Skip to contentSkip to conversation

New mtd-utils release?

DISCUSSION Daniel Mack 8 participants 20 messages Dec 14, 2009 – Jan 19, 2010 Memory technology devices (MTD)

lore
Share & tools
lore
#0 lore raw

Hi,

we're building mtd-utils with buildroot, and as we need a recent version of it to get the ubi features, we currently check out a git snapshot as tarball. Last week, for some reason the SHA1 ID we used (e783e75e0) disappeared from the repository. (Just out of curiosity: what happened? Was the tree rebased or filtered?).

However, this approach is somewhat hackish anyway, and we would much more like to check out a version that is officially tagged. Seeing the last tag has been done ~17 month ago, I wonder whether there is any change to get a new one soon?

Thanks,
Daniel

#1 lore raw
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Daniel" == Daniel Mack <[email protected]> writes:

Hi,

Daniel> we're building mtd-utils with buildroot, and as we need a Daniel> recent version of it to get the ubi features, we currently Daniel> check out a git snapshot as tarball. Last week, for some reason Daniel> the SHA1 ID we used (e783e75e0) disappeared from the Daniel> repository. (Just out of curiosity: what happened? Was the Daniel> tree rebased or filtered?).

FYI, the e783e SHA1 must be a local change. We have been using fcb52ccc996 since April 2008. What did change recently though was that the git installation on infradead got updated recently, and gitweb >= 1.6.6-rc1 uses a slightly different snapshot tarball format than before (it appends the 7 first chars of SHA1 to directory name).

Daniel> However, this approach is somewhat hackish anyway, and we would much Daniel> more like to check out a version that is officially tagged. Seeing the Daniel> last tag has been done ~17 month ago, I wonder whether there is any Daniel> change to get a new one soon?

Me too - It would be very good to see a stable release with ubi support.

Signature · 1 line
Bye, Peter Korsgaard
#2 lore raw
In reply to Daniel Mack
Quoted 7 lines · Daniel Mack

Hi,

we're building mtd-utils with buildroot, and as we need a recent version of it to get the ubi features, we currently check out a git snapshot as tarball. Last week, for some reason the SHA1 ID we used (e783e75e0) disappeared from the repository. (Just out of curiosity: what happened? Was the tree rebased or filtered?).

Dunno, but it must be me how made this crew-up, sorry. I cannot say exactly how it happened, but it was non-intentional. Probably pushed one patch version, then later amended it, and pushed it with --force. This should never be done, sorry.

Quoted 2 lines

However, this approach is somewhat hackish anyway, and we would much more like to check out a version that is officially tagged.

OK, I can tag them at the beginning of the next year. I'm having holydays in 10 min.

Quoted 3 lines

Seeing the
last tag has been done ~17 month ago, I wonder whether there is any change to get a new one soon?

We usually use the master branch. And there is simply no one who cares about tags. But I can do this. You'll get a new tag when I come back from holydays, unless someone else tag it earlier.

Signature · 2 lines
Best Regards,
Artem Bityutskiy (Артём Битюцкий)
#3 lore raw
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Artem" == Artem Bityutskiy <[email protected]> writes:

Hi,

Quoted 2 lines
Quoted 2 lines

However, this approach is somewhat hackish anyway, and we would much more like to check out a version that is officially tagged.

Artem> OK, I can tag them at the beginning of the next year. I'm having Artem> holydays in 10 min.

Thanks - Enjoy your holidays!

Signature · 1 line
Bye, Peter Korsgaard
#4 lore raw
Quoted 11 lines · Peter Korsgaard
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Artem" == Artem Bityutskiy <[email protected]> writes:

Hi,

Quoted 2 lines
Quoted 2 lines

However, this approach is somewhat hackish anyway, and we would much more like to check out a version that is officially tagged.

Artem> OK, I can tag them at the beginning of the next year. I'm having Artem> holydays in 10 min.

