sending ioctl warnings for a partition

sending ioctl warnings for a partition

[Tullio Dovera]

My platform is based on linux v3.2.2 + Debian stable zfs-fuse-0.6.9-1
application

When I use a dedicated scsi partition for zfs-fuse I see the following
warnings at kernel level:

kernel: [   12.881949] zfs-fuse: sending ioctl 2285 to a partition!

The SG_IO is blocked inside scsi_verify_blk_ioctl() function. In order to permit
SG_IO for a partition I suggest to add it in the white list:

--- a/block/scsi_ioctl.c	2012-02-01 11:10:41.000000000 +0100
+++ b/block/scsi_ioctl.c	2012-02-01 11:31:09.000000000 +0100
@@ -710,6 +710,7 @@
 	case SG_GET_RESERVED_SIZE:
 	case SG_SET_RESERVED_SIZE:
 	case SG_EMULATED_HOST:
+	case SG_IO:
 		return 0;
 	case CDROM_GET_CAPABILITY:
 		/* Keep this until we remove the printk below.  udev sends it

Best regards,
   Tullio

Re: sending ioctl warnings for a partition

[Douglas Gilbert]

On 12-02-04 11:05 AM, Tullio Dovera wrote:
> My platform is based on linux v3.2.2 + Debian stable zfs-fuse-0.6.9-1
> application
>
> When I use a dedicated scsi partition for zfs-fuse I see the following
> warnings at kernel level:
>
> kernel: [   12.881949] zfs-fuse: sending ioctl 2285 to a partition!
>
> The SG_IO is blocked inside scsi_verify_blk_ioctl() function. In order to permit
> SG_IO for a partition I suggest to add it in the white list:
>
> --- a/block/scsi_ioctl.c	2012-02-01 11:10:41.000000000 +0100
> +++ b/block/scsi_ioctl.c	2012-02-01 11:31:09.000000000 +0100
> @@ -710,6 +710,7 @@
>   	case SG_GET_RESERVED_SIZE:
>   	case SG_SET_RESERVED_SIZE:
>   	case SG_EMULATED_HOST:
> +	case SG_IO:
>   		return 0;
>   	case CDROM_GET_CAPABILITY:
>   		/* Keep this until we remove the printk below.  udev sends it

If I'm not mistaken, the whole point of a series of patches
titled: "fail SCSI passthrough ioctls on partition devices"
was to flag and ultimately break any user space code that
attempted to use the SG_IO ioctl on a partition block
device (e.g. send a SCSI INQUIRY via SG_IO to /dev/sda2).

This is for security reasons. So when the kernel enforces
that, zfs-fuse will break. Please alert the zfs-fuse
maintainers.

Doug Gilbert

Re: sending ioctl warnings for a partition

[Paolo Bonzini]

On 02/04/2012 07:28 PM, Douglas Gilbert wrote:
>
> If I'm not mistaken, the whole point of a series of patches
> titled: "fail SCSI passthrough ioctls on partition devices"
> was to flag and ultimately break any user space code that
> attempted to use the SG_IO ioctl on a partition block
> device (e.g. send a SCSI INQUIRY via SG_IO to /dev/sda2).
>
> This is for security reasons. So when the kernel enforces
> that, zfs-fuse will break. Please alert the zfs-fuse
> maintainers.

That's correct.  zfs-fuse is trying to send a SYNCHRONIZE CACHE command, 
they should just use BLKFLSBUF (which will also give a warning due to a 
quirk in block/ioctl.c, but it's a false positive and the warning will 
go away soonish).

In principle, SYNCHRONIZE CACHE is "safe" even if sent to a partition 
and doesn't really leak anything, but I'm not sure we want to look into 
the CDB in scsi_ioctl.c.

Paolo

Re: sending ioctl warnings for a partition

[Paolo Bonzini]

On 02/05/2012 12:30 AM, Paolo Bonzini wrote:
> That's correct.  zfs-fuse is trying to send a SYNCHRONIZE CACHE command,
> they should just use BLKFLSBUF (which will also give a warning due to a
> quirk in block/ioctl.c, but it's a false positive and the warning will
> go away soonish).

