ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Frank Haverkamp]

[-- Attachment #1: Type: text/plain, Size: 962 bytes --]

Hi Artem,

I saw that you moved the old ubi-toolchain to the old-tools directory
with the comment "sort-me-out" in your git-commits.

We need the tool-chain to generate PFI files for the code-update
process. Those pfi-files are suited to be used in streaming mode
together with the pfiflash utility. We use the for NOR and NAND flash
system the same pfi. Therefor the pfi format does not contain flash
geometry information e.g. vid-hdr-offset, erase-block size.

I, and maybe some others on the CC list, would like to see those tools
moved from the "sort-me-out" section again into the ubi-tools directory.
Please let us know, what you think, needs to be done to get this done.

Thanks,

Frank
-- 
IBM Deutschland Entwicklung GmbH, Schoenaicher Str. 220, 71032
Boeblingen, Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Herbert Kircher, Sitz der Gesellschaft: Böblingen,
Registergericht: Amtsgericht Stuttgart, HRB 243294

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5269 bytes --]

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Artem Bityutskiy]

On Mon, 2008-02-18 at 14:49 +0100, Frank Haverkamp wrote:
> We need the tool-chain to generate PFI files for the code-update
> process. Those pfi-files are suited to be used in streaming mode
> together with the pfiflash utility. We use the for NOR and NAND flash
> system the same pfi. Therefor the pfi format does not contain flash
> geometry information e.g. vid-hdr-offset, erase-block size.

Yeah. I tried to use the tools but they did not generate correct UBI
images. I tried to fix this and found out that the internals are too
messy to be maintainable, at least by me.

I cleaned-up what I could, made the utilities consistent, small, and
easy to maintain.

Vs. the PFI stuff, I also tried to clean-up this, but gave up after few
days, and just implemented another utility - ubinize.

I know what is PFI and I decided not to do this, because pfi =
configuration ini file + ubi images. I figured that it does not make
much sense to inroduce this additional magic format.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Thomas Gleixner]

On Mon, 18 Feb 2008, Frank Haverkamp wrote:

> Hi Artem,
> 
> I saw that you moved the old ubi-toolchain to the old-tools directory
> with the comment "sort-me-out" in your git-commits.
> 
> We need the tool-chain to generate PFI files for the code-update
> process. Those pfi-files are suited to be used in streaming mode
> together with the pfiflash utility. We use the for NOR and NAND flash
> system the same pfi. Therefor the pfi format does not contain flash
> geometry information e.g. vid-hdr-offset, erase-block size.
> 
> I, and maybe some others on the CC list, would like to see those tools
> moved from the "sort-me-out" section again into the ubi-tools directory.
> Please let us know, what you think, needs to be done to get this done.

I really do not understand why this move was done anyway. We can move
stuff into retirement when we have a full replacement there.

Just shuffling functional tools around is useless and makes the life
of developers, users and package maintainers harder for no good
reason.

I did not update for some time, but did now and noticed that my
packaging scripts need a real overhaul for no benefit.

Yours grumpy,

	tglx

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Frank Haverkamp]

[-- Attachment #1: Type: text/plain, Size: 965 bytes --]

Hi Thomas,

On Mon, 2008-02-18 at 17:01 +0100, Thomas Gleixner wrote:
> I did not update for some time, but did now and noticed that my
> packaging scripts need a real overhaul for no benefit.
> 

I talked to Artem. He told me that the tools had deficiencies that
prevented him using them. E.g. that he could not put the vid-hdr the
offset he wanted it to put. He promissed me to write up the deficiencies
so that we can work on them. I am sure that after fixing, we can move
the tools back to their old location.

Of course it would be nice, if we could move the tools back even before
the fixup, to avoid changing our build scripts. Let us discuss that with
Artem tomorrow.