Thanks - Enjoy your holidays!

I've tagged the current head with v1.3.0. I did not carefully write the change-log, though, because I was not prepared for doing the tagging. Anyway, I'll try to keep this in mind in the future and provide new tags one there is a major change, as well as better change-log.

Signature · 2 lines
Best Regards,
Artem Bityutskiy (Артём Битюцкий)
#5 lore raw
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Artem" == Artem Bityutskiy <[email protected]> writes:

Hi,

Artem> I've tagged the current head with v1.3.0. I did not carefully Artem> write the change-log, though, because I was not prepared for Artem> doing the tagging. Anyway, I'll try to keep this in mind in the Artem> future and provide new tags one there is a major change, as well Artem> as better change-log.

Great, thanks. I'll give it a spin.

Signature · 1 line
Bye, Peter Korsgaard
#6 lore raw
In reply to Artem Bityutskiy

Artem,

Quoted 4 lines · Artem Bityutskiy

I've tagged the current head with v1.3.0. I did not carefully write the change-log, though, because I was not prepared for doing the tagging. Anyway, I'll try to keep this in mind in the future and provide new tags one there is a major change, as well as better change-log.

Could you upload a tarball into ftp://ftp.infradead.org/pub/mtd-utils/? It would make the life of the distribution people easier.

Cheers,
Robert

Signature · 4 lines
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
#7 lore raw
Quoted 10 lines · Robert Schwebel

Artem,

Quoted 4 lines · Artem Bityutskiy

I've tagged the current head with v1.3.0. I did not carefully write the change-log, though, because I was not prepared for doing the tagging. Anyway, I'll try to keep this in mind in the future and provide new tags one there is a major change, as well as better change-log.

Could you upload a tarball into ftp://ftp.infradead.org/pub/mtd-utils/? It would make the life of the distribution people easier.

No sorry, I do not have access there, and I also think distributions can easily fetch from git trees, which is not a problem at all.

Signature · 2 lines
Best Regards,
Artem Bityutskiy (Артём Битюцкий)
#8 lore raw
Quoted 4 lines
Quoted 4 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Artem" == Artem Bityutskiy <[email protected]> writes:

Could you upload a tarball into ftp://ftp.infradead.org/pub/mtd-utils/? It would make the life of the distribution people easier.

Artem> No sorry, I do not have access there, and I also think distributions can Artem> easily fetch from git trees, which is not a problem at all.

Who does, David?

I DO think normal tarballs is a lot nicer than a tag somewhere in $VCS. For meta distributions like ptxdist, OE or buildroot tarballs are easier to use and mirror, and if nothing else it stops people from missing the new release when they browse ftp://ftp.infradead.org/pub/mtd-utils/.

Signature · 1 line
Bye, Peter Korsgaard
#9 lore raw
Quoted 9 lines · Peter Korsgaard
Quoted 4 lines
Quoted 4 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Artem" == Artem Bityutskiy <[email protected]> writes:

Could you upload a tarball into ftp://ftp.infradead.org/pub/mtd-utils/? It would make the life of the distribution people easier.

Artem> No sorry, I do not have access there, and I also think distributions can Artem> easily fetch from git trees, which is not a problem at all.

Who does, David?

Artem, Thomas, Josh and myself.

Quoted 5 lines

I DO think normal tarballs is a lot nicer than a tag somewhere in $VCS. For meta distributions like ptxdist, OE or buildroot tarballs are easier to use and mirror, and if nothing else it stops people from missing the new release when they browse ftp://ftp.infradead.org/pub/mtd-utils/.

Yeah, I agree.

Signature · 2 lines
David Woodhouse                            Open Source Technology Centre
[email protected]                              Intel Corporation
#10 lore raw
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"David" == David Woodhouse <[email protected]> writes:

Hi,

Quoted 3 lines
Quoted 3 lines
Quoted 2 lines
Quoted 2 lines

Could you upload a tarball into ftp://ftp.infradead.org/pub/mtd-utils/? It would make the life of the distribution people easier.

