- 23 Jul, 2016 1 commit
-
-
Eric W. Biederman authored
Seth Forshee reported a mount regression in nfs autmounts with "fs: Add user namespace member to struct super_block". It turns out that the assumption that current->cred is something reasonable during mount while necessary to improve support of unprivileged mounts is wrong in the automount path. To fix the existing filesystems override current->cred with the init_cred before calling d_automount and restore current->cred after d_automount completes. To support unprivileged mounts would require a more nuanced cred selection, so fail on unprivileged mounts for the time being. As none of the filesystems that currently set FS_USERNS_MOUNT implement d_automount this check is only good for preventing future problems. Fixes: 6e4eab57 ("fs: Add user namespace member to struct super_block") Tested-by:
Seth Forshee <seth.forshee@canonical.com> Signed-off-by:
"Eric W. Biederman" <ebiederm@xmission.com>
-
- 05 Jul, 2016 7 commits
-
-
Seth Forshee authored
For filesystems mounted from a user namespace on-disk ids should be translated relative to s_users_ns rather than init_user_ns. When an id in the filesystem doesn't exist in s_user_ns the associated id in the inode will be set to INVALID_[UG]ID, which turns these into de facto "nobody" ids. This actually maps pretty well into the way most code already works, and those places where it didn't were fixed in previous patches. Moving forward vfs code needs to be careful to handle instances where ids in inodes may be invalid. Signed-off-by:
-
Seth Forshee authored
The EVM HMAC should be calculated using the on disk user and group ids, so the k[ug]ids in the inode must be translated relative to the s_user_ns of the inode's super block. Signed-off-by:
-
Eric W. Biederman authored
Mostly supporting filesystems outside of init_user_ns is s/&init_usre_ns/dquot->dq_sb->s_user_ns/. An actual need for supporting quotas on filesystems outside of s_user_ns is quite a ways away and to be done responsibily needs an audit on what can happen with hostile quota files. Until that audit is complete don't attempt to support quota files on filesystems outside of s_user_ns. Cc: Jan Kara <jack@suse.cz> Acked-by:
-
Eric W. Biederman authored
In Q_XSETQLIMIT use sb->s_user_ns to detect when we are dealing with the filesystems notion of id 0. Cc: Jan Kara <jack@suse.cz> Acked-by:
-
Eric W. Biederman authored
Introduce the helper qid_has_mapping and use it to ensure that the quota system only considers qids that map to the filesystems s_user_ns. In practice for quota supporting filesystems today this is the exact same check as qid_valid. As only 0xffffffff aka (qid_t)-1 does not map into init_user_ns. Replace the qid_valid calls with qid_has_mapping as values come in from userspace. This is harmless today and it prepares the quota system to work on filesystems with quotas but mounted by unprivileged users. Call qid_has_mapping from dqget. This ensures the passed in qid has a prepresentation on the underlying filesystem. Previously this was unnecessary as filesystesm never had qids that could not map. With the introduction of filesystems outside of s_user_ns this will not remain true. All of this ensures the quota code never has to deal with qids that don't map to the underlying filesystem. Cc: Jan Kara <jack@suse.cz> Acked-by:
-
Eric W. Biederman authored
It is expected that filesystems can not represent uids and gids from outside of their user namespace. Keep things simple by not even trying to create filesystem nodes with non-sense uids and gids. Acked-by:
-
Eric W. Biederman authored
When a filesystem outside of init_user_ns is mounted it could have uids and gids stored in it that do not map to init_user_ns. The plan is to allow those filesystems to set i_uid to INVALID_UID and i_gid to INVALID_GID for unmapped uids and gids and then to handle that strange case in the vfs to ensure there is consistent robust handling of the weirdness. Upon a careful review of the vfs and filesystems about the only case where there is any possibility of confusion or trouble is when the inode is written back to disk. In that case filesystems typically read the inode->i_uid and inode->i_gid and write them to disk even when just an inode timestamp is being updated. Which leads to a rule that is very simple to implement and understand inodes whose i_uid or i_gid is not valid may not be written. In dealing with access times this means treat those inodes as if the inode flag S_NOATIME was set. Reads of the inodes appear safe and useful, but any write or modification is disallowed. The only inode write that is allowed is a chown that sets the uid and gid on the inode to valid values. After such a chown the inode is normal and may be treated as such. Denying all writes to inodes with uids or gids unknown to the vfs also prevents several oddball cases where corruption would have occurred because the vfs does not have complete information. One problem case that is prevented is attempting to use the gid of a directory for new inodes where the directories sgid bit is set but the directories gid is not mapped. Another problem case avoided is attempting to update the evm hash after setxattr, removexattr, and setattr. As the evm hash includeds the inode->i_uid or inode->i_gid not knowning the uid or gid prevents a correct evm hash from being computed. evm hash verification also fails when i_uid or i_gid is unknown but that is essentially harmless as it does not cause filesystem corruption. Acked-by:
-
- 30 Jun, 2016 4 commits
-
-
Seth Forshee authored
Using INVALID_[UG]ID for the LSM file creation context doesn't make sense, so return an error if the inode passed to set_create_file_as() has an invalid id. Signed-off-by:
Serge Hallyn <serge.hallyn@canonical.com> Signed-off-by:
-
Seth Forshee authored
Filesystem uids which don't map into a user namespace may result in inode->i_uid being INVALID_UID. A symlink and its parent could have different owners in the filesystem can both get mapped to INVALID_UID, which may result in following a symlink when this would not have otherwise been permitted when protected symlinks are enabled. Signed-off-by:
-
