Skip to contentSkip to conversation

treat OOB as a single chunk of oobavail bytes

PATCH Sergey Kubushyn 4 participants 6 messages Jan 30, 2006 – Jan 31, 2006 1 message not in the archive Memory technology devices (MTD)

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

On Sun, 29 Jan 2006, Vitaly Wool wrote:

Guys, can you explain what is the need for such a change in the first place? Another round of confusion? Forcing everybody to upgrade their painfully crafted kernels (your CVS version won't apply to e.g. 2.6.12 kernel that was the last one before rather major change in ARM architecture, will it?)?

Why don't just fix YAFFS2 itself to make it work with a stock kernel? It's easy to do and the appropriate patch is out for quite a long time, tested and works for everyone. Patch for YAFFS2 it is, not for kernel.

Occam was a wise man so it does make sence to use his razor... Belly cover weared without a need is harmful (C) K.Prutkov ....

Quoted 65 lines

Hi Charles,

no it's not yet in CVS. I think I'll be able to commit it if we get any positive feedback on this patch (i. e. if it's actually working for anyone else than me :))
As for write_oobfree, it looks to be useful for JFFSx to stop messing with oobinfos in JFFSx code.

Best regards,
Vitaly

Quoted 10 lines · Charles Manning
Quoted 7 lines · Vitaly Wool

Hi Yuri,

well, lemme just summarize what you hafta do in order to make YAFFS2 work with my patches.

1. Apply the following patch to the mtd code:

http://lists.infradead.org/pipermail/linux-mtd/2005-December/014648.html

Quoted 2 lines
Quoted 2 lines

(yes, it's a single patch!)
2. Modify the fs/yaffs2/yaffs_mtdif2.c to use

read_oobfree/write_oobfree

Quoted 8 lines
Quoted 6 lines

where appropriate.

Hope that helps,

Vitaly

Quoted 8 lines

Hi Vitaly

Is this in CVS yet? If not, I hope it will be there soon.

From what I see, this is going to be just what we need.

YAFFS2 does not need to write_oobfree as a seperate function (it is always written as a single page write). YAFFS2 only needs

read_oobfree.

Quoted 4 lines

However, YAFFS1 could use the write_oobfree function at some future date if YAFFS1 gets moved to this interface to remove the current reliance on hanging the ugly ECC parameter stuff through the

interface.

Quoted 3 lines

I am pretty busy right now, but will try get the YAFFS2 side sorted ASAP. If the YAFFS2 stuff is CVSed before the mtd then I'll just

point

Quoted 9 lines

people to this patch.

Thanx!

Charles

Quoted 5 lines

_______________________________________________
yaffs mailing list
[email protected]
http://stoneboat.aleph1.co.uk/cgi-bin/mailman/listinfo/yaffs

---
******************************************************************

*  KSI@home    KOI8 Net  < >  The impossible we do immediately.  *
*  Las Vegas   NV, USA   < >  Miracles require 24-hour notice.   *

******************************************************************

#2 lore raw

Hello Sergey,

I'm afraid I haven't got your allusions... Anyway, you might want to learn more about main principles of modularity.

Best regards,
Vitaly

Quoted 120 lines · Sergey Kubushyn

On Sun, 29 Jan 2006, Vitaly Wool wrote:

Guys, can you explain what is the need for such a change in the first place? Another round of confusion? Forcing everybody to upgrade their painfully crafted kernels (your CVS version won't apply to e.g. 2.6.12 kernel that was the last one before rather major change in ARM architecture, will it?)?

Why don't just fix YAFFS2 itself to make it work with a stock kernel? It's easy to do and the appropriate patch is out for quite a long time, tested and works for everyone. Patch for YAFFS2 it is, not for kernel.

Occam was a wise man so it does make sence to use his razor... Belly cover weared without a need is harmful (C) K.Prutkov ....

Quoted 87 lines

Hi Charles,

no it's not yet in CVS. I think I'll be able to commit it if we get any positive feedback on this patch (i. e. if it's actually working for anyone else than me :))
As for write_oobfree, it looks to be useful for JFFSx to stop messing with oobinfos in JFFSx code.

Best regards,
Vitaly

Quoted 7 lines · Charles Manning
Quoted 2 lines · Vitaly Wool

Hi Yuri,

Quoted 7 lines
Quoted 7 lines

well, lemme just summarize what you hafta do in order to make YAFFS2 work with my patches.

1. Apply the following patch to the mtd code:

http://lists.infradead.org/pipermail/linux-mtd/2005-December/014648.html

Quoted 4 lines
Quoted 4 lines

(yes, it's a single patch!)
2. Modify the fs/yaffs2/yaffs_mtdif2.c to use

read_oobfree/write_oobfree

Quoted 8 lines
Quoted 8 lines

where appropriate.