Artem> No sorry, I do not have access there, and I also think distributions can Artem> easily fetch from git trees, which is not a problem at all.

Quoted 2 lines
Quoted 2 lines

Who does, David?

David> Artem, Thomas, Josh and myself.

It would be good if one of you guys could put 1.3.1 there as well.

Signature · 1 line
Bye, Peter Korsgaard
#11 lore raw
In reply to Peter Korsgaard
Quoted 9 lines · Peter Korsgaard
Quoted 4 lines
Quoted 4 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Artem" == Artem Bityutskiy <[email protected]> writes:

Could you upload a tarball into ftp://ftp.infradead.org/pub/mtd-utils/? It would make the life of the distribution people easier.

Artem> No sorry, I do not have access there, and I also think distributions can Artem> easily fetch from git trees, which is not a problem at all.

Who does, David?

Dunno.

Quoted 5 lines

I DO think normal tarballs is a lot nicer than a tag somewhere in $VCS. For meta distributions like ptxdist, OE or buildroot tarballs are easier to use and mirror, and if nothing else it stops people from missing the new release when they browse ftp://ftp.infradead.org/pub/mtd-utils/.

May be.

However, I do not want to be impolite, but I just do not want to do that. I someone else wants - no problem. I just do not volunteer to maintain the FTP directory with the tarballs, I only volunteer to provide the tags. Apologies.

Signature · 2 lines
Best Regards,
Artem Bityutskiy (Артём Битюцкий)
#12 lore raw
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Artem" == Artem Bityutskiy <[email protected]> writes:

Hi,

Quoted 5 lines
Quoted 5 lines

I DO think normal tarballs is a lot nicer than a tag somewhere in $VCS. For meta distributions like ptxdist, OE or buildroot tarballs are easier to use and mirror, and if nothing else it stops people from missing the new release when they browse ftp://ftp.infradead.org/pub/mtd-utils/.

Artem> May be.

Artem> However, I do not want to be impolite, but I just do not want to do Artem> that. I someone else wants - no problem. I just do not volunteer to Artem> maintain the FTP directory with the tarballs, I only volunteer to Artem> provide the tags. Apologies.

Ok, I see that someone (David?) already uploaded a tarball.

Thanks, both.

Signature · 1 line
Bye, Peter Korsgaard
#13 lore raw
In reply to Peter Korsgaard
Quoted 2 lines · Peter Korsgaard

missing the new release when they browse
ftp://ftp.infradead.org/pub/mtd-utils/.

Fwiw, I'm one of those people who browsers the FTP directory occasionally to get whatever appears to be the latest version, and would not have known to pick up the new release without the new tarball.

That's because my package build scripts, much like other distros (Ptxdist, OE, uClinux etc.), include a URL to the original source they are using, and to update the distro, it's natural to look in the same place for new versions.

If an FTP/HTTP directory _was_ apparently the primary location of some package, and then ceases to be kept up to date because it's moved or because git has taken over, at least please put a README or similar on the old site saying that the primary location has moved elsewhere.

However, in this case it appears that the FTP directory is still being kept up to date, so there is no need.

Thanks a lot,
-- Jamie

#14 lore raw
In reply to Artem Bityutskiy
Quoted 20 lines · Artem Bityutskiy
Quoted 7 lines · Daniel Mack

Hi,

we're building mtd-utils with buildroot, and as we need a recent version of it to get the ubi features, we currently check out a git snapshot as tarball. Last week, for some reason the SHA1 ID we used (e783e75e0) disappeared from the repository. (Just out of curiosity: what happened? Was the tree rebased or filtered?).

Dunno, but it must be me how made this crew-up, sorry. I cannot say exactly how it happened, but it was non-intentional. Probably pushed one patch version, then later amended it, and pushed it with --force. This should never be done, sorry.

