I have uploaded another tarball with the current WMT branch. (linuxwmt-12-09-25.tar.bz2)
Given that the git fetch instructions and build instructions are in the wiki, I expect this will be the last tarball. Creating tarball's for users is, well, annoying and unnecessary.
I recommend you try getting the git source instead - if you can't, let me know and I'll try help. It makes updates much easier and means you can have the current source as soon as it's pushed.
Compiled the image as per the wiki. The image boots from microsd and
usb but no display.
The earlier tar ball you provided always produced a display, but
failed to mount rootfs. from
either microsd or usb.
<senti...@digitalfantasy.co.nz> wrote:
> I have uploaded another tarball with the current WMT branch.
> (linuxwmt-12-09-25.tar.bz2)
> Given that the git fetch instructions and build instructions are in the
> wiki, I expect this will be the last tarball.
> Creating tarball's for users is, well, annoying and unnecessary.
> I recommend you try getting the git source instead - if you can't, let me
> know and I'll try help.
> It makes updates much easier and means you can have the current source as
> soon as it's pushed.
On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
> Followed the wiki and ended with * wmt.
> Compiled the image as per the wiki. The image boots from microsd and
> usb but no display.
> The earlier tar ball you provided always produced a display, but
> failed to mount rootfs. from
> either microsd or usb.
> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
> <senti...@digitalfantasy.co.nz> wrote:
> > I have uploaded another tarball with the current WMT branch.
> > (linuxwmt-12-09-25.tar.bz2)
> > Given that the git fetch instructions and build instructions are in the
> > wiki, I expect this will be the last tarball.
> > Creating tarball's for users is, well, annoying and unnecessary.
> > I recommend you try getting the git source instead - if you can't, let me
> > know and I'll try help.
> > It makes updates much easier and means you can have the current source as
> > soon as it's pushed.
> > Regards
> > Tony P
Do you have a 7" WM8650-based tablet? I assume so since you say it
worked on the tarball.
If so, I can't see any reason why it shouldn't have a display.
I *think* the driver changed between the 12-09-01 tarball and the
current git, but most of my testing is done on a 7" WM8650 so it should
'just work'.
Do you see any framebuffer related output in dmesg via serial?
On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz> wrote:
> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
>> Followed the wiki and ended with * wmt.
>> Compiled the image as per the wiki. The image boots from microsd and
>> usb but no display.
>> The earlier tar ball you provided always produced a display, but
>> failed to mount rootfs. from
>> either microsd or usb.
>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
>> <senti...@digitalfantasy.co.nz> wrote:
>> > I have uploaded another tarball with the current WMT branch.
>> > (linuxwmt-12-09-25.tar.bz2)
>> > Given that the git fetch instructions and build instructions are in the
>> > wiki, I expect this will be the last tarball.
>> > Creating tarball's for users is, well, annoying and unnecessary.
>> > I recommend you try getting the git source instead - if you can't, let me
>> > know and I'll try help.
>> > It makes updates much easier and means you can have the current source as
>> > soon as it's pushed.
>> > Regards
>> > Tony P
> Do you have a 7" WM8650-based tablet?
yes
> I assume so since you say it worked on the tarball.
display worked, though sdmmc and usb did not.
> If so, I can't see any reason why it shouldn't have a display.
> I *think* the driver changed between the 12-09-01 tarball and the
> current git, but most of my testing is done on a 7" WM8650 so it should
> 'just work'.
> Do you see any framebuffer related output in dmesg via serial?
> On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz> wrote:
>> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
>>> Followed the wiki and ended with * wmt.
>>> Compiled the image as per the wiki. The image boots from microsd and
>>> usb but no display.
>>> The earlier tar ball you provided always produced a display, but
>>> failed to mount rootfs. from
>>> either microsd or usb.
>>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
>>> <senti...@digitalfantasy.co.nz> wrote:
>>> > I have uploaded another tarball with the current WMT branch.
>>> > (linuxwmt-12-09-25.tar.bz2)
>>> > Given that the git fetch instructions and build instructions are in the
>>> > wiki, I expect this will be the last tarball.
>>> > Creating tarball's for users is, well, annoying and unnecessary.
>>> > I recommend you try getting the git source instead - if you can't, let me
>>> > know and I'll try help.
>>> > It makes updates much easier and means you can have the current source as
>>> > soon as it's pushed.
>>> > Regards
>>> > Tony P
>> Do you have a 7" WM8650-based tablet?
> yes
>> I assume so since you say it worked on the tarball.
> display worked, though sdmmc and usb did not.
>> If so, I can't see any reason why it shouldn't have a display.
>> I *think* the driver changed between the 12-09-01 tarball and the
>> current git, but most of my testing is done on a 7" WM8650 so it should
>> 'just work'.
>> Do you see any framebuffer related output in dmesg via serial?
> On Tue, Sep 25, 2012 at 10:36 AM, J T Dsouza <jtd1...@gmail.com> wrote:
> > On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz>
> wrote:
> >> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
> >>> Followed the wiki and ended with * wmt.
> >>> Compiled the image as per the wiki. The image boots from microsd and
> >>> usb but no display.
> >>> The earlier tar ball you provided always produced a display, but
> >>> failed to mount rootfs. from
> >>> either microsd or usb.
