libubiio : userspace UBI input output library

libubiio : userspace UBI input output library

[Sidney Amani]

Hi list,

UBIIO library is an user-space library which exports same functions in
UBI Linux kernel module.
It provides exactly the same API as the kernel does, that way kernel
code can be used in user-space without any change.
The code is widely inspired with libubi source code.

Documentation can be found at http://uffs.org/api/libubiio.html
It is obviously the same as UBI kernel API, but there is actually few
differences in return values.

The code can be found at http://git.uffs.org/

We would like to know if you are interested in it and want to include
it in mtdutils.

Regards.

-- 
Sidney Amani

Re: libubiio : userspace UBI input output library

[Artem Bityutskiy]

On Sun, 2009-03-01 at 15:45 +0100, Sidney Amani wrote:
> Hi list,
> 
> UBIIO library is an user-space library which exports same functions in
> UBI Linux kernel module.
> It provides exactly the same API as the kernel does, that way kernel
> code can be used in user-space without any change.
> The code is widely inspired with libubi source code.
> 
> Documentation can be found at http://uffs.org/api/libubiio.html
> It is obviously the same as UBI kernel API, but there is actually few
> differences in return values.
> 
> The code can be found at http://git.uffs.org/
> 
> We would like to know if you are interested in it and want to include
> it in mtdutils.

Hi, I guess this is useful. Just wonder, if you are targeted to
mtd-utils, why not to develop in there from the very beginning? Use
similar coding style, clean-up current code if needed, etc?

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: libubiio : userspace UBI input output library

[Corentin Chary]

On Tue, Mar 3, 2009 at 7:38 AM, Artem Bityutskiy <dedekind@infradead.org> wrote:
> On Sun, 2009-03-01 at 15:45 +0100, Sidney Amani wrote:
>> Hi list,
>>
>> UBIIO library is an user-space library which exports same functions in
>> UBI Linux kernel module.
>> It provides exactly the same API as the kernel does, that way kernel
>> code can be used in user-space without any change.
>> The code is widely inspired with libubi source code.
>>
>> Documentation can be found at http://uffs.org/api/libubiio.html
>> It is obviously the same as UBI kernel API, but there is actually few
>> differences in return values.
>>
>> The code can be found at http://git.uffs.org/
>>
>> We would like to know if you are interested in it and want to include
>> it in mtdutils.
>
> Hi, I guess this is useful. Just wonder, if you are targeted to
> mtd-utils, why not to develop in there from the very beginning? Use
> similar coding style, clean-up current code if needed, etc?
>

Hi,
We first made that lib for us. And as it can be considered as "usable" (wich is
not the case for the rest of the project) we posted it there.
If you think it can be merged in mtd-utils, we'll take some time for
the coding style
and cleanups... I wanted to do that before, but didn't had a lot of
time to do so.
Which part of mtd-utils could use this lib ? maybe mkfs.ubifs ?

-- 
Corentin Chary
http://xf.iksaif.net

Re: libubiio : userspace UBI input output library

[Artem Bityutskiy]

On Tue, 2009-03-03 at 08:48 +0100, Corentin Chary wrote:
> Hi,
> We first made that lib for us. And as it can be considered as "usable" (wich is
> not the case for the rest of the project) we posted it there.
> If you think it can be merged in mtd-utils, we'll take some time for
> the coding style
> and cleanups... I wanted to do that before, but didn't had a lot of
> time to do so.
> Which part of mtd-utils could use this lib ? maybe mkfs.ubifs ?

Well, I personally do not see any reason not to merge this. I just
suggest to work with mtd-utils from the very beginning. I know it looks
very messy and crappy. But you may make it better while working with it.
And that would be a great contribution as well. And this would be more
collaborative, I think.

