[PATCH] genericx86-common.inc: Stop autoloading uvesa module at boot

[PATCH] genericx86-common.inc: Stop autoloading uvesa module at boot

[Alejandro Hernandez]

After commit e8b1c653946ef921b65d47e52aea0dc530ef4286, we started seeing
errors like the following during boot:

uvesafb: failed to execute /sbin/v86d
uvesafb: probe of uvesafb.0 failed with error -22
uvesafb: vbe_init() failed with -22
uvesafb: Getting VBE info block failed (eax=0x4f00, err=-2)

These were caused because the uvesa module was being loaded during boot.

Since genericx86-common.inc includes qemuboot-x86, the module also tries
to be loaded on genericx86 MACHINES, this patch removes it when this is the
case, getting rid of the errors.

[YOCTO #11879]

Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
---
 meta-yocto-bsp/conf/machine/include/genericx86-common.inc | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/meta-yocto-bsp/conf/machine/include/genericx86-common.inc b/meta-yocto-bsp/conf/machine/include/genericx86-common.inc
index e35685cb22b..1aabc4e8bad 100644
--- a/meta-yocto-bsp/conf/machine/include/genericx86-common.inc
+++ b/meta-yocto-bsp/conf/machine/include/genericx86-common.inc
@@ -15,6 +15,8 @@ XSERVER ?= "${XSERVER_X86_BASE} \
 
 MACHINE_EXTRA_RRECOMMENDS += "linux-firmware eee-acpi-scripts"
 
+KERNEL_MODULE_AUTOLOAD_remove = "uvesafb"
+
 GLIBC_ADDONS = "nptl"
 
 EXTRA_OECONF_append_pn-matchbox-panel-2 = " --with-battery=acpi"
@@ -23,3 +25,4 @@ IMAGE_FSTYPES += "wic wic.bmap"
 WKS_FILE ?= "genericx86.wks"
 do_image_wic[depends] += "gptfdisk-native:do_populate_sysroot"
 do_image_wic[recrdeptask] += "do_bootimg"
+
-- 
2.12.3

Re: [PATCH] genericx86-common.inc: Stop autoloading uvesa module at boot

[Richard Purdie]

On Wed, 2017-08-16 at 17:25 -0700, Alejandro Hernandez wrote:
> After commit e8b1c653946ef921b65d47e52aea0dc530ef4286, we started
> seeing
> errors like the following during boot:
> 
> uvesafb: failed to execute /sbin/v86d
> uvesafb: probe of uvesafb.0 failed with error -22
> uvesafb: vbe_init() failed with -22
> uvesafb: Getting VBE info block failed (eax=0x4f00, err=-2)
> 
> These were caused because the uvesa module was being loaded during
> boot.
> 
> Since genericx86-common.inc includes qemuboot-x86, the module also
> tries
> to be loaded on genericx86 MACHINES, this patch removes it when this
> is the
> case, getting rid of the errors.
> 
> [YOCTO #11879]
> 
> Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.c
> om>
> ---
>  meta-yocto-bsp/conf/machine/include/genericx86-common.inc | 3 +++
>  1 file changed, 3 insertions(+)

No, we can't do this, sorry. I really dislike having any usage of
remove in the core code. The reason is that its near impossible to undo
a remove operation and if we need one, it suggests we really should
restructure something instead.

So please find a different way to fix this.

This commit message also doesn't really tell me why we need to stop
loading uvesa on genericx86.

Cheers,

Richard

Re: [PATCH] genericx86-common.inc: Stop autoloading uvesa module at boot

[Alejandro Hernandez]

Hey Richard,


On 08/17/2017 02:42 AM, Richard Purdie wrote:
> On Wed, 2017-08-16 at 17:25 -0700, Alejandro Hernandez wrote:
>> After commit e8b1c653946ef921b65d47e52aea0dc530ef4286, we started
>> seeing
>> errors like the following during boot:
>>
>> uvesafb: failed to execute /sbin/v86d
>> uvesafb: probe of uvesafb.0 failed with error -22
>> uvesafb: vbe_init() failed with -22
>> uvesafb: Getting VBE info block failed (eax=0x4f00, err=-2)
>>
>> These were caused because the uvesa module was being loaded during
>> boot.
>>
>> Since genericx86-common.inc includes qemuboot-x86, the module also
>> tries
>> to be loaded on genericx86 MACHINES, this patch removes it when this
>> is the
>> case, getting rid of the errors.
>>
>> [YOCTO #11879]
>>
>> Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.c
>> om>
>> ---
>>   meta-yocto-bsp/conf/machine/include/genericx86-common.inc | 3 +++
>>   1 file changed, 3 insertions(+)
> No, we can't do this, sorry. I really dislike having any usage of
> remove in the core code. The reason is that its near impossible to undo
> a remove operation and if we need one, it suggests we really should
> restructure something instead.
>
> So please find a different way to fix this.
>
> This commit message also doesn't really tell me why we need to stop
> loading uvesa on genericx86.
Alright, I'll see if I can find another way, the reason being that it 
simply gets rid of the error,
I actually referenced the commit where we took uvesa out of generic but 
I deleted it last minute,
my bad.

http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=6af89812e8a9931ffed63768ed85367519bf7aef

>
> Cheers,
>
> Richard