> >>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
> >>> <senti...@digitalfantasy.co.nz> wrote:
> >>> > I have uploaded another tarball with the current WMT branch.
> >>> > (linuxwmt-12-09-25.tar.bz2)
> >>> > Given that the git fetch instructions and build instructions are in
> the
> >>> > wiki, I expect this will be the last tarball.
> >>> > Creating tarball's for users is, well, annoying and unnecessary.
> >>> > I recommend you try getting the git source instead - if you can't,
> let me
> >>> > know and I'll try help.
> >>> > It makes updates much easier and means you can have the current
> source as
> >>> > soon as it's pushed.
> >>> > Regards
> >>> > Tony P
> >> Do you have a 7" WM8650-based tablet?
> > yes
> >> I assume so since you say it worked on the tarball.
> > display worked, though sdmmc and usb did not.
> >> If so, I can't see any reason why it shouldn't have a display.
> >> I *think* the driver changed between the 12-09-01 tarball and the
> >> current git, but most of my testing is done on a 7" WM8650 so it should
> >> 'just work'.
> >> Do you see any framebuffer related output in dmesg via serial?
>> On Tue, Sep 25, 2012 at 10:36 AM, J T Dsouza <jtd1...@gmail.com> wrote:
>> > On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz>
>> > wrote:
>> >> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
>> >>> Followed the wiki and ended with * wmt.
>> >>> Compiled the image as per the wiki. The image boots from microsd and
>> >>> usb but no display.
>> >>> The earlier tar ball you provided always produced a display, but
>> >>> failed to mount rootfs. from
>> >>> either microsd or usb.
>> >>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
>> >>> <senti...@digitalfantasy.co.nz> wrote:
>> >>> > I have uploaded another tarball with the current WMT branch.
>> >>> > (linuxwmt-12-09-25.tar.bz2)
>> >>> > Given that the git fetch instructions and build instructions are in
>> >>> > the
>> >>> > wiki, I expect this will be the last tarball.
>> >>> > Creating tarball's for users is, well, annoying and unnecessary.
>> >>> > I recommend you try getting the git source instead - if you can't,
>> >>> > let me
>> >>> > know and I'll try help.
>> >>> > It makes updates much easier and means you can have the current
>> >>> > source as
>> >>> > soon as it's pushed.
>> >>> > Regards
>> >>> > Tony P
>> >> Do you have a 7" WM8650-based tablet?
>> > yes
>> >> I assume so since you say it worked on the tarball.
>> > display worked, though sdmmc and usb did not.
>> >> If so, I can't see any reason why it shouldn't have a display.
>> >> I *think* the driver changed between the 12-09-01 tarball and the
>> >> current git, but most of my testing is done on a 7" WM8650 so it should
>> >> 'just work'.
>> >> Do you see any framebuffer related output in dmesg via serial?
>>> On Tue, Sep 25, 2012 at 10:36 AM, J T Dsouza <jtd1...@gmail.com> wrote:
>>> > On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz>
>>> > wrote:
>>> >> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
>>> >>> Followed the wiki and ended with * wmt.
>>> >>> Compiled the image as per the wiki. The image boots from microsd and
>>> >>> usb but no display.
>>> >>> The earlier tar ball you provided always produced a display, but
>>> >>> failed to mount rootfs. from
>>> >>> either microsd or usb.
>>> >>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
>>> >>> <senti...@digitalfantasy.co.nz> wrote:
>>> >>> > I have uploaded another tarball with the current WMT branch.
>>> >>> > (linuxwmt-12-09-25.tar.bz2)
>>> >>> > Given that the git fetch instructions and build instructions are in
>>> >>> > the
>>> >>> > wiki, I expect this will be the last tarball.
>>> >>> > Creating tarball's for users is, well, annoying and unnecessary.
>>> >>> > I recommend you try getting the git source instead - if you can't,
>>> >>> > let me
>>> >>> > know and I'll try help.
>>> >>> > It makes updates much easier and means you can have the current
>>> >>> > source as
>>> >>> > soon as it's pushed.
>>> >>> > Regards
>>> >>> > Tony P
>>> >> Do you have a 7" WM8650-based tablet?
>>> > yes
>>> >> I assume so since you say it worked on the tarball.
>>> > display worked, though sdmmc and usb did not.
>>> >> If so, I can't see any reason why it shouldn't have a display.
>>> >> I *think* the driver changed between the 12-09-01 tarball and the
>>> >> current git, but most of my testing is done on a 7" WM8650 so it should
>>> >> 'just work'.
>>> >> Do you see any framebuffer related output in dmesg via serial?
On Tue, 2012-09-25 at 11:32 +0530, J T Dsouza wrote:
> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
> > Any difference if you remove mem= argument?
> None.
> > On Sep 25, 2012 9:09 AM, "J T Dsouza" <jtd1...@gmail.com> wrote:
> >> On Tue, Sep 25, 2012 at 10:36 AM, J T Dsouza <jtd1...@gmail.com> wrote:
> >> > On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz>
> >> > wrote:
> >> >> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
> >> >>> Followed the wiki and ended with * wmt.
> >> >>> Compiled the image as per the wiki. The image boots from microsd and
> >> >>> usb but no display.