Not even that.  All that is needed is fsync.

Paolo

Re: sending ioctl warnings for a partition

[Tullio Dovera]

Thank you very much. I have opened an issue in zfs-fuse project.

Tullio

On Sun, Feb 5, 2012 at 3:14 PM, Paolo Bonzini <pbonzini@redhat.com> wrote:
> On 02/05/2012 12:30 AM, Paolo Bonzini wrote:
>>
>> That's correct.  zfs-fuse is trying to send a SYNCHRONIZE CACHE command,
>> they should just use BLKFLSBUF (which will also give a warning due to a
>> quirk in block/ioctl.c, but it's a false positive and the warning will
>> go away soonish).
>
>
> Not even that.  All that is needed is fsync.
>
> Paolo
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: sending ioctl warnings for a partition

[Milan Knížek]

Hello Tullio,

can you pls point me where in zfs-fuse project is the issue reported?

I have not found anything in zfs-fuse @ googlegroups and
http://zfs-fuse.net/issues

Thank you and regards,
Milan

-- 

To: Paolo Bonzini <pbonzini@xxxxxxxxxx>
Subject: Re: sending ioctl warnings for a partition
From: Tullio Dovera <tdovera@xxxxxxxxx>
Date: Tue, 7 Feb 2012 17:34:58 +0100
Cc: linux-scsi <linux-scsi@xxxxxxxxxxxxxxx>, zfs-fuse@xxxxxxx
In-reply-to: <4F2E8EB3.3010307@redhat.com>
Thank you very much. I have opened an issue in zfs-fuse project.

Tullio

On Sun, Feb 5, 2012 at 3:14 PM, Paolo Bonzini <pbonzini@xxxxxxxxxx>
wrote:
> On 02/05/2012 12:30 AM, Paolo Bonzini wrote:
>>
>> That's correct.  zfs-fuse is trying to send a SYNCHRONIZE CACHE
>> command, they should just use BLKFLSBUF (which will also give a
>> warning due to a quirk in block/ioctl.c, but it's a false positive
>> and the warning will go away soonish).
>
>
> Not even that.  All that is needed is fsync.
>
> Paolo

-- 
http://www.milan-knizek.net/
About linux and photography (Czech only)
O linuxu a fotografování
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: sending ioctl warnings for a partition

[Tullio Dovera]

Dear Milan,
   I have reopened the issue because I can't find it in zfs-fuse project.

The issue is #143

Best regards,
   Tullio

On Mon, Feb 27, 2012 at 11:48 AM, Milan Knížek <knizek.confy@gmail.com> wrote:
>
> Hello Tullio,
>
> can you pls point me where in zfs-fuse project is the issue reported?
>
> I have not found anything in zfs-fuse @ googlegroups and
> http://zfs-fuse.net/issues
>
> Thank you and regards,
> Milan
>
> --
>
> To: Paolo Bonzini <pbonzini@xxxxxxxxxx>
> Subject: Re: sending ioctl warnings for a partition
> From: Tullio Dovera <tdovera@xxxxxxxxx>
> Date: Tue, 7 Feb 2012 17:34:58 +0100
> Cc: linux-scsi <linux-scsi@xxxxxxxxxxxxxxx>, zfs-fuse@xxxxxxx
> In-reply-to: <4F2E8EB3.3010307@redhat.com>
> Thank you very much. I have opened an issue in zfs-fuse project.
>
> Tullio
>
> On Sun, Feb 5, 2012 at 3:14 PM, Paolo Bonzini <pbonzini@xxxxxxxxxx>
> wrote:
> > On 02/05/2012 12:30 AM, Paolo Bonzini wrote:
> >>
> >> That's correct.  zfs-fuse is trying to send a SYNCHRONIZE CACHE
> >> command, they should just use BLKFLSBUF (which will also give a
> >> warning due to a quirk in block/ioctl.c, but it's a false positive
> >> and the warning will go away soonish).
> >
> >
> > Not even that.  All that is needed is fsync.
> >
> > Paolo
>
> --
> http://www.milan-knizek.net/
> About linux and photography (Czech only)
> O linuxu a fotografování
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html