So far, it looks like ubi-utils/new-utils would be a better place for
this, isn't it? But it would also be nice to send the source code to the
ML. It will unlikely get much review, but still.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: libubiio : userspace UBI input output library

[Corentin Chary]

On Tue, Mar 3, 2009 at 8:59 AM, Artem Bityutskiy <dedekind@infradead.org> wrote:
> On Tue, 2009-03-03 at 08:48 +0100, Corentin Chary wrote:
>> Hi,
>> We first made that lib for us. And as it can be considered as "usable" (wich is
>> not the case for the rest of the project) we posted it there.
>> If you think it can be merged in mtd-utils, we'll take some time for
>> the coding style
>> and cleanups... I wanted to do that before, but didn't had a lot of
>> time to do so.
>> Which part of mtd-utils could use this lib ? maybe mkfs.ubifs ?
>
> Well, I personally do not see any reason not to merge this. I just
> suggest to work with mtd-utils from the very beginning. I know it looks
> very messy and crappy. But you may make it better while working with it.
> And that would be a great contribution as well. And this would be more
> collaborative, I think.
Indeed. I know there is a lot to do with mtd-utils, but my kernel hacking time
is full these days. I really should find a job where I could be paid for that !

> So far, it looks like ubi-utils/new-utils would be a better place for
> this, isn't it? But it would also be nice to send the source code to the
> ML. It will unlikely get much review, but still.
Ok, in the next weeks we'll make some cleans. Then we'll send patchs
to merge it in mtd-utils.
In the mean time, you can "git clone git://git.uffs.org/libubiio.git"

Thanks
-- 
Corentin Chary
http://xf.iksaif.net

Re: libubiio : userspace UBI input output library

[Artem Bityutskiy]

On Tue, 2009-03-03 at 09:37 +0100, Corentin Chary wrote:
> On Tue, Mar 3, 2009 at 8:59 AM, Artem Bityutskiy <dedekind@infradead.org> wrote:
> > On Tue, 2009-03-03 at 08:48 +0100, Corentin Chary wrote:
> >> Hi,
> >> We first made that lib for us. And as it can be considered as "usable" (wich is
> >> not the case for the rest of the project) we posted it there.
> >> If you think it can be merged in mtd-utils, we'll take some time for
> >> the coding style
> >> and cleanups... I wanted to do that before, but didn't had a lot of
> >> time to do so.
> >> Which part of mtd-utils could use this lib ? maybe mkfs.ubifs ?
> >
> > Well, I personally do not see any reason not to merge this. I just
> > suggest to work with mtd-utils from the very beginning. I know it looks
> > very messy and crappy. But you may make it better while working with it.
> > And that would be a great contribution as well. And this would be more
> > collaborative, I think.
> Indeed. I know there is a lot to do with mtd-utils, but my kernel hacking time
> is full these days. I really should find a job where I could be paid for that !

You mean your uffs project is not a paid full-time stuff?

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: libubiio : userspace UBI input output library

[Corentin Chary]

On Tue, Mar 3, 2009 at 9:39 AM, Artem Bityutskiy <dedekind@infradead.org> wrote:
> On Tue, 2009-03-03 at 09:37 +0100, Corentin Chary wrote:
>> On Tue, Mar 3, 2009 at 8:59 AM, Artem Bityutskiy <dedekind@infradead.org> wrote:
>> > On Tue, 2009-03-03 at 08:48 +0100, Corentin Chary wrote:
>> >> Hi,
>> >> We first made that lib for us. And as it can be considered as "usable" (wich is
>> >> not the case for the rest of the project) we posted it there.
>> >> If you think it can be merged in mtd-utils, we'll take some time for
>> >> the coding style
>> >> and cleanups... I wanted to do that before, but didn't had a lot of
>> >> time to do so.
>> >> Which part of mtd-utils could use this lib ? maybe mkfs.ubifs ?
>> >
>> > Well, I personally do not see any reason not to merge this. I just
>> > suggest to work with mtd-utils from the very beginning. I know it looks
>> > very messy and crappy. But you may make it better while working with it.
>> > And that would be a great contribution as well. And this would be more
>> > collaborative, I think.
>> Indeed. I know there is a lot to do with mtd-utils, but my kernel hacking time
>> is full these days. I really should find a job where I could be paid for that !
>
> You mean your uffs project is not a paid full-time stuff?

