Le 7 sept. 10 à 01:18, Ben Hutchings a écrit :
On Mon, 2010-09-06 at 13:40 +0200, Thibaut VARENE wrote:reopen 595502 thanks Version: 2.6.32-21 Please, explain to me how changing a Panic into a Warning without effectively fixing the root cause of the bug "fixes" anything?And, oh, if by "graceful failure" you mean: "it will blow up your raiddevices and kill your filesystems", then yes, the failure has been very graceful for me, thankyouverymuch. I could have lived with a little less gracefulness, though ;-/[...] There's no need to take your frustration out on me. I know nothing
I don't think I've taken out my frustration on you, I would have been a lot less ironic and a lot more postal if I had been to take the frustration resulting from the pain this caused me on anyone... I must say tho I'm quite frustrated that incremental kernel upgrades to the now frozen next stable release introduce *regressions*...
about IA64 and have to assume that upstream developers know what they are doing.
That remains an open question ;P
Please test Linux 2.6.35 as packaged in experimental.
I'm afraid this is absolutely not happening, this machine is a webserver and I've just spent enough time bringing it back to life... Besides squeeze is not shipping with 2.6.35, right? I'm hoping you plan on releasing a working kernel for the upcoming stable release?
One thing I don't get is why the need to break a perfectly working kernel (2.6.32-9)? Can't the specific change that introduced this bug be simply reverted?