Hope that helps,

Vitaly

Quoted 10 lines

Hi Vitaly

Is this in CVS yet? If not, I hope it will be there soon.

From what I see, this is going to be just what we need.

YAFFS2 does not need to write_oobfree as a seperate function (it is always written as a single page write). YAFFS2 only needs

read_oobfree.

Quoted 5 lines

However, YAFFS1 could use the write_oobfree function at some future date if YAFFS1 gets moved to this interface to remove the current reliance on hanging the ugly ECC parameter stuff through the

interface.

Quoted 4 lines

I am pretty busy right now, but will try get the YAFFS2 side sorted ASAP. If the YAFFS2 stuff is CVSed before the mtd then I'll just

point

Quoted 11 lines

people to this patch.

Thanx!

Charles

Quoted 7 lines

_______________________________________________
yaffs mailing list
[email protected]
http://stoneboat.aleph1.co.uk/cgi-bin/mailman/listinfo/yaffs

---
******************************************************************

*  KSI@home    KOI8 Net  < >  The impossible we do immediately.  *
*  Las Vegas   NV, USA   < >  Miracles require 24-hour notice.   *

******************************************************************

#3 lore raw

On Mon, 30 Jan 2006, Vitaly Wool wrote:

You guys are creating unneeded entities. There is no need to change the kernel for YAFFS2 to work. It can be done with a rather simple mtdiff2.c rewrite in YAFFS2 tree that took me less than a day to do. Furthermore, my patch is in the list so you can even save that day by simply applying it and make the thing work with a _stock_ kernel, right now.

You're taking a long path instead adding unnecessary functionality to Linux kernel. How d'ya think, how much time will it take for all those CVS MTD changes to make it into the mainstream kernel? And you can get YAFFS2 working for everybody with _any_ stock kernel right now by just applying the patch to mtdiff2.c.

There is another unnecessary complication -- your CVS MTD changes would go into something like, say, 2.6.22 kernel and they would not fit into the older ones. That means that everybody will be forced to switch to a newer kernel that is not such an easy task for those who use highly customized kernels they are happy with. E.g. ARM architecture underwent quite a substantial changes after 2.6.12 so it is NOT just a matter of easy patch fitting into a new kernel, it's quite a rewrite.

Furthermore, there are places where those kernels undergo some kind of regulatory body certification (e.g. Gaming Board or DOD) that takes significant time and effort to do over and over again. They simply don't like any updates, once the code is certified it's effectively frozen.

Unfortunately enough for YAFFS it's not driven by a reason or common sence, personal likes and dislikes are ruling here... Somebody got insulted so he rejects all reasonable arguments just because they come from a person he doesn't like. That might be fine for preschool kids, but we're supposedly professionals here, aren't we? And this is simply ridiculous for a professional to behave in such a way...

Quoted 59 lines

Hello Sergey,

I'm afraid I haven't got your allusions... Anyway, you might want to learn more about main principles of modularity.

Best regards,
Vitaly

Quoted 3 lines · Sergey Kubushyn

On Sun, 29 Jan 2006, Vitaly Wool wrote:

Guys, can you explain what is the need for such a change in the first

place?

Quoted 1 line

Another round of confusion? Forcing everybody to upgrade their

painfully

Quoted 1 line

crafted kernels (your CVS version won't apply to e.g. 2.6.12 kernel

that was

Quoted 1 line

the last one before rather major change in ARM architecture, will

it?)?

Quoted 2 lines

Why don't just fix YAFFS2 itself to make it work with a stock kernel?

It's

Quoted 1 line

easy to do and the appropriate patch is out for quite a long time,

tested

Quoted 3 lines

and works for everyone. Patch for YAFFS2 it is, not for kernel.

Occam was a wise man so it does make sence to use his razor... Belly

cover

Quoted 7 lines

weared without a need is harmful (C) K.Prutkov ....

Quoted 3 lines

Hi Charles,

no it's not yet in CVS. I think I'll be able to commit it if we get

any

Quoted 3 lines
Quoted 3 lines

positive feedback on this patch (i. e. if it's actually working for anyone else than me :))
As for write_oobfree, it looks to be useful for JFFSx to stop

messing

Quoted 9 lines
Quoted 9 lines

with oobinfos in JFFSx code.

Best regards,
Vitaly

Charles Manning wrote:

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines · Vitaly Wool
Quoted 7 lines

Hi Yuri,

well, lemme just summarize what you hafta do in order to make

YAFFS2

Quoted 83 lines
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
Quoted 6 lines

work with my patches.