> >> >>> The earlier tar ball you provided always produced a display, but
> >> >>> failed to mount rootfs. from
> >> >>> either microsd or usb.
> >> >>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
> >> >>> <senti...@digitalfantasy.co.nz> wrote:
> >> >>> > I have uploaded another tarball with the current WMT branch.
> >> >>> > (linuxwmt-12-09-25.tar.bz2)
> >> >>> > Given that the git fetch instructions and build instructions are in
> >> >>> > the
> >> >>> > wiki, I expect this will be the last tarball.
> >> >>> > Creating tarball's for users is, well, annoying and unnecessary.
> >> >>> > I recommend you try getting the git source instead - if you can't,
> >> >>> > let me
> >> >>> > know and I'll try help.
> >> >>> > It makes updates much easier and means you can have the current
> >> >>> > source as
> >> >>> > soon as it's pushed.
> >> >>> > Regards
> >> >>> > Tony P
> >> >> Do you have a 7" WM8650-based tablet?
> >> > yes
> >> >> I assume so since you say it worked on the tarball.
> >> > display worked, though sdmmc and usb did not.
> >> >> If so, I can't see any reason why it shouldn't have a display.
> >> >> I *think* the driver changed between the 12-09-01 tarball and the
> >> >> current git, but most of my testing is done on a 7" WM8650 so it should
> >> >> 'just work'.
> >> >> Do you see any framebuffer related output in dmesg via serial?
On Tue, Sep 25, 2012 at 11:43 AM, Alexey Charkov <alch...@gmail.com> wrote:
> Can you share your .config then?
> Best,
> Alexey
> 2012/9/25 J T Dsouza <jtd1...@gmail.com>:
>> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
>>> Any difference if you remove mem= argument?
>> None.
>>> On Sep 25, 2012 9:09 AM, "J T Dsouza" <jtd1...@gmail.com> wrote:
>>>> On Tue, Sep 25, 2012 at 10:36 AM, J T Dsouza <jtd1...@gmail.com> wrote:
>>>> > On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz>
>>>> > wrote:
>>>> >> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
>>>> >>> Followed the wiki and ended with * wmt.
>>>> >>> Compiled the image as per the wiki. The image boots from microsd and
>>>> >>> usb but no display.
>>>> >>> The earlier tar ball you provided always produced a display, but
>>>> >>> failed to mount rootfs. from
>>>> >>> either microsd or usb.
>>>> >>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
>>>> >>> <senti...@digitalfantasy.co.nz> wrote:
>>>> >>> > I have uploaded another tarball with the current WMT branch.
>>>> >>> > (linuxwmt-12-09-25.tar.bz2)
>>>> >>> > Given that the git fetch instructions and build instructions are in
>>>> >>> > the
>>>> >>> > wiki, I expect this will be the last tarball.
>>>> >>> > Creating tarball's for users is, well, annoying and unnecessary.
>>>> >>> > I recommend you try getting the git source instead - if you can't,
>>>> >>> > let me
>>>> >>> > know and I'll try help.
>>>> >>> > It makes updates much easier and means you can have the current
>>>> >>> > source as
>>>> >>> > soon as it's pushed.
>>>> >>> > Regards
>>>> >>> > Tony P
>>>> >> Do you have a 7" WM8650-based tablet?
>>>> > yes
>>>> >> I assume so since you say it worked on the tarball.
>>>> > display worked, though sdmmc and usb did not.
>>>> >> If so, I can't see any reason why it shouldn't have a display.
>>>> >> I *think* the driver changed between the 12-09-01 tarball and the
>>>> >> current git, but most of my testing is done on a 7" WM8650 so it should
>>>> >> 'just work'.
>>>> >> Do you see any framebuffer related output in dmesg via serial?
On Tue, 2012-09-25 at 21:47 +0530, J T Dsouza wrote:
> attached
> On Tue, Sep 25, 2012 at 11:43 AM, Alexey Charkov <alch...@gmail.com> wrote:
> > Can you share your .config then?
> > Best,
> > Alexey
> > 2012/9/25 J T Dsouza <jtd1...@gmail.com>:
> >> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
> >>> Any difference if you remove mem= argument?
> >> None.
> >>> On Sep 25, 2012 9:09 AM, "J T Dsouza" <jtd1...@gmail.com> wrote:
> >>>> On Tue, Sep 25, 2012 at 10:36 AM, J T Dsouza <jtd1...@gmail.com> wrote:
> >>>> > On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz>
> >>>> > wrote:
> >>>> >> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
> >>>> >>> Followed the wiki and ended with * wmt.
> >>>> >>> Compiled the image as per the wiki. The image boots from microsd and
> >>>> >>> usb but no display.
> >>>> >>> The earlier tar ball you provided always produced a display, but
> >>>> >>> failed to mount rootfs. from
> >>>> >>> either microsd or usb.
> >>>> >>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
> >>>> >>> <senti...@digitalfantasy.co.nz> wrote:
> >>>> >>> > I have uploaded another tarball with the current WMT branch.
> >>>> >>> > (linuxwmt-12-09-25.tar.bz2)
> >>>> >>> > Given that the git fetch instructions and build instructions are in
> >>>> >>> > the
> >>>> >>> > wiki, I expect this will be the last tarball.