Regards,

Frank

-- 
IBM Deutschland Entwicklung GmbH, Schoenaicher Str. 220, 71032
Boeblingen, Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Herbert Kircher, Sitz der Gesellschaft: Böblingen,
Registergericht: Amtsgericht Stuttgart, HRB 243294

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5269 bytes --]

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Josh Boyer]

On Mon, 18 Feb 2008 17:49:03 +0100
Frank Haverkamp <haver@vnet.ibm.com> wrote:

> Hi Thomas,
> 
> On Mon, 2008-02-18 at 17:01 +0100, Thomas Gleixner wrote:
> > I did not update for some time, but did now and noticed that my
> > packaging scripts need a real overhaul for no benefit.
> > 
> 
> I talked to Artem. He told me that the tools had deficiencies that
> prevented him using them. E.g. that he could not put the vid-hdr the
> offset he wanted it to put. He promissed me to write up the deficiencies
> so that we can work on them. I am sure that after fixing, we can move
> the tools back to their old location.

I have a patch for the vid header thing.  I also have a patch for
bin2nand/nand2bin since they are hard coded to use the SmartMedia ECC
byte ordering.

> Of course it would be nice, if we could move the tools back even before
> the fixup, to avoid changing our build scripts. Let us discuss that with
> Artem tomorrow.

I'm a bit grumpy about the move as well, not because they were moved,
but because there's a dozen or so completely useless commits in the
repo.  But, oh well.

josh

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Frank Haverkamp]

[-- Attachment #1: Type: text/plain, Size: 878 bytes --]

Hi Josh,

On Mon, 2008-02-18 at 10:53 -0600, Josh Boyer wrote:
> I also have a patch for
> bin2nand/nand2bin since they are hard coded to use the SmartMedia ECC
> byte ordering.

I added fixes to be more flexible in the fashion the tools place the
OOB-ECC bytes already some time ago. Please have a look if those changes
already do what you had in mind here.

Yesterday I sent a patch to move the location of bin2nand and nand2bin
since those tools are not UBI related. If David accepts the change, you
will find them at a different spot (but anyhow, you were on CC list for
that patch).

Regards,

Frank


-- 
IBM Deutschland Entwicklung GmbH, Schoenaicher Str. 220, 71032
Boeblingen, Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Herbert Kircher, Sitz der Gesellschaft: Böblingen,
Registergericht: Amtsgericht Stuttgart, HRB 243294

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5269 bytes --]

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Josh Boyer]

On Tue, 19 Feb 2008 09:47:36 +0100
Frank Haverkamp <haver@vnet.ibm.com> wrote:

> Hi Josh,
> 
> On Mon, 2008-02-18 at 10:53 -0600, Josh Boyer wrote:
> > I also have a patch for
> > bin2nand/nand2bin since they are hard coded to use the SmartMedia ECC
> > byte ordering.
> 
> I added fixes to be more flexible in the fashion the tools place the
> OOB-ECC bytes already some time ago. Please have a look if those changes
> already do what you had in mind here.

No, they don't.  That was a major part of what I needed (so thank you!),
but ecc layout isn't the same as ecc byte ordering.  I'll post a patch
soonish.

> Yesterday I sent a patch to move the location of bin2nand and nand2bin
> since those tools are not UBI related. If David accepts the change, you
> will find them at a different spot (but anyhow, you were on CC list for
> that patch).

Yes I saw that.  We don't need to wait for David.  I'll look it over
today.

josh

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Frank Haverkamp]

[-- Attachment #1: Type: text/plain, Size: 580 bytes --]

Hi Josh,

On Tue, 2008-02-19 at 06:29 -0600, Josh Boyer wrote:
> No, they don't.  That was a major part of what I needed (so thank
> you!),
> but ecc layout isn't the same as ecc byte ordering.  I'll post a patch
> soonish.

I already thought that ;-). Fine nice that we will get a new feature
into the tool.