1. Apply the following patch to the mtd code:

http://lists.infradead.org/pipermail/linux-mtd/2005-December/014648.ht

ml

Quoted 74 lines
Quoted 57 lines
Quoted 6 lines
Quoted 6 lines

(yes, it's a single patch!)
2. Modify the fs/yaffs2/yaffs_mtdif2.c to use

read_oobfree/write_oobfree

Quoted 8 lines
Quoted 8 lines

where appropriate.

Hope that helps,

Vitaly

Quoted 10 lines

Hi Vitaly

Is this in CVS yet? If not, I hope it will be there soon.

From what I see, this is going to be just what we need.

YAFFS2 does not need to write_oobfree as a seperate function (it is always written as a single page write). YAFFS2 only needs

read_oobfree.

Quoted 5 lines

However, YAFFS1 could use the write_oobfree function at some future date if YAFFS1 gets moved to this interface to remove the current reliance on hanging the ugly ECC parameter stuff through the

interface.

Quoted 4 lines

I am pretty busy right now, but will try get the YAFFS2 side sorted ASAP. If the YAFFS2 stuff is CVSed before the mtd then I'll just

point

Quoted 11 lines

people to this patch.

Thanx!

Charles

Quoted 7 lines

_______________________________________________
yaffs mailing list
[email protected]
http://stoneboat.aleph1.co.uk/cgi-bin/mailman/listinfo/yaffs

---
******************************************************************

*  KSI@home    KOI8 Net  < >  The impossible we do immediately.  *
*  Las Vegas   NV, USA   < >  Miracles require 24-hour notice.   *

******************************************************************

---
******************************************************************

*  KSI@home    KOI8 Net  < >  The impossible we do immediately.  *
*  Las Vegas   NV, USA   < >  Miracles require 24-hour notice.   *

******************************************************************

#4 lore raw

Can all people involved in this thread please read

http://www.infradead.org/~dwmw2/email.html

and the famous writeup "Why is top posting bad" which dates back to the start of usenet.

If you want that people responsible for that stuff can follow this discussions in their _SPARE_ time, then please comply to the rules.

Otherwise do not complain, when it is ignored

tglx

#5 lore raw
In reply to Sergey Kubushyn

+++ Sergey Kubushyn [06-01-30 09:53 -0800]:

Quoted 7 lines · Vitaly Wool

You guys are creating unneeded entities. There is no need to change the kernel for YAFFS2 to work. It can be done with a rather simple mtdiff2.c rewrite in YAFFS2 tree that took me less than a day to do. Furthermore, my patch is in the list so you can even save that day by simply applying it and make the thing work with a _stock_ kernel, right now.

You are wrong. There _is_ a need to change the kernel (MTD interface), in order to fix this problem properly. Your patch is a useful hack to work around a problem, but it is not a real long-term solution.

One of the significant advantages of free software is that it is always possible to get to a real engineering solution, not bodge in some hack to satisfy marketing and release schedules. Clearly there are disadvantages in working towards this proper solution (a clean and correct interface), which you list below (it takes longer, it doesn't necessarily solve the problem for older versions of Linux/MTD), but it is still better than not solving the problem, and putting in a workaround forevermore.

This has been explained before, and it's tiresome that you still don't seem to understand, so I'm having to explain it again.

see:
http://www.aleph1.co.uk/pipermail/yaffs/2005q4/001589.html
(the definitive history and explanation of why AUTOPLACE makes sense)

I also listed all the significant relevant threads I could find in this message:
http://lists.aleph1.co.uk/pipermail/yaffs/2006q1/001811.html

Quoted 2 lines

You're taking a long path instead adding unnecessary functionality to Linux kernel.

The functionality is not unnecessary. It was specified back in 2003 when the AUTOPLACE support in MTD for NAND for first worked out, but never actually made it into MTD. This was a mistake, but it took time to come to light, and realise its significance.

It is sensible to have a mechanism in MTD (AUTPLACE) that allows MTD clients to use the oob area without knowing the detailed layout of the data stored in there, given the incompatible requirements of different hardware and software ecc mechanisms, and different technologies and filesystems.

It is obviously not sensible to implement this mechanism in a way that allows reading and writing of data+oob but not oob only, which is the current state of afairs. For a filesystem like YAFFS, which is designed to read/write the oob without having to also read the coresponding data, this omission is serious.

As you say, your patch is a useful workaround for the time being, for which we are grateful, but its existence does not mean that a proper solution is not needed.