> >>>> >>> > Creating tarball's for users is, well, annoying and unnecessary.
> >>>> >>> > I recommend you try getting the git source instead - if you can't,
> >>>> >>> > let me
> >>>> >>> > know and I'll try help.
> >>>> >>> > It makes updates much easier and means you can have the current
> >>>> >>> > source as
> >>>> >>> > soon as it's pushed.
> >>>> >>> > Regards
> >>>> >>> > Tony P
> >>>> >> Do you have a 7" WM8650-based tablet?
> >>>> > yes
> >>>> >> I assume so since you say it worked on the tarball.
> >>>> > display worked, though sdmmc and usb did not.
> >>>> >> If so, I can't see any reason why it shouldn't have a display.
> >>>> >> I *think* the driver changed between the 12-09-01 tarball and the
> >>>> >> current git, but most of my testing is done on a 7" WM8650 so it should
> >>>> >> 'just work'.
> >>>> >> Do you see any framebuffer related output in dmesg via serial?
On Tue, Sep 25, 2012 at 10:41 PM, Tony Prisk <li...@prisktech.co.nz> wrote:
> On Tue, 2012-09-25 at 21:47 +0530, J T Dsouza wrote:
>> attached
>> On Tue, Sep 25, 2012 at 11:43 AM, Alexey Charkov <alch...@gmail.com> wrote:
>> > Can you share your .config then?
>> > Best,
>> > Alexey
>> > 2012/9/25 J T Dsouza <jtd1...@gmail.com>:
>> >> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
>> >>> Any difference if you remove mem= argument?
>> >> None.
>> >>> On Sep 25, 2012 9:09 AM, "J T Dsouza" <jtd1...@gmail.com> wrote:
>> >>>> On Tue, Sep 25, 2012 at 10:36 AM, J T Dsouza <jtd1...@gmail.com> wrote:
>> >>>> > On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz>
>> >>>> > wrote:
>> >>>> >> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
>> >>>> >>> Followed the wiki and ended with * wmt.
>> >>>> >>> Compiled the image as per the wiki. The image boots from microsd and
>> >>>> >>> usb but no display.
>> >>>> >>> The earlier tar ball you provided always produced a display, but
>> >>>> >>> failed to mount rootfs. from
>> >>>> >>> either microsd or usb.
>> >>>> >>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
>> >>>> >>> <senti...@digitalfantasy.co.nz> wrote:
>> >>>> >>> > I have uploaded another tarball with the current WMT branch.
>> >>>> >>> > (linuxwmt-12-09-25.tar.bz2)
>> >>>> >>> > Given that the git fetch instructions and build instructions are in
>> >>>> >>> > the
>> >>>> >>> > wiki, I expect this will be the last tarball.
>> >>>> >>> > Creating tarball's for users is, well, annoying and unnecessary.
>> >>>> >>> > I recommend you try getting the git source instead - if you can't,
>> >>>> >>> > let me
>> >>>> >>> > know and I'll try help.
>> >>>> >>> > It makes updates much easier and means you can have the current
>> >>>> >>> > source as
>> >>>> >>> > soon as it's pushed.
>> >>>> >>> > Regards
>> >>>> >>> > Tony P
>> >>>> >> Do you have a 7" WM8650-based tablet?
>> >>>> > yes
>> >>>> >> I assume so since you say it worked on the tarball.
>> >>>> > display worked, though sdmmc and usb did not.
>> >>>> >> If so, I can't see any reason why it shouldn't have a display.
>> >>>> >> I *think* the driver changed between the 12-09-01 tarball and the
>> >>>> >> current git, but most of my testing is done on a 7" WM8650 so it should
>> >>>> >> 'just work'.
>> >>>> >> Do you see any framebuffer related output in dmesg via serial?
> >> >>>> just incase i am missing something here.
> >> >>>> On Tue, Sep 25, 2012 at 10:36 AM, J T Dsouza <jtd1...@gmail.com> wrote:
> >> >>>> > On Tue, Sep 25, 2012 at 9:18 AM, Tony Prisk <li...@prisktech.co.nz>
> >> >>>> > wrote:
> >> >>>> >> On Tue, 2012-09-25 at 03:49 +0530, J T Dsouza wrote:
> >> >>>> >>> Followed the wiki and ended with * wmt.
> >> >>>> >>> Compiled the image as per the wiki. The image boots from microsd and
> >> >>>> >>> usb but no display.
> >> >>>> >>> The earlier tar ball you provided always produced a display, but
> >> >>>> >>> failed to mount rootfs. from
> >> >>>> >>> either microsd or usb.
> >> >>>> >>> On Tue, Sep 25, 2012 at 2:07 AM, Tony Prisk
> >> >>>> >>> <senti...@digitalfantasy.co.nz> wrote:
> >> >>>> >>> > I have uploaded another tarball with the current WMT branch.
> >> >>>> >>> > (linuxwmt-12-09-25.tar.bz2)
> >> >>>> >>> > Given that the git fetch instructions and build instructions are in
> >> >>>> >>> > the
> >> >>>> >>> > wiki, I expect this will be the last tarball.
> >> >>>> >>> > Creating tarball's for users is, well, annoying and unnecessary.
