[PATCHv4 0/7] zswap: compressed swap caching

Seth Jennings sjenning at linux.vnet.ibm.com
Wed Jan 30 16:01:30 UTC 2013


On 01/29/2013 10:32 PM, Minchan Kim wrote:
> On Tue, Jan 29, 2013 at 04:49:04PM -0600, Seth Jennings wrote:
>> On 01/29/2013 04:14 PM, Joe Perches wrote:
>>> On Tue, 2013-01-29 at 15:40 -0600, Seth Jennings wrote:
>>>> The code required for the flushing is in a separate patch now
>>>> as requested.
>>>
>>> What tree does this apply to?
>>> Both -next and linus fail to compile.
>>
>> Link to build instruction in the cover letter:
>>
>>>> NOTE: To build, read this:
>>>> http://lkml.org/lkml/2013/1/28/586
>>
>> The complexity is due to a conflict with a zsmalloc patch in Greg's
>> staging tree that has yet to make its way upstream.
>>
>> Sorry for the inconvenience.
> 
> Seth, Please don't ignore previous review if you didn't convince reviewer.
> I don't want to consume time with arguing trivial things.
> 
> Copy and Paste from previous discussion from zsmalloc pathset
> 
>>>> On Fri, Jan 25, 2013 at 11:46:14AM -0600, Seth Jennings wrote:
>>>>> These patches are the first 4 patches of the zswap patchset I
>>>>> sent out previously.  Some recent commits to zsmalloc and
>>>>> zcache in staging-next forced a rebase. While I was at it, Nitin
>>>>> (zsmalloc maintainer) requested I break these 4 patches out from
>>>>> the zswap patchset, since they stand on their own.
>>>>
>>>> [2/4] and [4/4] is okay to merge current zsmalloc in staging but
>>>> [1/4] and [3/4] is dependent on zswap so it should be part of
>>>> zswap patchset.
>>>
>>> Just to clarify, patches 1 and 3 are _not_ dependent on zswap.  They
>>> just introduce changes that are only needed by zswap.
>>
>> I don't think so. If zswap might be not merged, we don't need [1, 3]
>> at the moment. You could argue that [1, 3] make zsmalloc more flexible
>> and I agree. BUT I want it when we have needs. It would be not too late.
>> So [1,3] should be part of zswap patchset.

I apologize.  I am really trying to keep all the feedback straight,
and I didn't know what Greg was going to do with those zsmalloc
patches.  However, as of last night, he didn't accept the two you
mentioned as being tied to zswap-only functionality.

I'll bring them back into the patchset for v5 once I/we address
Andrew's feedback, which might take some time.

Thanks,
Seth




More information about the devel mailing list