It's a end study project, so not full-time, but still, taking a lot of =)



-- 
Corentin Chary
http://xf.iksaif.net

Re: libubiio : userspace UBI input output library

[Sidney Amani]

On Tue, Mar 3, 2009 at 8:59 AM, Artem Bityutskiy <dedekind@infradead.org> wrote:
> Well, I personally do not see any reason not to merge this. I just
> suggest to work with mtd-utils from the very beginning. I know it looks
> very messy and crappy. But you may make it better while working with it.
> And that would be a great contribution as well. And this would be more
> collaborative, I think.
>

Hi Artem,

Actually yes we shouldn't work on our own side, but as we didn't feel
like sending patches and patches..
It's may be possible to work on a specific branch?
Anyway next time we will try to do it more collaborative.

> So far, it looks like ubi-utils/new-utils would be a better place for
> this, isn't it?

I think so.
Our lib has some redundant functionalities with libubi (e.g
ubi_get_device_info), so we may want to take some functions off.

Yes, mtd-utils is very messy, when will the new utils replace old one?
We are probably not in the best position to do it.

> But it would also be nice to send the source code to the
> ML. It will unlikely get much review, but still.
>

Ok, we will do it.

-- 
Sidney Amani
http://uffs.org

Re: libubiio : userspace UBI input output library

[Corentin Chary]

> It's may be possible to work on a specific branch?
I think well just put a tree based on
http://git.infradead.org/mtd-utils.git?a=shortlog;h=refs/heads/ubi
at git.uffs.org, then anyone will be able to pull from us, and sending
patch will be easy thanks to git format-patch/git send-email


-- 
Corentin Chary
http://xf.iksaif.net

Re: libubiio : userspace UBI input output library

[Artem Bityutskiy]

On Tue, 2009-03-03 at 11:40 +0100, Sidney Amani wrote:
> Yes, mtd-utils is very messy, when will the new utils replace old one?
> We are probably not in the best position to do it.

Well, this is historical. I'll try to re-arrange this.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: libubiio : userspace UBI input output library

[Artem Bityutskiy]

On Tue, 2009-03-03 at 13:07 +0200, Artem Bityutskiy wrote:
> On Tue, 2009-03-03 at 11:40 +0100, Sidney Amani wrote:
> > Yes, mtd-utils is very messy, when will the new utils replace old one?
> > We are probably not in the best position to do it.
> 
> Well, this is historical. I'll try to re-arrange this.

Not sure what you mean. You may always clone the tree and have
your own. And send the bits which are ready and of general interest
out, making sure they are merged. Then you re-base. This is
how this stuff usually goes, at least in the kernel.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: libubiio : userspace UBI input output library

[Corentin Chary]

On Tue, Mar 3, 2009 at 12:09 PM, Artem Bityutskiy
<dedekind@infradead.org> wrote:
> On Tue, 2009-03-03 at 13:07 +0200, Artem Bityutskiy wrote:
>> On Tue, 2009-03-03 at 11:40 +0100, Sidney Amani wrote:
>> > Yes, mtd-utils is very messy, when will the new utils replace old one?
>> > We are probably not in the best position to do it.
>>
>> Well, this is historical. I'll try to re-arrange this.
>
> Not sure what you mean. You may always clone the tree and have
> your own. And send the bits which are ready and of general interest
> out, making sure they are merged. Then you re-base. This is
> how this stuff usually goes, at least in the kernel.

Yep, it's what he/I ment.
I'm already familiar with the process, it's how it works between
acpi4asus and the acpi subsystem :)



-- 
Corentin Chary
http://xf.iksaif.net

Re: libubiio : userspace UBI input output library

[Artem Bityutskiy]

On Tue, 2009-03-03 at 11:40 +0100, Sidney Amani wrote:
> Yes, mtd-utils is very messy, when will the new utils replace old one?

Done.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)