> >> >>>> >>> > I recommend you try getting the git source instead - if you can't,
> >> >>>> >>> > let me
> >> >>>> >>> > know and I'll try help.
> >> >>>> >>> > It makes updates much easier and means you can have the current
> >> >>>> >>> > source as
> >> >>>> >>> > soon as it's pushed.
> >> >>>> >>> > Regards
> >> >>>> >>> > Tony P
> >> >>>> >> Do you have a 7" WM8650-based tablet?
> >> >>>> > yes
> >> >>>> >> I assume so since you say it worked on the tarball.
> >> >>>> > display worked, though sdmmc and usb did not.
> >> >>>> >> If so, I can't see any reason why it shouldn't have a display.
> >> >>>> >> I *think* the driver changed between the 12-09-01 tarball and the
> >> >>>> >> current git, but most of my testing is done on a 7" WM8650 so it should
> >> >>>> >> 'just work'.
> >> >>>> >> Do you see any framebuffer related output in dmesg via serial?
On Tue, Sep 25, 2012 at 1:05 PM, Tony Prisk <li...@prisktech.co.nz> wrote:
> On Tue, 2012-09-25 at 11:32 +0530, J T Dsouza wrote:
>> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
>> > Any difference if you remove mem= argument?
> Two things stand out
> 1) You need to increase the coherent memory allocation - you are only
> reserving 256KB for coherent allocations. Increase it to at least 2MB
> for the framebuffer - more doesn't really hurt, it's still available for
> normal allocations - I think mine is on 16MB. This should give an error
> message if its a problem.
Is CMA applicable to ARMV6+.and above? If so will I have to use
something else instead of -march=armv4t -mtune=arm920t
> 2) The graphics driver isn't showing any messages at all. Either the
> devicetree isn't configured correctly (I would hope it is) or the driver
> isn't selected in .config.
> One thing that pop's to mind is to check the devicetree
> (arch/arm/boot/dts/wm8650.dtsi) compatible string = 'wm,wm8650-fb'
> On Tue, Sep 25, 2012 at 1:05 PM, Tony Prisk <li...@prisktech.co.nz> wrote:
>> On Tue, 2012-09-25 at 11:32 +0530, J T Dsouza wrote:
>>> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
>>> > Any difference if you remove mem= argument?
>> Two things stand out
>> 1) You need to increase the coherent memory allocation - you are only
>> reserving 256KB for coherent allocations. Increase it to at least 2MB
>> for the framebuffer - more doesn't really hurt, it's still available for
>> normal allocations - I think mine is on 16MB. This should give an error
>> message if its a problem.
> Is CMA applicable to ARMV6+.and above? If so will I have to use
> something else instead of -march=armv4t -mtune=arm920t
> What should I change in .config
I think that this dependency has been relaxed recently to also cover
ARMv5. Isn't it just a matter of appending coherent_pool=[size] to the
kernel command line?
On Thu, Sep 27, 2012 at 4:37 PM, Alexey Charkov <alch...@gmail.com> wrote:
> 2012/9/27 J T Dsouza <jtd1...@gmail.com>:
>> On Tue, Sep 25, 2012 at 1:05 PM, Tony Prisk <li...@prisktech.co.nz> wrote:
>>> On Tue, 2012-09-25 at 11:32 +0530, J T Dsouza wrote:
>>>> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
>>>> > Any difference if you remove mem= argument?
>>> Two things stand out
>>> 1) You need to increase the coherent memory allocation - you are only
>>> reserving 256KB for coherent allocations. Increase it to at least 2MB
>>> for the framebuffer - more doesn't really hurt, it's still available for
>>> normal allocations - I think mine is on 16MB. This should give an error
>>> message if its a problem.
>> Is CMA applicable to ARMV6+.and above? If so will I have to use
>> something else instead of -march=armv4t -mtune=arm920t
>> What should I change in .config
> I think that this dependency has been relaxed recently to also cover
> ARMv5. Isn't it just a matter of appending coherent_pool=[size] to the
> kernel command line?
The downloaded tarball compiled and boots from mmc and gives me a
login, but not from usb. Usb plug events are not discovered. USB is
compiled into kernel. So no keyboard.
have to check what driver will work in place of rt3070sta. Afaik it is
replaced by rt5070.
Looks like git errs.
I will use the new .config on the sources from git and verify.
On Thu, Sep 27, 2012 at 8:12 PM, J T Dsouza <jtd1...@gmail.com> wrote:
> Downloading the new tarball. will know if it's just me being stupid or
> git playing truant.
> On Thu, Sep 27, 2012 at 4:37 PM, Alexey Charkov <alch...@gmail.com> wrote:
>> 2012/9/27 J T Dsouza <jtd1...@gmail.com>:
>>> On Tue, Sep 25, 2012 at 1:05 PM, Tony Prisk <li...@prisktech.co.nz> wrote:
>>>> On Tue, 2012-09-25 at 11:32 +0530, J T Dsouza wrote:
>>>>> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
>>>>> > Any difference if you remove mem= argument?
>>>> Two things stand out
>>>> 1) You need to increase the coherent memory allocation - you are only
>>>> reserving 256KB for coherent allocations. Increase it to at least 2MB
>>>> for the framebuffer - more doesn't really hurt, it's still available for
>>>> normal allocations - I think mine is on 16MB. This should give an error
>>>> message if its a problem.
