Messages in this thread | | | Date | Fri, 04 May 2012 15:08:01 +0200 | From | Paolo Bonzini <> | Subject | Re: [PATCH] scsi: Silence unnecessary warnings about ioctl to partition |
| |
Il 04/05/2012 15:06, Mark Lord ha scritto: >> > I said which ioctls, not which options. I.e. cut-and-paste from dmesg. > Here are some:
Thanks!
> messages.1:Apr 28 14:08:09 zbox5 kernel: [ 268.800536] hdparm: sending ioctl 330 to a partition! > messages.1:Apr 28 14:08:09 zbox5 kernel: [ 268.800547] hdparm: sending ioctl 330 to a partition! > messages.1:Apr 28 14:08:57 zbox5 kernel: [ 316.413944] hdparm: sending ioctl 330 to a partition! > messages.1:Apr 28 14:08:57 zbox5 kernel: [ 316.413951] hdparm: sending ioctl 330 to a partition! > messages.1:Apr 28 14:20:10 zbox5 kernel: [ 989.525085] hdparm: sending ioctl 330 to a partition! > messages.1:Apr 28 14:20:10 zbox5 kernel: [ 989.525093] hdparm: sending ioctl 330 to a partition! > > The in <linux/hdreg.h> say this about 330: > > ... > /* 0x330 is reserved - used to be HDIO_GETGEO_BIG */ > ... > > So it's HDIO_GETGEO_BIG, which doesn't exist in newer kernels. > I wonder when that got removed? Minor userspace breakage there. > > hdparm issues it first as a backward-compatibility thing, > before falling back to the even-more obsolete HDIO_GETGEO, > which curiously enough is still in modern kernels.
Ok, so hdparm is not broken. My patch (which left the warning only for SG_IO, and failed all other ioctls) would have worked, too.
Paolo
| |