If you still cannot see this then please re-read the relevant YAFFS and MTD threads, or just accept that all the other important players in both YAFFS and MTD development disagree with you. It does not help anyone for you to keep repeating the same assertion that your way is right and our way is wrong. You have at least been reasonably civil this this time round, for which I am grateful, but you are still being part of the problem rather than part of the solution, because we are having to repeat old arguments rather than doing something more constructive.

Quoted 7 lines

There is another unnecessary complication -- your CVS MTD changes would go into something like, say, 2.6.22 kernel and they would not fit into the older ones. That means that everybody will be forced to switch to a newer kernel that is not such an easy task for those who use highly customized kernels they are happy with. E.g. ARM architecture underwent quite a substantial changes after 2.6.12 so it is NOT just a matter of easy patch fitting into a new kernel, it's quite a rewrite.

You are correct that this is a problem. I hope it will not be too difficult to use the new OOB autoplace with older kernels. If not then people will have to make do with some other solution, such as your patch. But not-fixing it for even longer would not improve matters. We were hoping this would be done a year ago, but sometimes things don't go the way one would like.

Quoted 4 lines

Unfortunately enough for YAFFS it's not driven by a reason or common sence, personal likes and dislikes are ruling here... Somebody got insulted so he rejects all reasonable arguments just because they come from a person he doesn't like.

Nope, that's simply untrue. Everyone involved apart from you is persuing (or waiting eagerly for) the correcting of the MTD API because _it is the right technical solution_. It has nothing to do with the fact that many people here think you are a tiresome idiot that I should ban from the list. I have not done that because I think civil discussion is important, and I still hope that one day you will understand this issue properly and stop calling us names.

Quoted 2 lines

That might be fine for preschool kids, but we're supposedly professionals here, aren't we?

We are, and we should behave as such. We should look at the design; read the discussion and consider the arguments; understand the benefits of modularity, hiding details behind stable interfaces, and design for the long term. We should know the difference between a quick-and-dirty fix and proper design. We should remain polite at all times, and we should accept a majority descision even if we don't agree with it.

Now can you please drop this pet peeve so I don't have to explain this _again_ in 3 months time?

Wookey

Signature · 2 lines
Aleph One Ltd, Bottisham, CAMBRIDGE, CB5 9BA, UK  Tel +44 (0) 1223 811679
work: http://www.aleph1.co.uk/     play: http://www.chaos.org.uk/~wookey/
#6 lore raw
Quoted 9 lines · Wookey

+++ Sergey Kubushyn [06-01-30 09:53 -0800]:

Quoted 3 lines · Vitaly Wool

You guys are creating unneeded entities. There is no need to change

the

Quoted 1 line

kernel for YAFFS2 to work. It can be done with a rather simple

mtdiff2.c

Quoted 1 line

rewrite in YAFFS2 tree that took me less than a day to do.

Furthermore, my

Quoted 8 lines
Quoted 1 line

patch is in the list so you can even save that day by simply applying

it and

Quoted 1 line

make the thing work with a _stock_ kernel, right now.

You are wrong. There _is_ a need to change the kernel (MTD interface), in
order to fix this problem properly. Your patch is a useful hack to work around a problem, but it is not a real long-term solution.

You did NOT look into that patch, did ya? It is NOT a hack, it uses proper MTD API and makes YAFFS follow the specs.

Quoted 13 lines

One of the significant advantages of free software is that it is always possible to get to a real engineering solution, not bodge in some hack to
satisfy marketing and release schedules. Clearly there are disadvantages in
working towards this proper solution (a clean and correct interface), which
you list below (it takes longer, it doesn't necessarily solve the problem
for older versions of Linux/MTD), but it is still better than not solving the
problem, and putting in a workaround forevermore.

Quoted 3 lines

This has been explained before, and it's tiresome that you still don't seem
to understand, so I'm having to explain it again.

It is NOT a workaround! This is following specs!

Quoted 7 lines

see:
http://www.aleph1.co.uk/pipermail/yaffs/2005q4/001589.html
(the definitive history and explanation of why AUTOPLACE makes sense)

I also listed all the significant relevant threads I could find in this message:
http://lists.aleph1.co.uk/pipermail/yaffs/2006q1/001811.html

NO, all that does NOT prove that kernel changes are necessary. That shows that had those changes were made the life of YAFFS developers would've been a little bit easier.

Quoted 18 lines
Quoted 1 line

You're taking a long path instead adding unnecessary functionality to

Linux

Quoted 1 line

kernel.

The functionality is not unnecessary. It was specified back in 2003 when
the AUTOPLACE support in MTD for NAND for first worked out, but never actually made it into MTD. This was a mistake, but it took time to come to
light, and realise its significance.

