From 5070e5311cb39be35fb059e9e0489070814c44a5 Mon Sep 17 00:00:00 2001 From: Chunwei Chen Date: Fri, 27 Jan 2017 10:43:39 -0800 Subject: [PATCH] Retire .write/.read file operations The .write/.read file operations callbacks can be retired since support for .read_iter/.write_iter and .aio_read/.aio_write has been added. The vfs_write()/vfs_read() entry functions will select the correct interface for the kernel. This is desirable because all VFS write/read operations now rely on common code. This change also add the generic write checks to make sure that ulimits are enforced correctly on write. Signed-off-by: Brian Behlendorf Signed-off-by: Chunwei Chen Closes #5587 Closes #5673 --- config/kernel-vfs-rw-iterate.m4 | 21 +++++++++ config/kernel.m4 | 1 + module/zfs/zpl_file.c | 83 ++++++++++++++++++++------------- 3 files changed, 72 insertions(+), 33 deletions(-) diff --git a/config/kernel-vfs-rw-iterate.m4 b/config/kernel-vfs-rw-iterate.m4 index f8dc422296..af44beb7bc 100644 --- a/config/kernel-vfs-rw-iterate.m4 +++ b/config/kernel-vfs-rw-iterate.m4 @@ -25,3 +25,24 @@ AC_DEFUN([ZFS_AC_KERNEL_VFS_RW_ITERATE], AC_MSG_RESULT(no) ]) ]) + +dnl # +dnl # Linux 4.1.x API +dnl # +AC_DEFUN([ZFS_AC_KERNEL_GENERIC_WRITE_CHECKS], + [AC_MSG_CHECKING([whether generic_write_checks() takes kiocb]) + ZFS_LINUX_TRY_COMPILE([ + #include + + ],[ + struct kiocb *iocb = NULL; + struct iov_iter *iov = NULL; + generic_write_checks(iocb, iov); + ],[ + AC_MSG_RESULT(yes) + AC_DEFINE(HAVE_GENERIC_WRITE_CHECKS_KIOCB, 1, + [generic_write_checks() takes kiocb]) + ],[ + AC_MSG_RESULT(no) + ]) +]) diff --git a/config/kernel.m4 b/config/kernel.m4 index fcbb686c27..5b8abf7516 100644 --- a/config/kernel.m4 +++ b/config/kernel.m4 @@ -102,6 +102,7 @@ AC_DEFUN([ZFS_AC_CONFIG_KERNEL], [ ZFS_AC_KERNEL_LSEEK_EXECUTE ZFS_AC_KERNEL_VFS_ITERATE ZFS_AC_KERNEL_VFS_RW_ITERATE + ZFS_AC_KERNEL_GENERIC_WRITE_CHECKS ZFS_AC_KERNEL_KMAP_ATOMIC_ARGS ZFS_AC_KERNEL_FOLLOW_DOWN_ONE ZFS_AC_KERNEL_MAKE_REQUEST_FN diff --git a/module/zfs/zpl_file.c b/module/zfs/zpl_file.c index a853754b1c..bdb37f29d1 100644 --- a/module/zfs/zpl_file.c +++ b/module/zfs/zpl_file.c @@ -259,20 +259,6 @@ zpl_read_common(struct inode *ip, const char *buf, size_t len, loff_t *ppos, flags, cr, 0)); } -static ssize_t -zpl_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos) -{ - cred_t *cr = CRED(); - ssize_t read; - - crhold(cr); - read = zpl_read_common(filp->f_mapping->host, buf, len, ppos, - UIO_USERSPACE, filp->f_flags, cr); - crfree(cr); - - return (read); -} - static ssize_t zpl_iter_read_common(struct kiocb *kiocb, const struct iovec *iovp, unsigned long nr_segs, size_t count, uio_seg_t seg, size_t skip) @@ -310,7 +296,14 @@ static ssize_t zpl_aio_read(struct kiocb *kiocb, const struct iovec *iovp, unsigned long nr_segs, loff_t pos) { - return (zpl_iter_read_common(kiocb, iovp, nr_segs, kiocb->ki_nbytes, + ssize_t ret; + size_t count; + + ret = generic_segment_checks(iovp, &nr_segs, &count, VERIFY_WRITE); + if (ret) + return (ret); + + return (zpl_iter_read_common(kiocb, iovp, nr_segs, count, UIO_USERSPACE, 0)); } #endif /* HAVE_VFS_RW_ITERATE */ @@ -348,6 +341,7 @@ zpl_write_common_iovec(struct inode *ip, const struct iovec *iovp, size_t count, return (wrote); } + inline ssize_t zpl_write_common(struct inode *ip, const char *buf, size_t len, loff_t *ppos, uio_seg_t segment, int flags, cred_t *cr) @@ -361,20 +355,6 @@ zpl_write_common(struct inode *ip, const char *buf, size_t len, loff_t *ppos, flags, cr, 0)); } -static ssize_t -zpl_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos) -{ - cred_t *cr = CRED(); - ssize_t wrote; - - crhold(cr); - wrote = zpl_write_common(filp->f_mapping->host, buf, len, ppos, - UIO_USERSPACE, filp->f_flags, cr); - crfree(cr); - - return (wrote); -} - static ssize_t zpl_iter_write_common(struct kiocb *kiocb, const struct iovec *iovp, unsigned long nr_segs, size_t count, uio_seg_t seg, size_t skip) @@ -395,16 +375,40 @@ zpl_iter_write_common(struct kiocb *kiocb, const struct iovec *iovp, static ssize_t zpl_iter_write(struct kiocb *kiocb, struct iov_iter *from) { + size_t count; ssize_t ret; uio_seg_t seg = UIO_USERSPACE; + +#ifndef HAVE_GENERIC_WRITE_CHECKS_KIOCB + struct file *file = kiocb->ki_filp; + struct address_space *mapping = file->f_mapping; + struct inode *ip = mapping->host; + int isblk = S_ISBLK(ip->i_mode); + + count = iov_iter_count(from); + ret = generic_write_checks(file, &kiocb->ki_pos, &count, isblk); +#else + /* + * XXX - ideally this check should be in the same lock region with + * write operations, so that there's no TOCTTOU race when doing + * append and someone else grow the file. + */ + ret = generic_write_checks(kiocb, from); + count = ret; +#endif + if (ret <= 0) + return (ret); + if (from->type & ITER_KVEC) seg = UIO_SYSSPACE; if (from->type & ITER_BVEC) seg = UIO_BVEC; + ret = zpl_iter_write_common(kiocb, from->iov, from->nr_segs, - iov_iter_count(from), seg, from->iov_offset); + count, seg, from->iov_offset); if (ret > 0) iov_iter_advance(from, ret); + return (ret); } #else @@ -412,7 +416,22 @@ static ssize_t zpl_aio_write(struct kiocb *kiocb, const struct iovec *iovp, unsigned long nr_segs, loff_t pos) { - return (zpl_iter_write_common(kiocb, iovp, nr_segs, kiocb->ki_nbytes, + struct file *file = kiocb->ki_filp; + struct address_space *mapping = file->f_mapping; + struct inode *ip = mapping->host; + int isblk = S_ISBLK(ip->i_mode); + size_t count; + ssize_t ret; + + ret = generic_segment_checks(iovp, &nr_segs, &count, VERIFY_READ); + if (ret) + return (ret); + + ret = generic_write_checks(file, &pos, &count, isblk); + if (ret) + return (ret); + + return (zpl_iter_write_common(kiocb, iovp, nr_segs, count, UIO_USERSPACE, 0)); } #endif /* HAVE_VFS_RW_ITERATE */ @@ -834,8 +853,6 @@ const struct file_operations zpl_file_operations = { .open = zpl_open, .release = zpl_release, .llseek = zpl_llseek, - .read = zpl_read, - .write = zpl_write, #ifdef HAVE_VFS_RW_ITERATE .read_iter = zpl_iter_read, .write_iter = zpl_iter_write,