c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2026-05-22 10:37:35 +03:00
Code Issues Packages Projects Releases Wiki Activity

Fix statfs(2) for 32-bit user space

Browse Source 
When handling a 32-bit statfs() system call the returned fields,
although 64-bit in the kernel, must be limited to 32-bits or an
EOVERFLOW error will be returned.

This is less of an issue for block counts since the default
reported block size in 128KiB. But since it is possible to
set a smaller block size, these values will be scaled as
needed to fit in a 32-bit unsigned long.

Unlike most other filesystems the total possible file counts
are more likely to overflow because they are calculated based
on the available free space in the pool. In order to prevent
this the reported value must be capped at 2^32-1. This is
only for statfs(2) reporting, there are no changes to the
internal ZFS limits.

Reviewed-by: Andreas Dilger <andreas.dilger@whamcloud.com>
Reviewed-by: Richard Yao <ryao@gentoo.org>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #7927 
Closes #7122 
Closes #7937
This commit is contained in:
Brian Behlendorf
2018-09-24 17:11:25 -07:00
committed by GitHub
parent 36e369ecb8
commit e897a23eb1
5 changed files with 64 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
module/zfs/zfs_vfsops.c
+3 -5
View File
@@ -1422,8 +1422,6 @@ zfs_statvfs(struct dentry *dentry, struct kstatfs *statp)
{
zfsvfs_t *zfsvfs = dentry->d_sb->s_fs_info;
uint64_t refdbytes, availbytes, usedobjs, availobjs;
uint64_t fsid;
uint32_t bshift;
int err = 0;
ZFS_ENTER(zfsvfs);
@@ -1431,7 +1429,7 @@ zfs_statvfs(struct dentry *dentry, struct kstatfs *statp)
dmu_objset_space(zfsvfs->z_os,
&refdbytes, &availbytes, &usedobjs, &availobjs);
fsid = dmu_objset_fsid_guid(zfsvfs->z_os);
uint64_t fsid = dmu_objset_fsid_guid(zfsvfs->z_os);
/*
* The underlying storage pool actually uses multiple block
* size. Under Solaris frsize (fragment size) is reported as
@@ -1443,7 +1441,7 @@ zfs_statvfs(struct dentry *dentry, struct kstatfs *statp)
*/
statp->f_frsize = zfsvfs->z_max_blksz;
statp->f_bsize = zfsvfs->z_max_blksz;
bshift = fls(statp->f_bsize) - 1;
uint32_t bshift = fls(statp->f_bsize) - 1;
/*
* The following report "total" blocks of various kinds in
@@ -1460,7 +1458,7 @@ zfs_statvfs(struct dentry *dentry, struct kstatfs *statp)
* static metadata. ZFS doesn't preallocate files, so the best
* we can do is report the max that could possibly fit in f_files,
* and that minus the number actually used in f_ffree.
* For f_ffree, report the smaller of the number of object available
* For f_ffree, report the smaller of the number of objects available
* and the number of blocks (each object will take at least a block).
*/
statp->f_ffree = MIN(availobjs, availbytes >> DNODE_SHIFT);
module/zfs/zpl_super.c
+22
View File
@@ -181,6 +181,28 @@ zpl_statfs(struct dentry *dentry, struct kstatfs *statp)
spl_fstrans_unmark(cookie);
ASSERT3S(error, <=, 0);
/*
* If required by a 32-bit system call, dynamically scale the
* block size up to 16MiB and decrease the block counts. This
* allows for a maximum size of 64EiB to be reported. The file
* counts must be artificially capped at 2^32-1.
*/
if (unlikely(zpl_is_32bit_api())) {
while (statp->f_blocks > UINT32_MAX &&
statp->f_bsize < SPA_MAXBLOCKSIZE) {
statp->f_frsize <<= 1;
statp->f_bsize <<= 1;
statp->f_blocks >>= 1;
statp->f_bfree >>= 1;
statp->f_bavail >>= 1;
}
uint64_t usedobjs = statp->f_files - statp->f_ffree;
statp->f_ffree = MIN(statp->f_ffree, UINT32_MAX - usedobjs);
statp->f_files = statp->f_ffree + usedobjs;
}
return (error);
}