Quoted 2 lines

However, this approach is somewhat hackish anyway, and we would much more like to check out a version that is officially tagged.

OK, I can tag them at the beginning of the next year. I'm having holydays in 10 min.

Quoted 7 lines
Quoted 3 lines

Seeing the
last tag has been done ~17 month ago, I wonder whether there is any change to get a new one soon?

We usually use the master branch. And there is simply no one who cares about tags. But I can do this. You'll get a new tag when I come back from holydays, unless someone else tag it earlier.

I've avoided doing a release because there isn't a really good way to correlate mtd-utils to kernel version. I don't believe there have been any on-disk (Ha!) UBIFS format changes, and UBI itself hasn't had format changes for a long time, so it is probably safe to do that now.

Still, I would very much like to avoid having a compatiblity matrix like the LTTng project does.

(Also, I believe nandwrite is _STILL_ broken with the OOB stuff.)

josh

#15 lore raw
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Josh" == Josh Boyer <[email protected]> writes:

Hi,

Josh> I've avoided doing a release because there isn't a really good Josh> way to correlate mtd-utils to kernel version. I don't believe Josh> there have been any on-disk (Ha!) UBIFS format changes, and UBI Josh> itself hasn't had format changes for a long time, so it is Josh> probably safe to do that now.

Well, you'll have those version issues anyway - With known-good git SHA1's or releases. Adding a release would just make more people use the same revision, instead of people using git head of whatever day they needed mtd-utils.

Josh> (Also, I believe nandwrite is _STILL_ broken with the OOB stuff.)

Ok, what is the issue? Anyone working on fixing it? Did it work back in 1.2.0?

Signature · 1 line
Bye, Peter Korsgaard
#16 lore raw
Quoted 10 lines · Peter Korsgaard
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

"Josh" == Josh Boyer <[email protected]> writes:

Hi,

Josh> I've avoided doing a release because there isn't a really good Josh> way to correlate mtd-utils to kernel version. I don't believe Josh> there have been any on-disk (Ha!) UBIFS format changes, and UBI Josh> itself hasn't had format changes for a long time, so it is Josh> probably safe to do that now.

Quoted 4 lines

Well, you'll have those version issues anyway - With known-good git SHA1's or releases. Adding a release would just make more people use the same revision, instead of people using git head of whatever day they needed mtd-utils.

This is true. I just feel better about saying 'use git head' or 'upgrade your kernel' than I do trying to come up with some kind of mapping.

Quoted 4 lines

Josh> (Also, I believe nandwrite is _STILL_ broken with the OOB stuff.)

Ok, what is the issue? Anyone working on fixing it? Did it work back in 1.2.0?

The MEMSETOOBSEL ioctl no longer exists in the kernel. It worked in 1.2.0 assuming you were using a sufficiently old enough kernel. I can't recall if anyone is working on fixing it or not, but Thomas has mentioned that there is a patch floating around somewhere.

josh

#17 lore raw

OMAP2 issue while compiling the latest mtd-2.6 git kernel

Hi ,

I am facing the following problem while compiling the mtd2.6 git kernel . I used the arch/arm/configs/omap_apollon_2420_defconfig as the config file and used the arm-2009q1/arm-none-linux-gnueabi- compiler for compiling.

Log :

# make
  CHK     include/linux/version.h
make[1]: `include/asm-arm/mach-types.h' is up to date.
  CHK     include/linux/utsrelease.h
  SYMLINK include/asm -> include/asm-arm
  CALL    scripts/checksyscalls.sh
<stdin>:1351:2: warning: #warning syscall migrate_pages not implemented
  CHK     include/linux/compile.h
  CC      arch/arm/mach-omap2/devices.o
arch/arm/mach-omap2/devices.c: In function 'omap_init_mcspi':
arch/arm/mach-omap2/devices.c:365: error: 'omap2_mcspi3_resources' undeclared (first use in this 
function)
arch/arm/mach-omap2/devices.c:365: error: (Each undeclared identifier is reported only once
arch/arm/mach-omap2/devices.c:365: error: for each function it appears in.)
arch/arm/mach-omap2/devices.c:367: error: 'omap2_mcspi4_resources' undeclared (first use in this 
function)
make[1]: *** [arch/arm/mach-omap2/devices.o] Error 1
make: *** [arch/arm/mach-omap2] Error 2

Regards,
Sarath.

