[Yaffs] PATCH!!! PATCH!!! (Was Sorry state of YAFFS2)

[Sergey Kubushyn]

On Thu, 27 Oct 2005, Martin Fouts wrote:

> > OK. I don't really care 'coz I'd probably fork my own FS
> > partially based on some ideas from YAFFS/YAFFS2.
>
> Perhaps that would be best Sergey, as you don't seem the least
> interested in playing well with others.
>
> > Please don't take it as an abuse, but YAFFS2 in its present
> > state is unusable.
>
> It's not abuse, Sergey.  It's also not true.  It's less than perfect,
> and you have an application for which it is not working, but others are
> using it well in different applications.  You need to recognize that.

This application is called Linux. Yes, I do agree that there are other
applications, not involving running Linux kernel and/or syslog.

> > It reproduceably fails as a root
> > filesystem -- after several reboots (even the proper ones,
> > less for a brutal powercycling a running
> > system) it leaves the FS in a state that makes it crashing
> > kernel on initial root FS mount (the crash is in hardlinks
> > list handling code, I can even point to exact line in the
> > source.)
>
> The difference, Sergey, is that instead of whining about it, you should
> point out the exact line in the source.  Preferably in a patch. I am
> glad you send patches.  You would do yourself a favor if you always
> included enough detail to reproduce your problem, and if you refrained
> from the social commentary.

Sorry, but there is no patch for that. It looks like a design flaw, the
logical error so there is nothing to patch there...

> > I reported that bug several times, I sent a couple
> > of decoded OOPSes but nobody seems to care.
>
> You have an inappropriate set of expectations Sergey.  You sent that
> message on the 12th.  It is now the 28th.  Meanwhile, you have sent
> many
> messages that seem designed to convince the yaffs developers to have
> nothing to do with you.
>
> Here is how the real world works, Sergey:  With respect to Linux yaffs
> development, Charles is a volunteer.  He works on Yaffs out of the
> kindness of his heart.  If you wish him to be kind to you, treat him
> kindly in return.
>
> With respect to Linux yaffs development, I am a professional.
> PalmSource pays me to, among other things, work on Yaffs.  But they pay
> me to work on the bits they care about.  Unfortunately for you, those
> bits all involve 512 byte pagesize parts. Any work I do on other bits
> of
> Yaffs is voluntary, and out of the kindess of my heart.  Also,
> unfortunately for you, these days, they're asking me to concentrate
> more
> on other parts of my job.  Finally, I do not have current access to 2k
> page NAND devices, so I can't test your patches.
>
> > And the CVS tree still keeps the source that does not work at all and
> > never ever worked.
>
> So here's the deal, Sergey: You've got problems with yaffs.  They need
> fixing.  They're not, however, the most pressing issues for the people
> who do fix things in Yaffs, and you're not making it at all desirable
> for them to fix your problems.
>
> Do not expect rapid turnaround on your problems.  I read the list every
> day, and I'm aware of your issues, but I don't have 2k parts, so I
> can't
> test your fixes, and I do have a list of other items that I'm working
> on, so your problems aren't going to be on the top of my list.
>
> Do expect that the best way to get responses is to post messages with
> clear explanations of how to reproduce the problem and with patches.
> The second best way is the clear explanation without the patch.  The
> worst way is to comment negatively.
>

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