Skip to contentSkip to conversation

AT91 NAND om AT91SAM9260EK

U-Boot-Users Ivan Kuten 5 participants 17 messages Feb 10, 2007 – Feb 12, 2007 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): mailman.12584.1171099992.16820.u-boot-users@lists.sourceforge.net

Ivan Kuten

author
#1 lore raw
Quoted 11 lines
Quoted 3 lines

But we agree that this *is* specific to a certain class of processors of the ARM family, right?

No. You can add dataflash to any processor with an SPI interface or even if you do not have an SPI interface you can access the device using standard I/O pins.

At the moment, inside U-Boot it is only used by AT91 processors but that is really not the same thing.

Hello, i'm afraid to be wrong, but at45.c is used also by Blackfin version of uboot at blackfin.uclinux.org to allow booting off Atmel Dataflash some of Blackfin boards. So at45.c is not a processor specific.

Best regards, Ivan

#2 lore raw
Quoted 8 lines
Quoted 8 lines
Quoted 3 lines

But we agree that this *is* specific to a certain class of processors of the ARM family, right?

No. You can add dataflash to any processor with an SPI interface or even if you do not have an SPI interface you can access the device using standard I/O pins.

Quoted 9 lines
Quoted 3 lines

At the moment, inside U-Boot it is only used by AT91 processors but that is really not the same thing.

Hello, i'm afraid to be wrong, but at45.c is used also by Blackfin version of uboot at blackfin.uclinux.org to allow booting off Atmel Dataflash some of Blackfin boards. So at45.c is not a processor specific.

The at45.c contains the at45 commands and the spi driver. For all other chips the spi driver seems to be
in the cpu/xxx directories.
That is why my patch merges the duplicated at45.c's into a single file, but the splits the file into two parts, one containing the at45 commands in driver/at45.c and one containing the spi driver for the at91rm9200 in cpu/arm920t/at91rm9200.

Any duplication causes by this is against code which is not yet present in the current source tree, but only against code Haavard wants to write.

Duplication of a few subroutines is not uncommon in U-boot Looking through the Power PC directories and the NIOS/NIOS2 you easily find subroutines beeing duplicated?between different architectures.

I do not see any "cpu/<vendor>" directories in U-Boot and to split up the spi driver into multiple files just because a few of the routines are duplicated will make maintenance harder, and will muddle the reponsibility.

I do not understand who will be responsible for maintaining the spi support for at91 chips if it is rewritten.

At least Wolfgang, which multiple times has said he wants to have a single Makefile to edit should appreciate this.

Please realize that, while I am providing the patches, I am just taking what the Atmel AT91 groups is delivering to the end customers.
The AT91 group is fed up with zero response
and has basically given up on trying to get patches approved.

I am in no way able to support fully testing significant modifications of these
patches and I belive that if this is done, there will be noone accepting responsibility for that these patches actually generate U-Boot binaries's which will actually work.

From what I have seen, the patches from the AT91 group fit closely into the current
way U-Boot is designed, and they should be reviewed by people that actually study the patches so that an unbiased opinion is given.

Best regards, Ivan

Quoted 2 lines

Best Regards
Ulf Samuelsson

#3 lore raw
Quoted 3 lines · Ulf Samuelsson

Any duplication causes by this is against code which is not yet present in the current source tree, but only against code Haavard wants to write.

Ok, I think there's some misunderstanding going on here. I didn't mean to say that the spi driver _must_ be made common for all chips _now_. Please disregard my suggestions about a common spi driver for now. Let's get this stuff moving.

Quoted 2 lines

I do not understand who will be responsible for maintaining the spi support for at91 chips if it is rewritten.

I'm responsible for the atmel_spi driver in the Linux kernel which is on its way into mainstream hopefully any day now (it's been accepted by the SPI maintainer.) This driver supports AT91RM9200, AT91SAM926x and AT32AP7000, so I don't see why I can't be responsible for the u-boot driver as well if nobody else are going to step up.

But let's do it your way for now and move the SPI parts of the dataflash driver to the CPU directory. We can always change it later, when support for all boards are in place and it becomes clearer which parts are shareable and which parts are not.

Quoted 4 lines

Please realize that, while I am providing the patches, I am just taking what the Atmel AT91 groups is delivering to the end customers.
The AT91 group is fed up with zero response
and has basically given up on trying to get patches approved.

