Re: [PATCH] rtslib/tcm: Add NULLIO support for rd_dr/rd_mcp

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Andy and Sagi,

On 04/29/2014 05:04 PM, Andy Grover wrote:
On 04/29/2014 05:01 AM, Sagi Grimberg wrote:
When the user creates a RAM device with size=0 we
interperet it as implicit NULLIO option (Instead
of modifying size to 1).

Sagi, thanks for reporting this.

It is true that some of 3.10+-kernel-specific have been a bit stalled in the main rtslib branch. The focus thus far has been to support current stable distributions (typically 3.2 kernels), preserve backward compatibility for users of older distribution bases, and prepare the move to the config framework for 3.10+ kernels.

This is only a transient situation, as the lio-config framework will soon take over the main rtslib configuration save and restore logic, and active tracking of recent kernel features will ensue.

However, I see no reason to hold back support for rd_mcp nullio as this is a small change that can be made backward-compatible, so I just pushed the necessary changes to the datera github repository in lio-config, rtslib and targetcli master branches. See below.


We still set the size to a fixed NULLIO_SIZE of 1T
in order to emulate a large storage device to the
fabric.

Signed-off-by: Sagi Grimberg <sagig@xxxxxxxxxxxx>
Reported-by: Slava Shwartsman <valyushash@xxxxxxxxx>

rtslib-fb already has support for nullio via a "nullio" boolean param to
RDMCPStorageObject, which I think is a little cleaner and allows
flexible nullio volume size. I'd encourage Datera rtslib to also adopt a
similar approach.

Yes, I agree. Within targetcli, a separate bool is the clean way to do it.

I ported your nullio changes to main targetcli and rtslib, in master branches. I also made sure this will not break on 3.2 kernels, and added a warning in case a user asks for nullio=true on a kernel where this is not supported.

Also, lio-utils have been updated too, so nullio support works for dump/restore both for older and 3.10+ kernels with datera's stack.


(Also, rd_dr is gone so rtslib can remove support for it.)

There is some - maybe even a lot of - legacy code in rtslib, not only rd_dr but support for even older stuff. Again, this is because the main rtslib has been giving long-term support to some of our users, notably appliances and embedded storage systems that require it.

Best Regards,
--
Jerome
--
To unsubscribe from this list: send the line "unsubscribe target-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux SCSI]     [Kernel Newbies]     [Linux SCSI Target Infrastructure]     [Share Photos]     [IDE]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux ATA RAID]     [Linux IIO]     [Device Mapper]

  Powered by Linux