Hi Dmitry, Yes I was in same impression, but for debugging, I disabled bcache and tried reading and writing block. Reads are successful but while writing a block bitmap or inode bitmap, I traced each call reaches to write syscall with correct blockbitmap and inodebitmap data. It shows, write is successful but when we re-read it, we read old data which is not what we have written. I suspected this issue with ata but now i'll try to run mentioned tests and see if it works! Anyway would you like to suggest me something, how we can better debug it or some tool to trace ata calls? im using qemu.
Thanks, Shivpratap On Mon, Aug 9, 2010 at 1:10 PM, Dmitry Gromada <[email protected]> wrote: > Hi Shiv, > > This is a problem of ext2fs. It has still not complete write support. > On the other hands, as well reading from as writing to disk has been > comprehensively tested, and has not any apparent problems. You can ensure > this > by launching the test named 't_libblkdev' residing at > 'jarios_applications/tests' against any ata block device. > > > Hi All, > > > > I am looking into ext2 fs, while writing files, dirs vfs calls > ext2_create > > method of etx2 fs but finally it calls write (users pace which uses rpc) > > for block devices which seems not working! > > can you pls explain, is there NO write support at block device ata? or > > there us bug in FS layer? > > > > Thanks, > > Shivpratap > > > Best regards, > Dmitry Gromada > _______________________________________________ > Jarios-dev mailing list > [email protected] > http://lists.jarios.org/cgi-bin/mailman/listinfo/jarios-dev >
_______________________________________________ Jarios-dev mailing list [email protected] http://lists.jarios.org/cgi-bin/mailman/listinfo/jarios-dev