Hmm...that sounds somehow familiar :-P

Hopefully, the new custodian model will improve this.

Quoted 5 lines

I am in no way able to support fully testing significant modifications of these
patches and I belive that if this is done, there will be noone accepting responsibility for that these patches actually generate U-Boot binaries's which will actually work.

U-boot is a community effort. We depend on an active community to test non-trivial changes. Michel has already volunteered, that's exactly what we need.

We can't depend _only_ on the AT91 group to make sure that dataflash will work on all boards using it, nor can we depend _only_ on the AVR32 group. As far as I know, none of us are testing any of this on Blackfin, for example.

Of course, Atmel needs to do testing as well, and it will probably make sense to offer special "Atmel blessed" versions of u-boot which have been tested thoroughly on all supported configurations for the customers that need this. But we can't say that nobody is allowed to make any changes to the official driver because it has been tested and we don't want to do it again.

Any changes Atmel does based on such testing should of course be submitted for inclusion upstream. But for this to work, we do need a timely response from the upstream maintainers.

Quoted 4 lines
Quoted 1 line

From what I have seen, the patches from the AT91 group fit closely into the

current
way U-Boot is designed, and they should be reviewed by people that actually study the patches so that an unbiased opinion is given.

I don't know what you're getting at here, but yes, in order to review a patch, you _do_ need to study it and give an unbiased opinion. But it works the other way around too -- if you want review, you _have_ to be willing to make changes based on that review.

Haavard

#4 lore raw
Quoted 19 lines · Ulf Samuelsson
Quoted 3 lines

Any duplication causes by this is against code which is not yet present in the current source tree, but only against code Haavard wants to write.

Ok, I think there's some misunderstanding going on here. I didn't mean to say that the spi driver _must_ be made common for all chips _now_. Please disregard my suggestions about a common spi driver for now. Let's get this stuff moving.

Quoted 2 lines

I do not understand who will be responsible for maintaining the spi support for at91 chips if it is rewritten.

I'm responsible for the atmel_spi driver in the Linux kernel which is on its way into mainstream hopefully any day now (it's been accepted by the SPI maintainer.) This driver supports AT91RM9200, AT91SAM926x and AT32AP7000, so I don't see why I can't be responsible for the u-boot driver as well if nobody else are going to step up.

Quoted 5 lines

But let's do it your way for now and move the SPI parts of the dataflash driver to the CPU directory. We can always change it later, when support for all boards are in place and it becomes clearer which parts are shareable and which parts are not.

This is exactly how I want to proceed.
There are a number of patches that needs to be sent
in order to have at91sam926x support.
I would like to submit them, have everything working. There are changes which I consider reasonable to request. There are also changes which I consider unreasonable to request.

Requesting me to modify the *contents* of a duplicated file because I want to have a single file I consider as an unreasonable request.

Quoted 16 lines
Quoted 5 lines

Please realize that, while I am providing the patches, I am just taking what
the Atmel AT91 groups is delivering to the end customers. The AT91 group is fed up with zero response
and has basically given up on trying to get patches approved.

Hmm...that sounds somehow familiar :-P

Hopefully, the new custodian model will improve this.

Quoted 5 lines

I am in no way able to support fully testing significant modifications of these
patches and I belive that if this is done, there will be noone accepting responsibility for that these patches actually generate U-Boot binaries's which will actually work.

Quoted 3 lines

U-boot is a community effort. We depend on an active community to test non-trivial changes. Michel has already volunteered, that's exactly what we need.

Yes and we need to have people preparing to test
* SAM9261
* SAM9263
* AT91RM9200DK
* AT91RM9200EK
* AT91RM9200DF
* CMC_PU2
as well.

Quoted 6 lines

We can't depend _only_ on the AT91 group to make sure that dataflash will work on all boards using it, nor can we depend _only_ on the AVR32 group. As far as I know, none of us are testing any of this on Blackfin, for example.

The blackfin support is not in the main tree as far as I could tell. It is available on a proprietary home page.

Quoted 6 lines

Of course, Atmel needs to do testing as well, and it will probably make sense to offer special "Atmel blessed" versions of u-boot which have been tested thoroughly on all supported configurations for the customers that need this. But we can't say that nobody is allowed to make any changes to the official driver because it has been tested and we don't want to do it again.

