[01:03] <ppd> BenC: It gives me the same eject error with the new kernel :(
[01:06] <BenC> what's the eject error?
[01:07] <ppd> Can't eject. Last error: Invalid argument.
[01:07] <BenC> invalid argument, that's not the bug error
[01:08] <BenC> it does EIO, Input/Output Error
[01:08] <BenC> does eject -s work for you?
[01:08] <ppd> wait
[01:09] <ppd> yes
[01:09] <ppd> from the console as root it works
[01:09] <BenC> then you did something wrong with the patch to scsi-ioctl.c
[01:09] <BenC> recheck it, and make sure the ioctl fro CDROMEJECT  matches, and that the functions are properly copied
[01:10] <ppd> I am using the archck patchset. Could that influence the result?
[01:12] <ppd> http://pastebin.com/489438
[01:12] <ppd> this compiled fine
[01:25] <ppd> :(
[01:29] <ppd> BenC: If you could have a look at it I'd be helped very much
[01:39] <BenC> ppd: looks ok to me, not sure why it isn't working
[01:40] <ppd> probably a newer eject version is needed?!
[01:42] <BenC> no, shouldn't be
[01:42] <BenC> are you sure you are using the built kernel?
[01:42] <BenC> eject -s should be the same as eject -r with that patch, IOW, if eject -s works, then normal eject should aswell
[01:43] <ppd> hm. I'll try eject from the console
[01:44] <ppd> max@max:/usr/src/linux-2.6.14$ eject /dev/sda1
[01:44] <ppd> eject: Kann nicht auswerfen! Letzter Fehler: Das Argument ist ungltig
[01:45] <ppd> umount was ok
[01:45] <ppd> probably I really started the wrong kernel... can't be....
[02:17] <ppd> BenC: sorry for annoying you again. But this time I'm absolutely sure I run the patched kernel and it doesn't work
[02:17] <ppd> BenC: I have done a strace eject /dev/sda1 but this didn't help me 
[02:24] <ogra> ogra@aleph:~$ mount |grep mapper
[02:24] <ogra> /dev/mapper/vg0-lvol0 on / type ext3 (rw,errors=remount-ro)
[02:24] <ogra> /dev/mapper/vg1-lvol1 on /var type ext3 (rw)
[02:24] <ogra> BenC, mjg59, thanks a lot !!
[02:24] <ogra> i owe you a drink or something ...
[02:26] <BenC> what for? :)
[02:26] <ogra> pointing me to the right places :) 
[02:26] <ogra> the via patch seems to run fine ...
[06:43] <infinity> Oh look, a new ABI.
[06:44] <infinity> BenC: I need to give LRM some TLC... Any reservations about holding off on the LRM/-meta uploads until tomorrow, when I can make it all happy?
[06:44] <infinity> (I can do -meta when LRM is done)
[06:44] <fabbione> morning guys
[06:45] <fabbione> oh oh oh
[06:50] <infinity> ?
[06:50] <fabbione> .15 final in the archive :)
[06:50] <infinity> Indeed.
[09:11] <Panda2> 
[09:12] <Panda2>   ?
[09:12] <infinity> Panda2: This is an English channel.
[09:13] <Panda2> sorry, i'm dont know
[09:14] <Panda2> Adam Conrad this cool nick
[09:14] <Panda2> talk to me Adam Conrad
[09:16] <Panda2> people?????????
[10:47] <CataEnry> hi all
[12:27] <pappan> BenC: hi
[01:28] <pappan> any 1 here
[01:28] <pappan> from kernel devel team
[01:45] <CataEnry> hi all
[02:05] <ppd> hi
[02:14] <ppd> is it possible the several fixes need to be backported in order to fix that eject bug?
[02:36] <BenC> no
[02:37] <ppd> BenC: then either I patched the file badly or this fix ust won't work with 2.6.14!?
