[Buildroot] [PATCH] package/rauc: needs kernel headers >= 3.0

[Buildroot] [PATCH] package/rauc: needs kernel headers >= 3.0

[Yann E. MORIN]

rauc includes <linux/mmc/ioctl.h> which was introduced only in
linux 3.0.

Also change order to have arch dependencies first.

Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Andrey Yurovsky <yurovsky@gmail.com>
---
 package/rauc/Config.in | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/package/rauc/Config.in b/package/rauc/Config.in
index 6433b827d3..864aa6b9c6 100644
--- a/package/rauc/Config.in
+++ b/package/rauc/Config.in
@@ -1,8 +1,9 @@
 config BR2_PACKAGE_RAUC
 	bool "rauc"
-	depends on BR2_TOOLCHAIN_HAS_THREADS # glib2
 	depends on BR2_USE_MMU # glib2
 	depends on BR2_USE_WCHAR # glib2
+	depends on BR2_TOOLCHAIN_HAS_THREADS # glib2
+	depends on BR2_TOOLCHAIN_HEADERS_AT_LEAST_3_0
 	select BR2_PACKAGE_LIBGLIB2
 	select BR2_PACKAGE_OPENSSL
 	select BR2_PACKAGE_DBUS # run-time dependency
@@ -32,6 +33,7 @@ config BR2_PACKAGE_RAUC_JSON
 
 endif
 
-comment "rauc needs a toolchain w/ wchar, threads"
-	depends on !BR2_USE_WCHAR || !BR2_TOOLCHAIN_HAS_THREADS
+comment "rauc needs a toolchain w/ wchar, threads, headers >= 3.0"
 	depends on BR2_USE_MMU
+	depends on !BR2_USE_WCHAR || !BR2_TOOLCHAIN_HAS_THREADS \
+		|| !BR2_TOOLCHAIN_HEADERS_AT_LEAST_3_0
-- 
2.14.1

[Buildroot] [PATCH] package/rauc: needs kernel headers >= 3.0

[Thomas Petazzoni]

Hello,

On Fri, 24 Aug 2018 10:02:17 +0200, Yann E. MORIN wrote:
> rauc includes <linux/mmc/ioctl.h> which was introduced only in
> linux 3.0.
> 
> Also change order to have arch dependencies first.
> 
> Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
> Cc: Andrey Yurovsky <yurovsky@gmail.com>

As discussed separately, this only solves the problem for the target
build, and not for the host build. So, I would prefer a solution that
adjusts the RAUC code to build with headers < 3.0, so that we solve
both the host and the target problem@the same time.

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

[Buildroot] [PATCH] package/rauc: needs kernel headers >= 3.0

[Yann E. MORIN]

Thomas, All,

On 2018-08-24 13:49 +0200, Thomas Petazzoni spake thusly:
> On Fri, 24 Aug 2018 10:02:17 +0200, Yann E. MORIN wrote:
> > rauc includes <linux/mmc/ioctl.h> which was introduced only in
> > linux 3.0.
> > 
> > Also change order to have arch dependencies first.
> > 
> > Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
> > Cc: Andrey Yurovsky <yurovsky@gmail.com>
> 
> As discussed separately, this only solves the problem for the target
> build, and not for the host build. So, I would prefer a solution that
> adjusts the RAUC code to build with headers < 3.0, so that we solve
> both the host and the target problem at the same time.

That is not really possible, because emmc support is core to rauc; it is
not even optional in the current code. Removing it (by making it
conditional) does not make sense IMHO (but I'm not using rauc).

And then, technically, it would require quite some overhaul in rauc,
because access to the emmc code is made from a lot of places.

So I'm not the one that will be doing that.

Regards,
Yann E. MORIN.

> Best regards,
> 
> Thomas
> -- 
> Thomas Petazzoni, CTO, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] [PATCH] package/rauc: needs kernel headers >= 3.0

[Thomas Petazzoni]

Hello,

On Fri, 24 Aug 2018 14:01:19 +0200, Yann E. MORIN wrote:

> That is not really possible, because emmc support is core to rauc; it is
> not even optional in the current code. Removing it (by making it
> conditional) does not make sense IMHO (but I'm not using rauc).
> 
> And then, technically, it would require quite some overhaul in rauc,
> because access to the emmc code is made from a lot of places.

Tadada. The problematic piece of code is not used for eMMC support in
general, just for eMMC boot partition support. It was added by a single
commit between v0.3 and v0.4 and is pretty self-contained and
well-isolated.

So I took what you said as a challenge, and came up with:

  http://patchwork.ozlabs.org/patch/961842/

Which has the added benefit of fixing the build issue of host-rauc :-)

Since it took me 36 minutes between your e-mail and this e-mail being,
I don't think "quite some overhaul in rauc" was really needed :-P

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

[Buildroot] [PATCH] package/rauc: needs kernel headers >= 3.0

[Yann E. MORIN]

Thomas, All,

On 2018-08-24 14:37 +0200, Thomas Petazzoni spake thusly:
> On Fri, 24 Aug 2018 14:01:19 +0200, Yann E. MORIN wrote:
> > That is not really possible, because emmc support is core to rauc; it is
> > not even optional in the current code. Removing it (by making it
> > conditional) does not make sense IMHO (but I'm not using rauc).
> > 
> > And then, technically, it would require quite some overhaul in rauc,
> > because access to the emmc code is made from a lot of places.
> 
> Tadada. The problematic piece of code is not used for eMMC support in
> general, just for eMMC boot partition support. It was added by a single
> commit between v0.3 and v0.4 and is pretty self-contained and
> well-isolated.
> 
> So I took what you said as a challenge, and came up with:
> 
>   http://patchwork.ozlabs.org/patch/961842/
> 
> Which has the added benefit of fixing the build issue of host-rauc :-)
> 
> Since it took me 36 minutes between your e-mail and this e-mail being,
> I don't think "quite some overhaul in rauc" was really needed :-P

Wooo... You took the bait quite easily! ;-)

But seriously, I indeed only skimmed for 'emmc' in the code. and there
are quite a few call sites. Now, does it make sense to disable this emmc
boot support at all? Let's see what upstream has to say about that. ;-)

Regards,
Yann E. MORIN.

> Best regards,
> 
> Thomas
> -- 
> Thomas Petazzoni, CTO, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'