No, but I am only asking for two things:

1) that I can apply the complete patchset for the at91sam926x and have that accepted before people start this activity
2) That people make sure that they do not destroy the U-boot support for at91 by providing untested patches.

Quoted 6 lines

Any changes Atmel does based on such testing should of course be submitted for inclusion upstream. But for this to work, we do need a timely response from the upstream maintainers.

Quoted 2 lines
Quoted 2 lines

From what I have seen, the patches from the AT91 group fit closely into the

Quoted 8 lines
Quoted 3 lines

current
way U-Boot is designed, and they should be reviewed by people that actually study the patches so that an unbiased opinion is given.

I don't know what you're getting at here, but yes, in order to review a patch, you _do_ need to study it and give an unbiased opinion. But it works the other way around too -- if you want review, you _have_ to be willing to make changes based on that review.

Yes, I am prepared to put the spi part and the at45 support parts of at45.c in any suggested directory.
That is a reasonable thing to ask for.

To ask me to modify the *contents* of at45.c
just because I want to remove duplication is out of line, IMHO. I dont say it is a bad thing, just that it should be done after the patchset is applied, and then by a group willing to maintain and test the code.

If I am not allowed to remove the duplication, then there is no possibility to have a common at45.c and I will be forced to modify the drivers/at45.c in the at91sam926x patchset to be located

board/at91sam9260ek/at45.c
board/at91sam9261ek/at45.c
board/at91sam9260ek/at45.c

so we will have 5 at45.c instead of 1.
I doubt that patch is going to be accepted.

Quoted 3 lines

Haavard

Best Regards
Ulf Samuelsson

#5 lore raw
Quoted 2 lines · Ulf Samuelsson

Requesting me to modify the *contents* of a duplicated file because I want to have a single file I consider as an unreasonable request.

I agree, but I don't recall anyone ever suggesting that (except that my initial proposal did schedule some changes before the at91sam926x board patches, but I also said that I was perfectly fine with doing them later.)

Quoted 11 lines
Quoted 6 lines

Of course, Atmel needs to do testing as well, and it will probably make sense to offer special "Atmel blessed" versions of u-boot which have been tested thoroughly on all supported configurations for the customers that need this. But we can't say that nobody is allowed to make any changes to the official driver because it has been tested and we don't want to do it again.

No, but I am only asking for two things:

1) that I can apply the complete patchset for the at91sam926x and have that accepted before people start this activity

That's up to Wolfgang I guess. I've already said that I can wait for the at91sam926x support to go in before I start modifying the at45 driver.

Quoted 2 lines

2) That people make sure that they do not destroy the U-boot support for at91 by providing untested patches.

Well, I do agree that submitting totally untested patches is unacceptable unless they are clearly marked as such. On the other hand, it's also totally unreasonable to demand that people submitting patches test them on boards they don't have.

If someone submits a patch that's not fully tested, we'll just have to test it before it enters mainline. It's really that simple.

Quoted 19 lines
Quoted 4 lines

I don't know what you're getting at here, but yes, in order to review a patch, you _do_ need to study it and give an unbiased opinion. But it works the other way around too -- if you want review, you _have_ to be willing to make changes based on that review.

Yes, I am prepared to put the spi part and the at45 support parts of at45.c in any suggested directory.
That is a reasonable thing to ask for.

To ask me to modify the *contents* of at45.c
just because I want to remove duplication is out of line, IMHO. I dont say it is a bad thing, just that it should be done after the patchset is applied, and then by a group willing to maintain and test the code.

If I am not allowed to remove the duplication, then there is no possibility to have a common at45.c and I will be forced to modify the drivers/at45.c in the at91sam926x patchset to be located

Quoted 6 lines

board/at91sam9260ek/at45.c
board/at91sam9261ek/at45.c
board/at91sam9260ek/at45.c

so we will have 5 at45.c instead of 1.
I doubt that patch is going to be accepted.

You keep repeating this as if anyone actually disagreed with you that we want a single unified at45 driver. Or that we should only start changing it _after_ it's been consolidated.

And yes, I happen to be part of a group willing to maintain and test the code. Not on every single board out there, but if we can manage to find a tester for at least one RM board and one SAM board, I think we'll be pretty well covered.

