Date: Sat, 16 Jul 2011 14:10:31 +0300 From: Gleb Kurtsou <gleb.kurtsou@gmail.com> To: soc-status@freebsd.org Subject: ino64 status report 8 Message-ID: <20110716111031.GA85084@tops>
next in thread | raw e-mail | index | archive | help
ino64 project aims at extending ino_t to 64 bits and nlink_t to 32 bits integers providing compatibility shims for both kernel and libc. Double buffering was removed from ufs_readdir, buffer cache is now used directly. To prevent a readdir race when offset from previous readdir becomes stale ufs dirrect's are parsed starting at DIRBLKSIZ boundary (512 bytes). Skip entries with zero inode number. NFS readdir offset and buffer size alignment was fixed: getdirentries() call with not DIRBLKSIZ aligned offset and/or buffer size will result in file offset that doesn't start on dirent boundary, i.e. next getdirentries() call will return garbled result. readdir-lint updated to better handle NFS' readdir output.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110716111031.GA85084>