[Squashfs-devel] Question about the use of SQUASHFS_I in squashfs2_0.c

[Guang Z. <ze...@ho...>]

Hi all,I am totoally newbie to filesystem development in Linux, so please k=
indly provide more details/advices in replying my question, thanks!I am rea=
ding the 3.0 code, and find in function squashfs_iget_2, there are quit a f=
ew use of macro SQUASHFS_I on the inode structure named i. As far as I can =
understand, SQUASHFS_I is a wrapper macro to get the squashfs_inode_info st=
ructure which the inode structure is embedded in. What I cannot understand =
is that the variable i, which is of type struct inode, is created from scra=
tch by calling squashfs_new_inode, not embedded within some squashfs_inode_=
info structure (see implementation of squashfs_new_inode in inode.c). So it=
 seems that we have no knowledge of the memory space preceeding or succeedi=
ng of the inode structure, but why it seems to be legitimate to reversely l=
ookup the enclosing squashfs_inode_info structure (because the code just wo=
rked!). Can anyone shed some light on this? Thanks.Regards,Guang Zhu=