Haavard

#6 lore raw
Quoted 9 lines · Ulf Samuelsson
Quoted 2 lines

Requesting me to modify the *contents* of a duplicated file because I want to have a single file I consider as an unreasonable request.

I agree, but I don't recall anyone ever suggesting that (except that my initial proposal did schedule some changes before the at91sam926x board patches, but I also said that I was perfectly fine with doing them later.)

No, Wolfgang implied that I could not move/split a file unless I promised to "clean" it up.

I think we have basic agreement and we now should await the U-Boot maintainers reaction.

/Ulf

#7 lore raw
Quoted 4 lines · Ulf Samuelsson
Quoted 4 lines · Ulf Samuelsson
Quoted 2 lines

Requesting me to modify the *contents* of a duplicated file because I want to have a single file I consider as an unreasonable request.

Quoted 8 lines
Quoted 5 lines

I agree, but I don't recall anyone ever suggesting that (except that my initial proposal did schedule some changes before the at91sam926x board patches, but I also said that I was perfectly fine with doing them later.)

No, Wolfgang implied that I could not move/split a file unless I promised to "clean" it up.

Oh, yes, I do recall that. But he did say "later", and that's what I'm offering to do.

Quoted 2 lines

I think we have basic agreement and we now should await the U-Boot maintainers reaction.

Agreed

Haavard

#8 lore raw
In reply to Haavard Skinnemoen
Quoted 7 lines · In message <[email protected]> you…
Quoted 3 lines

1) that I can apply the complete patchset for the at91sam926x and have that accepted before people start this activity

That's up to Wolfgang I guess. I've already said that I can wait for the at91sam926x support to go in before I start modifying the at45 driver.

Ulf has made it pretty clear that he thinks that I am just a show stopper for all his efforts, and that I don't have a clue what he's talking about. It seems I cannot help that, so I keep my mouth shut here.

I promise that we will provide the infrastructure (git repos) for the custodians as fast as possible. As far as I can see, at the moment this affects only ARM boards, so please run this throught he ARM custodian then (Peter Pearse). If it's acceptable to him, and nobody else complains, I will accept this, too.

Quoted 7 lines
Quoted 2 lines

2) That people make sure that they do not destroy the U-boot support for at91 by providing untested patches.

Well, I do agree that submitting totally untested patches is unacceptable unless they are clearly marked as such. On the other hand, it's also totally unreasonable to demand that people submitting patches test them on boards they don't have.

It's unreasonable, and impossible.

We have a pretty big selection of boards in our lab, but nevertheless I've never been able to test any change on all architectures and boards. And nobody else will ever do that.

Quoted 2 lines

If someone submits a patch that's not fully tested, we'll just have to test it before it enters mainline. It's really that simple.

And actually in some cases we will simply have to push it into mainline - in the past, there has often been very little of even zero feedback for patches or test branches. Some people simply don't wake up before their board does not compile any more in the standard tree. I cannot help this, and we will definitely not wait to see an ACK for each and every board that is listed in the U-Boot tree.

Best regards,

Wolfgang Denk

Signature · 6 lines
DENX Software Engineering GmbH, HRB 165235 Munich, CEO: Wolfgang Denk
Office:  Kirchenstr. 5,       D-82194 Groebenzell,            Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Conceptual integrity in turn dictates that the  design  must  proceed
from  one  mind,  or  from  a  very small number of agreeing resonant
minds.               - Frederick Brooks Jr., "The Mythical Man Month" 
#9 lore raw
Quoted 6 lines

I promise that we will provide the infrastructure (git repos) for the custodians as fast as possible. As far as I can see, at the moment this affects only ARM boards, so please run this throught he ARM custodian then (Peter Pearse). If it's acceptable to him, and nobody else complains, I will accept this, too.

OK, acceptable

Quoted 9 lines
Quoted 7 lines
Quoted 2 lines

2) That people make sure that they do not destroy the U-boot support for at91 by providing untested patches.

Well, I do agree that submitting totally untested patches is unacceptable unless they are clearly marked as such. On the other hand, it's also totally unreasonable to demand that people submitting patches test them on boards they don't have.

It's unreasonable, and impossible.

If you are working on, let's say, "common" files, then
you affect all boards, and then you should test on a reasonable subset.