It is sensible to have a mechanism in MTD (AUTPLACE) that allows MTD clients
to use the oob area without knowing the detailed layout of the data stored
in there, given the incompatible requirements of different hardware and software ecc mechanisms, and different technologies and filesystems.

Quoted 8 lines

It is obviously not sensible to implement this mechanism in a way that allows
reading and writing of data+oob but not oob only, which is the current state
of afairs. For a filesystem like YAFFS, which is designed to read/write the
oob without having to also read the coresponding data, this omission is serious.

No, it's not. The existing MTD API provides you with the OOB _LAYOUT_ so you can use those bytes not taken by anything else. It is simple as a shovel, you only need one additional step to read/write data using that layout. Everybody else (JFFSx etc) is already doing that. Why YAFFS is so different and exclusive that it can't do such a trivial thing itself is out of my understanding. You have a power outlet that everybody else uses without any problems. One of the prongs in your plug is slightly bent so it doesn't fit that outlet. Every other intellectual person would've straightened that slightly bent prong and forgot about the entire issue forever. But you guys don't wanna do such a simple thing, you're blaming the outlet!

Quoted 6 lines

As you say, your patch is a useful workaround for the time being, for which
we are grateful, but its existence does not mean that a proper solution is
not needed.

Quoted 14 lines

If you still cannot see this then please re-read the relevant YAFFS and MTD
threads, or just accept that all the other important players in both YAFFS
and MTD development disagree with you. It does not help anyone for you to
keep repeating the same assertion that your way is right and our way is wrong. You have at least been reasonably civil this this time round, for
which I am grateful, but you are still being part of the problem rather than
part of the solution, because we are having to repeat old arguments rather
than doing something more constructive.

No, I can't understand neither you guys not those "important players"... I have a strong feeling that it's not for real, that we're in some kind of a movie like "Hot Shots" where they suffered because their shoe laces were tied together...

Quoted 13 lines
Quoted 1 line

There is another unnecessary complication -- your CVS MTD changes

would go

Quoted 1 line

into something like, say, 2.6.22 kernel and they would not fit into

the

Quoted 1 line

older ones. That means that everybody will be forced to switch to a

newer

Quoted 1 line

kernel that is not such an easy task for those who use highly

customized

Quoted 2 lines

kernels they are happy with. E.g. ARM architecture underwent quite a substantial changes after 2.6.12 so it is NOT just a matter of easy

patch

Quoted 1 line

fitting into a new kernel, it's quite a rewrite.

Quoted 9 lines

You are correct that this is a problem. I hope it will not be too difficult
to use the new OOB autoplace with older kernels. If not then people will
have to make do with some other solution, such as your patch. But not-fixing
it for even longer would not improve matters. We were hoping this would be
done a year ago, but sometimes things don't go the way one would like.

And in the meantime you kept in CVS the source that NEVER EVER worked with ANY kernel. Nice touch.... Especially with "Everything's fine, we're of production quality" drumming.

Quoted 8 lines
Quoted 1 line

Unfortunately enough for YAFFS it's not driven by a reason or common

sence,

Quoted 1 line

personal likes and dislikes are ruling here... Somebody got insulted

so he

Quoted 1 line

rejects all reasonable arguments just because they come from a person

he

Quoted 1 line

doesn't like.

Quoted 13 lines

Nope, that's simply untrue. Everyone involved apart from you is persuing (or
waiting eagerly for) the correcting of the MTD API because _it is the right
technical solution_. It has nothing to do with the fact that many people
here think you are a tiresome idiot that I should ban from the list. I have
not done that because I think civil discussion is important, and I still
hope that one day you will understand this issue properly and stop calling
us names.

I'm not calling you names. I simply can't understand why you just won't untie your shoe laces. I'm not even trying to insult anybody, I genuinely can't understand. I would've even understood if it were difficult to do, but it is SUCH a trivial thing that I can't understand it...

Quoted 13 lines
Quoted 2 lines

That might be fine for preschool kids, but we're supposedly professionals here, aren't we?

We are, and we should behave as such. We should look at the design; read the
discussion and consider the arguments; understand the benefits of modularity, hiding details behind stable interfaces, and design for the long
term. We should know the difference between a quick-and-dirty fix and proper
design. We should remain polite at all times, and we should accept a majority descision even if we don't agree with it.

Quoted 2 lines

Now can you please drop this pet peeve so I don't have to explain this _again_ in 3 months time?

Yeah, I'd better go for a time being... It's simply unbelievable...

---
******************************************************************

*  KSI@home    KOI8 Net  < >  The impossible we do immediately.  *
*  Las Vegas   NV, USA   < >  Miracles require 24-hour notice.   *

******************************************************************