Hi Cyrille,
Fetching the spi-nor tree
(git://git.infradead.org/l2-mtd.git#spi-nor/next) this morning, I get
theses errors:
remote: error: Could not read 6e408e3aa1643624ea7da50ad5b27a226a16a654
remote: fatal: bad tree object 6e408e3aa1643624ea7da50ad5b27a226a16a654
remote: aborting due to possible repository corruption on the remote side.
fatal: protocol error: bad pack header
I wil continue to use the previously fetched tree.
--
Cheers,
Stephen Rothwell
Hi Stephen,
I've just added 2 commits about 8 hours ago, nothing special.
Could the issue be explained by an unlikely bad timing?
I don't see anything wrong when fetching the tree or when browsing the
web server:
http://git.infradead.org/l2-mtd.git/shortlog/refs/heads/spi-nor/next
Please, let me know whether it was a spurious error or if it still doesn't
work. Currently, I don't know what's going on.
Anyway, thanks for your report.
Best regards,
Cyrille
Le 12/12/2017 à 21:58, Stephen Rothwell a écrit :
> Hi Cyrille,
>
> Fetching the spi-nor tree
> (git://git.infradead.org/l2-mtd.git#spi-nor/next) this morning, I get
> theses errors:
>
> remote: error: Could not read 6e408e3aa1643624ea7da50ad5b27a226a16a654
> remote: fatal: bad tree object 6e408e3aa1643624ea7da50ad5b27a226a16a654
> remote: aborting due to possible repository corruption on the remote side.
> fatal: protocol error: bad pack header
>
> I wil continue to use the previously fetched tree.
>
Hi Cyrille,
[cc'ing Linus on the off chance he knows something relevant]
On Tue, 12 Dec 2017 22:24:17 +0100 Cyrille Pitchen <[email protected]> wrote:
>
> I've just added 2 commits about 8 hours ago, nothing special.
> Could the issue be explained by an unlikely bad timing?
No idea. The current top commit in my copy of your branch is:
b3fbf5e23f8c ("mtd: spi-nor: Check that BP bits are set properly")
> I don't see anything wrong when fetching the tree or when browsing the
> web server:
> http://git.infradead.org/l2-mtd.git/shortlog/refs/heads/spi-nor/next
>
> Please, let me know whether it was a spurious error or if it still doesn't
> work. Currently, I don't know what's going on.
I am still getting the error. I even tried cloning your tree and got
the same error.
> Le 12/12/2017 à 21:58, Stephen Rothwell a écrit :
> > Hi Cyrille,
> >
> > Fetching the spi-nor tree
> > (git://git.infradead.org/l2-mtd.git#spi-nor/next) this morning, I get
> > theses errors:
> >
> > remote: error: Could not read 6e408e3aa1643624ea7da50ad5b27a226a16a654
This SHA1 refers to a tree object (representing the drivers directory)
that exists in my tree somewhere (I don't know how to figure out which
commit it is related to (if any)).
> > remote: fatal: bad tree object 6e408e3aa1643624ea7da50ad5b27a226a16a654
> > remote: aborting due to possible repository corruption on the remote side.
> > fatal: protocol error: bad pack header
> >
> > I wil continue to use the previously fetched tree.
--
Cheers,
Stephen Rothwell
On Tue, Dec 12, 2017 at 1:24 PM, Cyrille Pitchen
<[email protected]> wrote:
>
> I don't see anything wrong when fetching the tree or when browsing the
> web server:
> http://git.infradead.org/l2-mtd.git/shortlog/refs/heads/spi-nor/next
There's definitely something funky there.
http://git.infradead.org/l2-mtd.git/commitdiff/f3203415105860e811f4ee9508efd009c9e6525e
http://git.infradead.org/l2-mtd.git/commitdiff/7a6843af9089b01eab266bbee0ed19f98f09a94a
both show "no differences found". I suspect that's because of git diff
failing due to missing objects.
Most likely reason? Something like an unsafe garbage collect / prune
operation at infradead.org at the same time as you pushed. I don't
know what infradead.org does, but you can definitely do bad things
maintaining a git tree.
I would suggest doing a force-push back to some known-good state, and
then pushing it out again.
Ie, something like
git push --force infradead.org:l2-mtd.git v4.15-rc1:master
to reset the master branch to v4.15-rc1, and then pushing out again.
And then asking whoever maintains the infradead git trees if something
odd happened..
Linus
Hi Cyrille,
On Wed, 13 Dec 2017 09:35:30 +1100 Stephen Rothwell <[email protected]> wrote:
>
> [cc'ing Linus on the off chance he knows something relevant]
>
> On Tue, 12 Dec 2017 22:24:17 +0100 Cyrille Pitchen <[email protected]> wrote:
> >
> > I've just added 2 commits about 8 hours ago, nothing special.
> > Could the issue be explained by an unlikely bad timing?
I assume you did something as the fetch works fine now (including the 2
new commits).
--
Cheers,
Stephen Rothwell
Hi all,
+ MTD maintainers
Le 12/12/2017 à 23:53, Linus Torvalds a écrit :
> On Tue, Dec 12, 2017 at 1:24 PM, Cyrille Pitchen
> <[email protected]> wrote:
>>
>> I don't see anything wrong when fetching the tree or when browsing the
>> web server:
>> http://git.infradead.org/l2-mtd.git/shortlog/refs/heads/spi-nor/next
>
> There's definitely something funky there.
>
> http://git.infradead.org/l2-mtd.git/commitdiff/f3203415105860e811f4ee9508efd009c9e6525e
> http://git.infradead.org/l2-mtd.git/commitdiff/7a6843af9089b01eab266bbee0ed19f98f09a94a
>
> both show "no differences found". I suspect that's because of git diff
> failing due to missing objects.
>
> Most likely reason? Something like an unsafe garbage collect / prune
> operation at infradead.org at the same time as you pushed. I don't
> know what infradead.org does, but you can definitely do bad things
> maintaining a git tree.
>
> I would suggest doing a force-push back to some known-good state, and
> then pushing it out again.
>
> Ie, something like
>
> git push --force infradead.org:l2-mtd.git v4.15-rc1:master
>
> to reset the master branch to v4.15-rc1, and then pushing out again.
> And then asking whoever maintains the infradead git trees if something
> odd happened..
>
I think I've fixed the issue following your advice.
However I don't know what happened to these old commits.
I've added MTD in Cc if they have some idea.
Anyway, sorry for this mess and thanks for your help!
Best regards,
Cyrille
> Linus
>