If you enter and do significant modification on a file in the board area, then I think things changes.
A contributor has no business changing a board file if he cannot reasonably verify that this will not break a board.

The same for files in the cpu/xxx directories.
You should not be changing those without testing on boards using that CPU.

If someone spends time to thoroughly test boards and has this included in the mainstream, then anyone changing this code should be aware that breaking this code has a price to end users.

Splitting a driver for a peripheral into several files just because a few subroutined can be shared has a price
to the maintainer, since this will reduce the overview of the code.

A few kB of duplication is not a high price to pay for working U-Boot and easier maintenance.
If I look at the PowerPC/NIOS code, this is exactly how it is done. Duplications in several areas.

Quoted 8 lines

We have a pretty big selection of boards in our lab, but nevertheless I've never been able to test any change on all architectures and boards. And nobody else will ever do that.

Quoted 2 lines

If someone submits a patch that's not fully tested, we'll just have to test it before it enters mainline. It's really that simple.

Quoted 6 lines

And actually in some cases we will simply have to push it into mainline - in the past, there has often been very little of even zero feedback for patches or test branches. Some people simply don't wake up before their board does not compile any more in the standard tree. I cannot help this, and we will definitely not wait to see an ACK for each and every board that is listed in the U-Boot tree.

No, but as mentioned above, different parts of U-Boot should require the submitter to focus on different subsets of boards.

The proposed dataflash patches in our other heated discussion for instance should only be accepted once proven that it actually works with dataflash memory chips.
The proposer seemed to be happy if the stuff compiled... Since that mainly affects Atmel boards, he should make sure to get Atmel boards or sign up testers for it,
and have it tested before even thinking about submit patches.

It really does not make sense to test his patches only? on boards which does not use the dataflash.

Quoted 2 lines

Best Regards
Ulf Samuelsson

#10 lore raw

Dear Ulf,

Quoted 3 lines · in message <00f201c74e25$71fdbf80$01c4af0a@Glamdring> you wrote:

A contributor has no business changing a board file if he cannot reasonably verify that this will not break a board.

You are right in principle, but live is not always black and white.

Quoted 5 lines
Quoted 2 lines

I cannot help this, and we will definitely not wait to see an ACK for each and every board that is listed in the U-Boot tree.

No, but as mentioned above, different parts of U-Boot should require the submitter to focus on different subsets of boards.

It ain't that easy. Take for example this stupid ". = .;" addition to the linker scripts that was needed to make some versions of the toolchain happy - of course I add this to ALL boards without being able to test even half of them.

And as mentioned before: if I have a global change to make, and I
cannot get response from board / architecture maintainers in
reasonable time, that change will be made globally, even if it breaks their boards. It's IMHO better to have a board clearly broken than having all boards in N different patch states.

Quoted 7 lines

The proposed dataflash patches in our other heated discussion for instance should only be accepted once proven that it actually works with dataflash memory chips.
The proposer seemed to be happy if the stuff compiled... Since that mainly affects Atmel boards, he should make sure to get Atmel boards or sign up testers for it,
and have it tested before even thinking about submit patches.

I disagree with you. You cannot expect any maintainer or custodian or other person volunteering to contribute to U-Boot code (or any other free software project) to "get Atmel boards or sign up testers for it". There are board maintainers, and it is their task to help testing things. And yes, this includes that they will sometimes have to help sorting out problems created by other people. That's how it has been, and will remain, even if Atmel was shipping test systems for free.

Or do you intend to get a cmc_pu2 board or sign up testers for it to check your patches? No, it will be me who has to run the tests, right?

Quoted 2 lines

It really does not make sense to test his patches only on boards which does not use the dataflash.

If that's all he has, then he has no other choice.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH, HRB 165235 Munich, CEO: Wolfgang Denk
Office:  Kirchenstr. 5,       D-82194 Groebenzell,            Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"If the code and  the  comments  disagree,  then  both  are  probably
wrong."                                                - Norm Schryer
#11 lore raw
Quoted 12 lines

Dear Ulf,

Quoted 3 lines · in message <00f201c74e25$71fdbf80$01c4af0a@Glamdring> you wrote:

A contributor has no business changing a board file if he cannot reasonably verify that this will not break a board.

You are right in principle, but live is not always black and white.

