On Wed, May 23, 2012 at 09:02:30AM +0900, Minchan Kim wrote: > On 05/21/2012 11:19 PM, Seth Jennings wrote: > > > On 05/20/2012 09:23 PM, Minchan Kim wrote: > > > >> We should use unsigned long as handle instead of void * to avoid any > >> confusion. Without this, users may just treat zs_malloc return value as > >> a pointer and try to deference it. > > > > > > I wouldn't have agreed with you about the need for this change as people > > should understand a void * to be the address of some data with unknown > > structure. > > > > However, I recently discussed with Dan regarding his RAMster project > > where he assumed that the void * would be an address, and as such, > > 4-byte aligned. So he has masked two bits into the two LSBs of the > > handle for RAMster, which doesn't work with zsmalloc since the handle is > > not an address. > > > > So really we do need to convey as explicitly as possible to the user > > that the handle is an _opaque_ value about which no assumption can be made. > > > > Also, I wanted to test this but is doesn't apply cleanly on > > zsmalloc-main.c on v3.4 or what I have as your latest patch series. > > What is the base for this patch? > > > It's based on next-20120518. > I have always used linux-next tree for staging. > Greg, What's the convenient tree for you? linux-next is fine. But note, I'm ignoring all patches for the next 2 weeks, especially staging patches, as this is the merge window time, and I can't apply anything to my trees, sorry. After 3.5-rc1 is out, then I will look at new stuff like this again. thanks, greg k-h -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>