Frank

-- 
IBM Deutschland Entwicklung GmbH, Schoenaicher Str. 220, 71032
Boeblingen, Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Herbert Kircher, Sitz der Gesellschaft: Böblingen,
Registergericht: Amtsgericht Stuttgart, HRB 243294

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5269 bytes --]

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Thomas Gleixner]

On Mon, 18 Feb 2008, Frank Haverkamp wrote:
> Hi Thomas,
> 
> On Mon, 2008-02-18 at 17:01 +0100, Thomas Gleixner wrote:
> > I did not update for some time, but did now and noticed that my
> > packaging scripts need a real overhaul for no benefit.
> > 
> 
> I talked to Artem. He told me that the tools had deficiencies that
> prevented him using them.

I really give a sh*t whether Artem can use them or not. He is not the
center of the universe. If he needs them changed, then he should send
a patch or a suitable replacement for the tool, which does not force
others to fixup their scripts and build tools for no benefit. This is
not a sandbox, this are tools used in production environments.

Artem, please stop this nonsense right now. You can create your own
world if you have the urge to do so, but please keep stuff others use
intact. This "I don't like that old crap" attitude is an arrogant and
at the same time ignorant childish behaviour, which has no place in a
serious project.

> E.g. that he could not put the vid-hdr the
> offset he wanted it to put. He promissed me to write up the deficiencies
> so that we can work on them. I am sure that after fixing, we can move
> the tools back to their old location.

No, they go back to their place immediately, period.

Yours even more grumpy,

      tglx

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Artem Bityutskiy]

On Mon, 2008-02-18 at 22:26 +0100, Thomas Gleixner wrote:
> I really give a sh*t whether Artem can use them or not. He is not the
> center of the universe. If he needs them changed, then he should send
> a patch or a suitable replacement for the tool, which does not force
> others to fixup their scripts and build tools for no benefit. This is
> not a sandbox, this are tools used in production environments.
> 
> Artem, please stop this nonsense right now. You can create your own
> world if you have the urge to do so, but please keep stuff others use
> intact. This "I don't like that old crap" attitude is an arrogant and
> at the same time ignorant childish behaviour, which has no place in a
> serious project.
> 
> > E.g. that he could not put the vid-hdr the
> > offset he wanted it to put. He promissed me to write up the deficiencies
> > so that we can work on them. I am sure that after fixing, we can move
> > the tools back to their old location.
> 
> No, they go back to their place immediately, period.
> 
> Yours even more grumpy,

I apologize for this. Indeed that was not nice and arrogant. The tools
are returned to their previous place.

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Artem Bityutskiy]

On Mon, 2008-02-18 at 17:49 +0100, Frank Haverkamp wrote:
> I talked to Artem. He told me that the tools had deficiencies that
> prevented him using them. E.g. that he could not put the vid-hdr the
> offset he wanted it to put. He promissed me to write up the deficiencies
> so that we can work on them. I am sure that after fixing, we can move
> the tools back to their old location.
> 
> Of course it would be nice, if we could move the tools back even before
> the fixup, to avoid changing our build scripts. Let us discuss that with
> Artem tomorrow.

I can return the tools to the state they were before I touched them, is
it OK for you?

-- 
Best regards,
Artem Bityutskiy (Битюцкий Артём)

Re: ubi-utils: Moving mkpfi, pfiflash, ... to old-tools

[Frank Haverkamp]

[-- Attachment #1: Type: text/plain, Size: 616 bytes --]

Hi Artem,

On Tue, 2008-02-19 at 11:34 +0200, Artem Bityutskiy wrote:
> I can return the tools to the state they were before I touched them,
> is
> it OK for you?

that would be very nice of you. And if you could still create the list
of deficiencies e.g. that the modification of the vid-hdr-offset is
missing it would be also nice.

Thanks,

Frank

-- 
IBM Deutschland Entwicklung GmbH, Schoenaicher Str. 220, 71032
Boeblingen, Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Herbert Kircher, Sitz der Gesellschaft: Böblingen,
Registergericht: Amtsgericht Stuttgart, HRB 243294

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 5269 bytes --]