OK, I see and understand your point.

Still, defaults are cumbersome to manage when said default gains or
loses dependencies; that's another location where we need to track
those...

Besides, the probability tht we gain a third implementation of openssl
is... slim. Very slim...

So, I would still drop the default here.

> [snip]
> >  config BR2_PACKAGE_HOSTAPD_WPA3
> >  	bool "Enable WPA3 support"
> > +	depends on BR2_PACKAGE_LIBOPENSSL_ARCH_SUPPORTS
>  We normally add a comment of depends on. But I guess it's pretty obvious here.

Yes, it is pretty obvious. So much so that it really does not warrant a
comment, because it is really is overloading.

> >  	select BR2_PACKAGE_OPENSSL
> >  	select BR2_PACKAGE_OPENSSL_FORCE_LIBOPENSSL
> >  	help
> 
> [snip]
> > diff --git a/package/tpm2-tss/Config.in b/package/tpm2-tss/Config.in
> > index 2fc56dbd76..b1ab00dd38 100644
> > --- a/package/tpm2-tss/Config.in
> > +++ b/package/tpm2-tss/Config.in
> > @@ -1,6 +1,7 @@
> >  config BR2_PACKAGE_TPM2_TSS
>  Unfortunately, TPM2_TSS is selected by ima-evm-utils, tpm2-abrmd, tpm2-pkcs11,
> tpm2-tools, tpm2-totp. Fortunately, it doesn't recurse any further.

Ah, right, well spotted. I was not carefull enough on that one.

>  With the propagation of the TPM2_TSS dependency (and for those, a comment
> *should* be added),

s/should/must/

> you can add my
> 
>  Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
> 
> and apply directly to master.

Since you did not insist on keeping the default as a condition for your
rev-by tag, I'll still drop it. ;-)

Thanks!

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'