Quoted 3 lines
Quoted 2 lines

I cannot help this, and we will definitely not wait to see an ACK for each and every board that is listed in the U-Boot tree.

Quoted 7 lines
Quoted 2 lines

No, but as mentioned above, different parts of U-Boot should require the submitter to focus on different subsets of boards.

It ain't that easy. Take for example this stupid ". = .;" addition to the linker scripts that was needed to make some versions of the toolchain happy - of course I add this to ALL boards without being able to test even half of them.

I think that is something which I could live with.
(This can be verified by building the board using MAKEALL) Going in and writing a new UART driver and replacing the old one is not something I could live with if they do not test this.

Quoted 6 lines

And as mentioned before: if I have a global change to make, and I
cannot get response from board / architecture maintainers in
reasonable time, that change will be made globally, even if it breaks their boards. It's IMHO better to have a board clearly broken than having all boards in N different patch states.

Global changes needs to go ahead, but I already point that out.

Quoted 9 lines
Quoted 8 lines

The proposed dataflash patches in our other heated discussion for instance should only be accepted once proven that it actually works with dataflash memory chips.
The proposer seemed to be happy if the stuff compiled... Since that mainly affects Atmel boards, he should make sure to get Atmel boards or sign up testers for it,
and have it tested before even thinking about submit patches.

Quoted 8 lines

I disagree with you. You cannot expect any maintainer or custodian or other person volunteering to contribute to U-Boot code (or any other free software project) to "get Atmel boards or sign up testers for it". There are board maintainers, and it is their task to help testing things. And yes, this includes that they will sometimes have to help sorting out problems created by other people. That's how it has been, and will remain, even if Atmel was shipping test systems for free.

I expect anyone that significantly modifies anything to ensure that the submitted patch is tested on *something*.
I expect that the fact that things compiles is not considered a measure of success
for significant code changes.

From this follows that anything which is specific to an Atmel target and not used anywhere else cannot be tested on anything but an Atmel target. Your changes above does not change code.

Quoted 3 lines

Or do you intend to get a cmc_pu2 board or sign up testers for it to check your patches? No, it will be me who has to run the tests, right?

No, I know that this is a weak point of the argument, and that is why I would like to avoid changing significantly the contents of the patch.

The only change I see I do, is the bugfix in form of a type cast. The current at45.c only exercises the internal resources of the at91rm9200, so this fix should be enough to exercise on any at9rm9200 board. There are board specific things like which chip select to use etc, but all these parts are part of other files, and not part of at45.c.

I really just change where the code is located in the source tree. I have tested building cmc_pu2 to ensure that is not a problem. Have also tested the patch so that non-Atmel arm targets and non-Arm targets can build (Power PC target).

Quoted 5 lines
Quoted 3 lines

It really does not make sense to test his patches only on boards which does not use the dataflash.

If that's all he has, then he has no other choice.

He has always the choice of realizing that he maybe should leave this possible improvement to someone capable of testing it and concentrate on stuff which he can test.

Quoted 5 lines

Best regards,

Wolfgang Denk

Best Regards
Ulf Samuelsson ulf at atmel.com
Atmel Nordic AB

Mail:  Box 2033, 174 02 Sundbyberg, Sweden
Visit:  Kavalleriv?gen 24, 174 58 Sundbyberg, Sweden

Phone +46 (8) 441 54 22 Fax +46 (8) 441 54 29
GSM +46 (706) 22 44 57

Technical support when I am not available:
AT89 C51 Applications Group: mailto:micro.hotline at nto.atmel.com AT90 AVR Applications Group: mailto:avr at atmel.com AT91 ARM Applications Group: mailto:at91support at atmel.com FPSLIC Application Group: mailto:fpslic at atmel.com Best AVR
link: www.avrfreaks.net

#12 lore raw
Quoted 8 lines · Ulf Samuelsson
Quoted 4 lines
Quoted 2 lines

It really does not make sense to test his patches only on boards which does not use the dataflash.

If that's all he has, then he has no other choice.

He has always the choice of realizing that he maybe should leave this possible improvement to someone capable of testing it and concentrate on stuff which he can test.

Yeah, that's a nice attitude. "Go away, we don't want your patches." Without even looking at them?

Btw, I can't help but feel that you're using me as an example here, since I _did_ at one point suggest that one particular patch would only need compile-testing. If you are, please stop it because you're misrepresenting what I said.