>>> Is CMA applicable to ARMV6+.and above? If so will I have to use
>>> something else instead of -march=armv4t -mtune=arm920t
>>> What should I change in .config
>> I think that this dependency has been relaxed recently to also cover
>> ARMv5. Isn't it just a matter of appending coherent_pool=[size] to the
>> kernel command line?
On Fri, 2012-09-28 at 01:46 +0530, J T Dsouza wrote:
> The downloaded tarball compiled and boots from mmc and gives me a
> login, but not from usb. Usb plug events are not discovered. USB is
> compiled into kernel. So no keyboard.
> have to check what driver will work in place of rt3070sta. Afaik it is
> replaced by rt5070.
> Looks like git errs.
> I will use the new .config on the sources from git and verify.
> On Thu, Sep 27, 2012 at 8:12 PM, J T Dsouza <jtd1...@gmail.com> wrote:
> > Attached the new .config. No console or screen.
> > Downloading the new tarball. will know if it's just me being stupid or
> > git playing truant.
> > On Thu, Sep 27, 2012 at 4:37 PM, Alexey Charkov <alch...@gmail.com> wrote:
> >> 2012/9/27 J T Dsouza <jtd1...@gmail.com>:
> >>> On Tue, Sep 25, 2012 at 1:05 PM, Tony Prisk <li...@prisktech.co.nz> wrote:
> >>>> On Tue, 2012-09-25 at 11:32 +0530, J T Dsouza wrote:
> >>>>> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
> >>>>> > Any difference if you remove mem= argument?
> >>>> Two things stand out
> >>>> 1) You need to increase the coherent memory allocation - you are only
> >>>> reserving 256KB for coherent allocations. Increase it to at least 2MB
> >>>> for the framebuffer - more doesn't really hurt, it's still available for
> >>>> normal allocations - I think mine is on 16MB. This should give an error
> >>>> message if its a problem.
> >>> Is CMA applicable to ARMV6+.and above? If so will I have to use
> >>> something else instead of -march=armv4t -mtune=arm920t
CMA is for coherent-memory-allocations and is required for allocating
the framebuffer (rather than reserving it early). You can set the
coherent reservation size @:
Device Drivers->Generic Driver Options->Contiguous Memory Allocator [*]
Device Drivers->Generic Driver Options->Size in Mega bytes [16]
> >> I think that this dependency has been relaxed recently to also cover
> >> ARMv5. Isn't it just a matter of appending coherent_pool=[size] to the
> >> kernel command line?
> On Fri, 2012-09-28 at 01:46 +0530, J T Dsouza wrote:
>> The downloaded tarball compiled and boots from mmc and gives me a
>> login, but not from usb. Usb plug events are not discovered. USB is
>> compiled into kernel. So no keyboard.
>> have to check what driver will work in place of rt3070sta. Afaik it is
>> replaced by rt5070.
>> Looks like git errs.
>> I will use the new .config on the sources from git and verify.
>> On Thu, Sep 27, 2012 at 8:12 PM, J T Dsouza <jtd1...@gmail.com> wrote:
>> > Attached the new .config. No console or screen.
>> > Downloading the new tarball. will know if it's just me being stupid or
>> > git playing truant.
>> > On Thu, Sep 27, 2012 at 4:37 PM, Alexey Charkov <alch...@gmail.com> wrote:
>> >> 2012/9/27 J T Dsouza <jtd1...@gmail.com>:
>> >>> On Tue, Sep 25, 2012 at 1:05 PM, Tony Prisk <li...@prisktech.co.nz> wrote:
>> >>>> On Tue, 2012-09-25 at 11:32 +0530, J T Dsouza wrote:
>> >>>>> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
>> >>>>> > Any difference if you remove mem= argument?
>> >>>> Two things stand out
>> >>>> 1) You need to increase the coherent memory allocation - you are only
>> >>>> reserving 256KB for coherent allocations. Increase it to at least 2MB
>> >>>> for the framebuffer - more doesn't really hurt, it's still available for
>> >>>> normal allocations - I think mine is on 16MB. This should give an error
>> >>>> message if its a problem.
>> >>> Is CMA applicable to ARMV6+.and above? If so will I have to use
>> >>> something else instead of -march=armv4t -mtune=arm920t
> CMA is for coherent-memory-allocations and is required for allocating
> the framebuffer (rather than reserving it early). You can set the
> coherent reservation size @:
> Device Drivers->Generic Driver Options->Contiguous Memory Allocator [*]
> Device Drivers->Generic Driver Options->Size in Mega bytes [16]
Would it make sense to add some sort of Kconfig dependency on these
for the framebuffer driver, or at least make it print explicit
suggestions upon allocation failure?
On Fri, 2012-09-28 at 11:59 +0400, Alexey Charkov wrote:
> 2012/9/28 Tony Prisk <senti...@digitalfantasy.co.nz>:
> > On Fri, 2012-09-28 at 01:46 +0530, J T Dsouza wrote:
> >> The downloaded tarball compiled and boots from mmc and gives me a
> >> login, but not from usb. Usb plug events are not discovered. USB is
> >> compiled into kernel. So no keyboard.