[02:38] <BenC> Your error is not the same as what is in the bug report, but I have no idea why you are getting the error
[02:38] <BenC> No sure how eject -s could work for you, but normal eject with the patch doesn't
[02:39] <ppd> eject -s works with root rights
[02:39] <BenC> do this
[02:39] <BenC> strace -o eject.out eject
[02:39] <BenC> and pastebin th eject.out
[02:39] <ppd> two seconds :)
[02:42] <ppd> http://pastebin.com/490104
[02:43] <ppd> :)
[02:45] <BenC> ah, ok, the CDROMEJECT ioctl is getting EIO
[02:46] <BenC> I didn't realize that eject tries every method if they keep failing
[02:46] <BenC> ls -l /dev/sda1
[02:46] <BenC> let me see that
[02:47] <ppd> brw-r-----  1 root plugdev 8, 1 2006-01-04 14:46 /dev/sda1
[02:47] <BenC> ok, so updated eject wont help (sg_io needs write perms)
[02:47] <BenC> from now on when testing the fix, do "eject -r", since that only does the CDROMEJECT ioctl that we are testing
[02:48] <BenC> what type of device is this you are ejecting?
[02:48] <ppd> an usb plugged harddisk
[02:49] <ppd> eject -r gives inpout/output error
[02:49] <BenC> and why exactly do you want to eject a harddisk?
[02:49] <ppd> i don't want. gnome does...
[02:50] <BenC> and gnome is complaining because it fails?
[02:51] <ppd> gnome pops up an error message telling me "Eject. ... Invalid argument"
[02:51] <BenC> can you do "sudo strace -o eject.out eject -s"?
[02:51] <ppd> sure
[02:51] <BenC> I think the issue here is that this device cannot be ejected
[02:53] <ppd> http://pastebin.com/490110
[02:55] <BenC> does that make the device inaccessible afterwards?
[02:56] <ppd> if I plug it in again?
[02:56] <BenC> no, right after eject
[02:57] <BenC> can you remount it after ejecting it with eject -s?
[02:58] <ppd> max@max:~$ sudo mount /dev/sda1 test/
[02:58] <ppd> max@max:~$
[02:58] <ppd> no problem
[03:00] <BenC> then it's not really ejecting
[03:01] <ppd> I think so. a harddisk can't be eject, or?
[03:06] <BenC> well some usb devices can be, like ipods
[03:06] <BenC> it basically puts the device offline
[03:06] <ppd> my usb stick doesn't give an error when umounting/ejecting with gnome
[03:07] <BenC> but eject -r still should not be returning EIO, since eject -s is not either
[03:07] <BenC> can you try eject -r on the device?
[03:07] <ppd> as user?
[03:07] <BenC> yeah
[03:07] <ppd> with strace?
[03:09] <ppd> http://pastebin.com/490132
[03:09] <ppd> here's the strace output
[03:18] <ppd> i/o error
[03:19] <BenC> ah, I see why eject -s isn't returning an error
[03:19] <BenC> it actually is getting an error, just that eject is ignoring it
[03:19] <ppd> that would be an option too :)
[03:21] <BenC> does "sudo eject -s" take your USB device offline (can't remount it)?
[03:24] <ppd> max@max:~$ sudo mount /dev/sdb1 test
[03:24] <ppd> max@max:~$ cd test && ls
[03:24] <ppd> lost+found  nvidia_patched  user
[03:24] <ppd> max@max:~/test$
[03:24] <ppd> stays online
[03:26] <BenC> ok, then this is a different bug altogether, in that gnome should do something better
[03:27] <BenC> I wonder if gnome calls eject, or calls it's own internal ioctl's
[03:27] <ppd> hm, probably I can find something in libnautilus source
[03:28] <ppd> or can iI strace nautilus eject ?
[03:28] <BenC> you can try, but this is getting out of my expertise
[03:29] <ppd> ok. thank you very much. So there's nothing I can do?
[03:37] <zul> heylo
[03:39] <ppd> BenC: a new version of eject wouldn't help? or hacking eject so that it ignores that error?
[03:47] <ppd> scsi eject works as root
[03:48] <BenC> it doesn't work, it just ignores the error
[03:48] <ppd> so who causes the error?
[03:48] <BenC> eject -s is the same as eject -r functionally in the kernel, after patching scsi_ioctl.c
[03:48] <BenC> it's just that eject ignores the error from sg_io
[03:48] <BenC> the device causes the error
[03:49] <BenC> it's scsi eject that is ignoring things
[03:49] <BenC> in the eject program that is
[03:49] <BenC> if it checked hdr->status, it would see the error
[03:50] <ppd> ok. so the error is neither in the kernel nor in eject?
[03:50] <ppd> It's just that you shouldn't call eject for a usb disk?
[07:06] <CataEnry> bye all
[10:40] <fabbione> BenC: ping?
[11:20] <BenC> fabbione: pong