#18 lore raw

Re: OMAP2 issue while compiling the latest mtd-2.6 git kernel

Quoted 17 lines · Tadimarri Sarath Babu

Hi ,

I am facing the following problem while compiling the mtd2.6 git kernel . I used the arch/arm/configs/omap_apollon_2420_defconfig as the config file and used the arm-2009q1/arm-none-linux-gnueabi- compiler for compiling.

Log :

# make
  CHK     include/linux/version.h
make[1]: `include/asm-arm/mach-types.h' is up to date.
  CHK     include/linux/utsrelease.h
  SYMLINK include/asm -> include/asm-arm
  CALL    scripts/checksyscalls.sh
<stdin>:1351:2: warning: #warning syscall migrate_pages not implemented
  CHK     include/linux/compile.h
  CC      arch/arm/mach-omap2/devices.o
arch/arm/mach-omap2/devices.c: In function 'omap_init_mcspi':
Quoted 8 lines
arch/arm/mach-omap2/devices.c:365: error: 'omap2_mcspi3_resources' undeclared (first use in this 
function)
arch/arm/mach-omap2/devices.c:365: error: (Each undeclared identifier is reported only once
arch/arm/mach-omap2/devices.c:365: error: for each function it appears in.)
arch/arm/mach-omap2/devices.c:367: error: 'omap2_mcspi4_resources' undeclared (first use in this 
function)
make[1]: *** [arch/arm/mach-omap2/devices.o] Error 1
make: *** [arch/arm/mach-omap2] Error 2

You should CC the OMAP mailing list and the OMAP maintainer. See the MAINTAINERS file.

Signature · 2 lines
Best Regards,
Artem Bityutskiy (Артём Битюцкий)
#19 lore raw
In reply to Artem Bityutskiy
Quoted 11 lines · Artem Bityutskiy
Quoted 5 lines · Daniel Mack

we're building mtd-utils with buildroot, and as we need a recent version of it to get the ubi features, we currently check out a git snapshot as tarball. Last week, for some reason the SHA1 ID we used (e783e75e0) disappeared from the repository. (Just out of curiosity: what happened? Was the tree rebased or filtered?).

Dunno, but it must be me how made this crew-up, sorry. I cannot say exactly how it happened, but it was non-intentional. Probably pushed one patch version, then later amended it, and pushed it with --force. This should never be done, sorry.

No big thing, I was just wondering. Probably the commit was dead (ie, not lined-up with an exposed head) for a longer time and just got vacuumed last week, dunno.

Quoted 6 lines
Quoted 2 lines

However, this approach is somewhat hackish anyway, and we would much more like to check out a version that is officially tagged.

OK, I can tag them at the beginning of the next year. I'm having holydays in 10 min.

Quoted 7 lines
Quoted 3 lines

Seeing the
last tag has been done ~17 month ago, I wonder whether there is any change to get a new one soon?

We usually use the master branch. And there is simply no one who cares about tags. But I can do this. You'll get a new tag when I come back from holydays, unless someone else tag it earlier.

Great, thank you. The background of my question was also whether there is a version which is reliable, in terms of representing a stable set of features, defining a larger merge of new features and the like, so it deserves the term 'release' :)

Not saying that any of the snapshops caused any trouble so far, it would just feel better to check out something that has been tagged by the maintainers.

Thanks,
Daniel