Haavard

#13 lore raw
Quoted 12 lines · Ulf Samuelsson
Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

It really does not make sense to test his patches only on boards which does not use the dataflash.

If that's all he has, then he has no other choice.

He has always the choice of realizing that he maybe should leave this possible improvement to someone capable of testing it and concentrate on stuff which he can test.

Yeah, that's a nice attitude. "Go away, we don't want your patches." Without even looking at them?

I dont want patches that stops U-boot from functioning, that is correct. If someone creates a patch which will improve functionality, tries it out on a reasonable number of targets and then it later turns out that there are some problems in corner cases, I will not call for a public hanging.

Quoted 4 lines

Btw, I can't help but feel that you're using me as an example here, since I _did_ at one point suggest that one particular patch would only need compile-testing. If you are, please stop it because you're misrepresenting what I said.

No, I was thinking of the guys that wants to change the complete support for dataflash with no intention to test it on any board with dataflash

Quoted 3 lines

Haavard

Best Regards
Ulf Samuelsson ulf at atmel.com
Atmel Nordic AB

Mail:  Box 2033, 174 02 Sundbyberg, Sweden
Visit:  Kavalleriv?gen 24, 174 58 Sundbyberg, Sweden

Phone +46 (8) 441 54 22 Fax +46 (8) 441 54 29
GSM +46 (706) 22 44 57

Technical support when I am not available:
AT89 C51 Applications Group: mailto:micro.hotline at nto.atmel.com AT90 AVR Applications Group: mailto:avr at atmel.com AT91 ARM Applications Group: mailto:at91support at atmel.com FPSLIC Application Group: mailto:fpslic at atmel.com Best AVR link: www.avrfreaks.net

#14 lore raw
Quoted 8 lines · Ulf Samuelsson
Quoted 2 lines

Yeah, that's a nice attitude. "Go away, we don't want your patches." Without even looking at them?

I dont want patches that stops U-boot from functioning, that is correct. If someone creates a patch which will improve functionality, tries it out on a reasonable number of targets and then it later turns out that there are some problems in corner cases, I will not call for a public hanging.

Sometimes you have to make some changes without being able to really test them on real hardware. That's definitely nothing we like to do, but sometimes there is no other way. Or we would keep "status quo" and I think we all agree that we want to keep this project moving, even with more pace than in the last months.

Quoted 8 lines
Quoted 4 lines

Btw, I can't help but feel that you're using me as an example here, since I _did_ at one point suggest that one particular patch would only need compile-testing. If you are, please stop it because you're misrepresenting what I said.

No, I was thinking of the guys that wants to change the complete support for dataflash with no intention to test it on any board with dataflash

That sounds a little like he (Grant) really wanted to _break_ dataflash support intentionally. His idea was to clean up some ugly places and replace them with something more readable and maintainable (Grant, please jump in if I misinterpret you here). If he is unable to test the results (compile clean is of course mandatory) then other have to "jump in" here. IIRC, he volunteered to clean up possible bugs too.

So please don't discourage people who want to improve this project. Thanks.

Best regards,
Stefan

=====================================================================
DENX Software Engineering GmbH, HRB 165235 Munich, CEO: Wolfgang Denk
Office: Kirchenstr. 5, D-82194 Groebenzell, Germany
=====================================================================

#15 lore raw
In reply to Ulf Samuelsson
Quoted 10 lines · Ulf Samuelsson
Quoted 10 lines · Ulf Samuelsson
Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

It really does not make sense to test his patches only on boards which does not use the dataflash.

If that's all he has, then he has no other choice.

He has always the choice of realizing that he maybe should leave this possible improvement to someone capable of testing it and concentrate on stuff which he can test.

Quoted 8 lines
Quoted 2 lines

Yeah, that's a nice attitude. "Go away, we don't want your patches." Without even looking at them?

I dont want patches that stops U-boot from functioning, that is correct. If someone creates a patch which will improve functionality, tries it out on a reasonable number of targets and then it later turns out that there are some problems in corner cases, I will not call for a public hanging.

I do agree with this -- all changes should be tested by _someone_. However, I don't think we should categorically reject all patches that haven't been tested by the author. If someone sees something which looks wrong, or an opportunity for optimization, not having the board in question shouldn't stop him from submitting a patch IMO. The board maintainer (or cpu/arch/subsystem maintainer, depending on the code in question) may test it himself, or send it to someone else for testing.