> >> have to check what driver will work in place of rt3070sta. Afaik it is
> >> replaced by rt5070.
> >> Looks like git errs.
> >> I will use the new .config on the sources from git and verify.
> >> On Thu, Sep 27, 2012 at 8:12 PM, J T Dsouza <jtd1...@gmail.com> wrote:
> >> > Attached the new .config. No console or screen.
> >> > Downloading the new tarball. will know if it's just me being stupid or
> >> > git playing truant.
> >> > On Thu, Sep 27, 2012 at 4:37 PM, Alexey Charkov <alch...@gmail.com> wrote:
> >> >> 2012/9/27 J T Dsouza <jtd1...@gmail.com>:
> >> >>> On Tue, Sep 25, 2012 at 1:05 PM, Tony Prisk <li...@prisktech.co.nz> wrote:
> >> >>>> On Tue, 2012-09-25 at 11:32 +0530, J T Dsouza wrote:
> >> >>>>> On Tue, Sep 25, 2012 at 11:05 AM, Alexey Charkov <alch...@gmail.com> wrote:
> >> >>>>> > Any difference if you remove mem= argument?
> >> >>>> Two things stand out
> >> >>>> 1) You need to increase the coherent memory allocation - you are only
> >> >>>> reserving 256KB for coherent allocations. Increase it to at least 2MB
> >> >>>> for the framebuffer - more doesn't really hurt, it's still available for
> >> >>>> normal allocations - I think mine is on 16MB. This should give an error
> >> >>>> message if its a problem.
> >> >>> Is CMA applicable to ARMV6+.and above? If so will I have to use
> >> >>> something else instead of -march=armv4t -mtune=arm920t
> > CMA is for coherent-memory-allocations and is required for allocating
> > the framebuffer (rather than reserving it early). You can set the
> > coherent reservation size @:
> > Device Drivers->Generic Driver Options->Contiguous Memory Allocator [*]
> > Device Drivers->Generic Driver Options->Size in Mega bytes [16]
> Would it make sense to add some sort of Kconfig dependency on these
> for the framebuffer driver, or at least make it print explicit
> suggestions upon allocation failure?
> Best,
> Alexey
Its weird, but quite often I can have CMA disabled and it will still
work :) Other times, it gives a -NOMEM fault. I guess it just depends on
which options are enabled and where the memory gets allocated from.
I need to patch the framebuffer code for 3.8, so I'll add a more
explicit message for framebuffer memory allocation failures.
Also I have pulled last repo state as it says on Wiki, but got everything different (nothing about "GE acceleration", only "WM8505 fb support" and "VT8500 LCD driver" under "Support for fb devices"). Fact: display doesn't work for me using the SF's git repo, only with gitorious.
On Monday, 24 September 2012 17:37:32 UTC-3, Tony Prisk wrote:
> I have uploaded another tarball with the current WMT branch. > (linuxwmt-12-09-25.tar.bz2)
> Given that the git fetch instructions and build instructions are in the > wiki, I expect this will be the last tarball. > Creating tarball's for users is, well, annoying and unnecessary.
> I recommend you try getting the git source instead - if you can't, let me > know and I'll try help. > It makes updates much easier and means you can have the current source as > soon as it's pushed.
On Wed, 2012-10-03 at 21:03 -0700, Marcos Silva Cunha wrote:
> Also I have pulled last repo state as it says on Wiki, but got
> everything different (nothing about "GE acceleration", only "WM8505 fb
> support" and "VT8500 LCD driver" under "Support for fb devices").
> Fact: display doesn't work for me using the SF's git repo, only with
> gitorious.
> On Monday, 24 September 2012 17:37:32 UTC-3, Tony Prisk wrote:
> I have uploaded another tarball with the current WMT branch.
> (linuxwmt-12-09-25.tar.bz2)
> Given that the git fetch instructions and build instructions
> are in the wiki, I expect this will be the last tarball.
> Creating tarball's for users is, well, annoying and
> unnecessary.
> I recommend you try getting the git source instead - if you
> can't, let me know and I'll try help.
> It makes updates much easier and means you can have the
> current source as soon as it's pushed.
> Regards
> Tony P
When did you pull it?
If it was in the last day or two, then...
I have updated the config instructions on the wiki to reflect the fact
that I pulled past v3.6 in mainline and so it includes the new device
tree files from mainline (the latest patchset I sent to -next).
Mainline doesn't include the new WM8505 driver that *was* in the wmt
branch. As a result (and as mentioned in the git log), the new driver
was dropped and replaced with the mainline version (the old one) and the
config instructions were changed accordingly.
So yes, you are correct - only vt8500 and wm8505 framebuffers are now
available - no seperate GE acceleration. The wiki reflects this.
If you join the linuxwmt project on sourceforge, you would recieve
notifications of changes (to the downloads and wiki at least - I don't
think git updates result in email notifications).
As for it not working... I'll pull the sourceforge code down into a
clean directory and test it on my wm8650.
On Thu, 2012-10-04 at 18:53 +1300, Tony Prisk wrote:
> On Wed, 2012-10-03 at 21:03 -0700, Marcos Silva Cunha wrote:
> > Also I have pulled last repo state as it says on Wiki, but got
> > everything different (nothing about "GE acceleration", only "WM8505 fb
> > support" and "VT8500 LCD driver" under "Support for fb devices").
