On Sat, Jan 10, 2009 at 1:31 PM, Sandeep K Sinha <sandeepksinha@xxxxxxxxx> wrote: > Well the way to avoid memcopy to copy the physical data from the > source data block to destination data block can be following, > > I have tested this and it works perfectly fine. > > 1. get the buffer head of the source inode's data blocks in 'x'. > 2. allocate a new destination block. > 3. get the buffer head of the newly allocated destination block into 'y'. > 4. y->b_data = x->b_data ; > 5. Mark the new buffer dirty and sync if you wish, still it would be ok. > > You are done !!! > > Remember that to allocate new data blocks, you would require to > allocate a dummy inode for the same. > Which you may release later on. > > Hope this helps. > > @greg/rohit > Reply inline. > > On Sat, Jan 10, 2009 at 12:22 PM, Rohit Sharma <imreckless@xxxxxxxxx> wrote: >> The idea given by Greg looks perfectly fine, >> i am going ahead to implement it, >> I expect to face many challenges while >> implementing Greg's algorithm because of >> the way ext2 is designed. >> I will update you guys soon. >> > > You will need to be very particular in the write code path, as the > actual for keeping system live lies there. > > Also, see how you can read the complete list of blocks allocated to a > particular inode. If it has indirect blocks, it would require extra > work as I didn't not fine any similar implementation/support in th > current ext2 code. ext2_get_blocks() does all that. See particularly the function ext2_block_to_path(). It fills up the array of the pointers you need to traverse and then does a add_chain to load that particular buffer. See point 2 in http://nngfs.pbwiki.com/Some+random+notes thanks - Manish >> On Fri, Jan 9, 2009 at 6:22 AM, Greg Freemyer <greg.freemyer@xxxxxxxxx> wrote: >>> HSM guys, >>> >>> Looking at your website, I don't see a way to join a mailing list. >>> > > We are also starting to host on sourceforge, which will be done soon. > currently we are hosting it on google code which doesn't have support > for the same as yet. >>> http://code.google.com/p/fscops/ >>> >>> Is there one? I added in fscops@xxxxxxxxx under the assumption is was a list. >>> > > No not as yet. But we will surely have one soon. > We communicate through issues on the google code web interface. we > will cc you the issues updates henceforth. > >>> Anyway, if you have a list I'd like to join it. >>> > > Surely, as soon as we have it, we will update you. > Thanks for you help and interest. >>> Greg >>> -- >>> Greg Freemyer >>> Litigation Triage Solutions Specialist >>> http://www.linkedin.com/in/gregfreemyer >>> First 99 Days Litigation White Paper - >>> http://www.norcrossgroup.com/forms/whitepapers/99%20Days%20whitepaper.pdf >>> >>> The Norcross Group >>> The Intersection of Evidence & Technology >>> http://www.norcrossgroup.com >>> >> > > > > -- > Regards, > Sandeep. > > > > > > > "To learn is to change. Education is a process that changes the learner." > -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