I disagree with your requirement that the author and the tester _always_ has to be the same person. Having done thorough testing, benchmarking, etc. and providing lots of nice numbers will of course increase the chances of acceptance, but none of it is an absolute requirement IMO.

Quoted 8 lines
Quoted 4 lines

Btw, I can't help but feel that you're using me as an example here, since I _did_ at one point suggest that one particular patch would only need compile-testing. If you are, please stop it because you're misrepresenting what I said.

No, I was thinking of the guys that wants to change the complete support for dataflash with no intention to test it on any board with dataflash

Ok, I don't know who you're talking about, then. Sorry for taking it personal ;-)

Haavard

#16 lore raw
Quoted 28 lines · Ulf Samuelsson
Quoted 10 lines
Quoted 10 lines · Ulf Samuelsson
Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

It really does not make sense to test his patches only on boards which does not use the dataflash.

If that's all he has, then he has no other choice.

He has always the choice of realizing that he maybe should leave this possible improvement to someone capable of testing it and concentrate on stuff which he can test.

Quoted 8 lines
Quoted 2 lines

Yeah, that's a nice attitude. "Go away, we don't want your patches." Without even looking at them?

I dont want patches that stops U-boot from functioning, that is correct. If someone creates a patch which will improve functionality, tries it out on a reasonable number of targets and then it later turns out that there are some problems in corner cases, I will not call for a public hanging.

I do agree with this -- all changes should be tested by _someone_. However, I don't think we should categorically reject all patches that haven't been tested by the author. If someone sees something which looks wrong, or an opportunity for optimization, not having the board in question shouldn't stop him from submitting a patch IMO. The board maintainer (or cpu/arch/subsystem maintainer, depending on the code in question) may test it himself, or send it to someone else for testing.

Quoted 6 lines

I disagree with your requirement that the author and the tester _always_ has to be the same person. Having done thorough testing, benchmarking, etc. and providing lots of nice numbers will of course increase the chances of acceptance, but none of it is an absolute requirement IMO.

I do not require that the author and the tester is the same person, If someone wants to fix something, it is perfectly OK that someone else tests the patch.

What I have seen of the dataflash interface discussion, there is no plan to test the dataflash, the goal was to get rid of the dataflash from the memory commands, write something which
appeared to replace the dataflash command, ensure it compiled and submit that.
And since the current interface is undesirable, those patches would get a fast acceptance

This plan leaves dataflash users to sort out the mess and with the current time to get patches approved, at91 support would be broken for a long time.

Best Regards
Ulf Samuelsson ulf at atmel.com
Atmel Nordic AB

Mail:  Box 2033, 174 02 Sundbyberg, Sweden
Visit:  Kavalleriv?gen 24, 174 58 Sundbyberg, Sweden

Phone +46 (8) 441 54 22 Fax +46 (8) 441 54 29
GSM +46 (706) 22 44 57

Technical support when I am not available:
AT89 C51 Applications Group: mailto:micro.hotline at nto.atmel.com AT90 AVR Applications Group: mailto:avr at atmel.com AT91 ARM Applications Group: mailto:at91support at atmel.com FPSLIC Application Group: mailto:fpslic at atmel.com Best AVR link: www.avrfreaks.net

#17 lore raw
In reply to Haavard Skinnemoen

For people that has not looked into what the three submitted patches do:

They create the same end result as if you had done the following:

1) Fix a bug in "board/cmc_pu2/at45.c" making it a true duplicate of ."board/at91rm9200dk/at45.c" (currently they differ in one line)
2) Move "board/at91rm9200dk/at45.c" to "drivers/at45.c"
3) Remove "board/cmc_pu2/at45.c".
4) Remove references to at45.c in the "board/xxx/Makefile"'s.
5) Ensure that the "drivers/Makefile" build "drivers/at45.c"
6) Move those parts of at45 which refer to at91rm9200 H/W

    to "cpu/arm920t/at91rm9200/spi.c"

7) Ensure "cpu/arm920t/at91rm9200/Makefile" builds the spi.c

The end result is that there is no duplication left in U-boot of either the at45.c nor the spi.c contents.

Best Regards
Ulf Samuelsson