> > Fact: display doesn't work for me using the SF's git repo, only with
> > gitorious.
> > On Monday, 24 September 2012 17:37:32 UTC-3, Tony Prisk wrote:
> > I have uploaded another tarball with the current WMT branch.
> > (linuxwmt-12-09-25.tar.bz2)
> > Given that the git fetch instructions and build instructions
> > are in the wiki, I expect this will be the last tarball.
> > Creating tarball's for users is, well, annoying and
> > unnecessary.
> > I recommend you try getting the git source instead - if you
> > can't, let me know and I'll try help.
> > It makes updates much easier and means you can have the
> > current source as soon as it's pushed.
> > Regards
> > Tony P
> When did you pull it?
> If it was in the last day or two, then...
> I have updated the config instructions on the wiki to reflect the fact
> that I pulled past v3.6 in mainline and so it includes the new device
> tree files from mainline (the latest patchset I sent to -next).
> Mainline doesn't include the new WM8505 driver that *was* in the wmt
> branch. As a result (and as mentioned in the git log), the new driver
> was dropped and replaced with the mainline version (the old one) and the
> config instructions were changed accordingly.
> So yes, you are correct - only vt8500 and wm8505 framebuffers are now
> available - no seperate GE acceleration. The wiki reflects this.
> If you join the linuxwmt project on sourceforge, you would recieve
> notifications of changes (to the downloads and wiki at least - I don't
> think git updates result in email notifications).
> As for it not working... I'll pull the sourceforge code down into a
> clean directory and test it on my wm8650.
> Regards
> Tony P
I pulled the source down into a clean directory, and compiled.
I found two problems:
1) There was a unresoved merge conflict in wm8650.dtsi which prevented
it from compiling. I have pushed a fix for this.
2) The UHCI driver has been modified between me submitting it and it
being applied to mainline which results in it not working with attached
devices. I have pushed a fix to WMT.
Hi Tony, thanks for you attention. I pulled and recompiled my kernel but still no video. I'm attaching my config. These are the commands I run to prepare the MMC:
mkimage -A arm -O linux -T script -C gzip -a 1 -e 0 -n "Script" -d wmt_script.source wmt_scriptcmd make mrproper make distclean make ARCH=arm menuconfig make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- uImage -j5 cp arch/arm/boot/uImage /media/MMC2GB_P1/ umount /dev/sdc*
On Monday, 24 September 2012 17:37:32 UTC-3, Tony Prisk wrote:
> I have uploaded another tarball with the current WMT branch. > (linuxwmt-12-09-25.tar.bz2)
> Given that the git fetch instructions and build instructions are in the > wiki, I expect this will be the last tarball. > Creating tarball's for users is, well, annoying and unnecessary.
> I recommend you try getting the git source instead - if you can't, let me > know and I'll try help. > It makes updates much easier and means you can have the current source as > soon as it's pushed.
On Thu, 2012-10-04 at 07:10 -0700, Marcos Silva Cunha wrote:
> Hi Tony, thanks for you attention. I pulled and recompiled my kernel
> but still no video. I'm attaching my config. These are the commands I
> run to prepare the MMC:
> mkimage -A arm -O linux -T script -C gzip -a 1 -e 0 -n "Script" -d
> wmt_script.source wmt_scriptcmd > make mrproper
> make distclean
> make ARCH=arm menuconfig
> make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- uImage -j5
> cp arch/arm/boot/uImage /media/MMC2GB_P1/
> umount /dev/sdc*
> Regards,
> On Monday, 24 September 2012 17:37:32 UTC-3, Tony Prisk wrote:
> I have uploaded another tarball with the current WMT branch.
> (linuxwmt-12-09-25.tar.bz2)
> Given that the git fetch instructions and build instructions
> are in the wiki, I expect this will be the last tarball.
> Creating tarball's for users is, well, annoying and
> unnecessary.
> I recommend you try getting the git source instead - if you
> can't, let me know and I'll try help.
> It makes updates much easier and means you can have the
> current source as soon as it's pushed.
> Regards
> Tony P
Straight away I can see that you haven't followed/understoof the
instructions on the wiki.
1) Where is the dtb being compiled? (This only needs to be done once
unless you change it so understandable that its not in the script).
2) Where is the dtb being appended to the image? (This needs to be done
everytime you rebuild zImage).
Step 4 and 5 from the wiki:
=====
4) Build the devicetree files (<-- my point 1) above).
Select one of the following files based on your target device.
make ARCH=arm vt8500-bv07.dtb
make ARCH=arm wm8505-ref.dtb
make ARCH=arm wm8650-mid.dtb
5) Build the image <-- my point 2) above)
cat arch/arm/boot/zImage arch/arm/boot/wm8650-mid.dtb >
arch/arm/boot/zImage_w_dtb
replace wm8650-mid.dtb with the appropriate file from step 3
mkimage -A arm -O linux -T kernel -C none -a 0x8000 -e 0x8000 -n "My
Linux" -d arch/arm/boot/zImage_w_dtb ~/uzImage.bin
=====
It would appear you are missing 2 steps in the build process.