Eric W. Biederman authored
Update posix_acl_valid to verify that an acl is within a user namespace. Update the callers of posix_acl_valid to pass in an appropriate user namespace. For posix_acl_xattr_set and v9fs_xattr_set_acl pass in inode->i_sb->s_user_ns to posix_acl_valid. For md_unpack_acl pass in &init_user_ns as no inode or superblock is in sight. Acked-by:
-
Eric W. Biederman authored
Refuse to admit any user namespace has a mapping of the INVALID_UID and the INVALID_GID when !CONFIG_USER_NS. Acked-by:
-
- 28 Jun, 2016 1 commit
-
-
Seth Forshee authored
Add checks to notify_change to verify that uid and gid changes will map into the superblock's user namespace. If they do not fail with -EOVERFLOW. This is mandatory so that fileystems don't have to even think of dealing with ia_uid and ia_gid that --EWB Moved the test from inode_change_ok to notify_change Signed-off-by:
-
- 24 Jun, 2016 5 commits
-
-
Seth Forshee authored
Security labels from unprivileged mounts in user namespaces must be ignored. Force superblocks from user namespaces whose labeling behavior is to use xattrs to use mountpoint labeling instead. For the mountpoint label, default to converting the current task context into a form suitable for file objects, but also allow the policy writer to specify a different label through policy transition rules. Pieced together from code snippets provided by Stephen Smalley. Signed-off-by:
Stephen Smalley <sds@tycho.nsa.gov> Acked-by:
James Morris <james.l.morris@oracle.com> Signed-off-by:
-
Seth Forshee authored
The SMACK64, SMACK64EXEC, and SMACK64MMAP labels are all handled differently in untrusted mounts. This is confusing and potentically problematic. Change this to handle them all the same way that SMACK64 is currently handled; that is, read the label from disk and check it at use time. For SMACK64 and SMACK64MMAP access is denied if the label does not match smk_root. To be consistent with suid, a SMACK64EXEC label which does not match smk_root will still allow execution of the file but will not run with the label supplied in the xattr. Signed-off-by:
Casey Schaufler <casey@schaufler-ca.com> Signed-off-by:
-
Seth Forshee authored
Security labels from unprivileged mounts cannot be trusted. Ideally for these mounts we would assign the objects in the filesystem the same label as the inode for the backing device passed to mount. Unfortunately it's currently impossible to determine which inode this is from the LSM mount hooks, so we settle for the label of the process doing the mount. This label is assigned to s_root, and also to smk_default to ensure that new inodes receive this label. The transmute property is also set on s_root to make this behavior more explicit, even though it is technically not necessary. If a filesystem has existing security labels, access to inodes is permitted if the label is the same as smk_root, otherwise access is denied. The SMACK64EXEC xattr is completely ignored. Explicit setting of security labels continues to require CAP_MAC_ADMIN in init_user_ns. Altogether, this ensures that filesystem objects are not accessible to subjects which cannot already access the backing store, that MAC is not violated for any objects in the fileystem which are already labeled, and that a user cannot use an unprivileged mount to gain elevated MAC privileges. sysfs, tmpfs, and ramfs are already mountable from user namespaces and support security labels. We can't rule out the possibility that these filesystems may already be used in mounts from user namespaces with security lables set from the init namespace, so failing to trust lables in these filesystems may introduce regressions. It is safe to trust labels from these filesystems, since the unprivileged user does not control the backing store and thus cannot supply security labels, so an explicit exception is made to trust labels from these filesystems. Signed-off-by:
-
Andy Lutomirski authored
If a process gets access to a mount from a different user namespace, that process should not be able to take advantage of setuid files or selinux entrypoints from that filesystem. Prevent this by treating mounts from other mount namespaces and those not owned by current_user_ns() or an ancestor as nosuid. This will make it safer to allow more complex filesystems to be mounted in non-root user namespaces. This does not remove the need for MNT_LOCK_NOSUID. The setuid, setgid, and file capability bits can no longer be abused if code in a user namespace were to clear nosuid on an untrusted filesystem, but this patch, by itself, is insufficient to protect the system from abuse of files that, when execed, would increase MAC privilege. As a more concrete explanation, any task that can manipulate a vfsmount associated with a given user namespace already has capabilities in that namespace and all of its descendents. If they can cause a malicious setuid, setgid, or file-caps executable to appear in that mount, then that executable will only allow them to elevate privileges in exactly the set of namespaces in which they are already privileges. On the other hand, if they can cause a malicious executable to appear with a dangerous MAC label, running it could change the caller's security context in a way that should not have been possible, even inside the namespace in which the task is confined. As a hardening measure, this would have made CVE-2014-5207 much more difficult to exploit. Signed-off-by:
Andy Lutomirski <luto@amacapital.net> Signed-off-by:
-
Seth Forshee authored
Capability sets attached to files must be ignored except in the user namespaces where the mounter is privileged, i.e. s_user_ns and its descendants. Otherwise a vector exists for gaining privileges in namespaces where a user is not already privileged. Add a new helper function, current_in_user_ns(), to test whether a user namespace is the same as or a descendant of another namespace. Use this helper to determine whether a file's capability set should be applied to the caps constructed during exec. --EWB Replaced in_userns with the simpler current_in_userns. Acked-by:
-
- 23 Jun, 2016 12 commits
-
-
Eric W. Biederman authored
Now that SB_I_NODEV controls the nodev behavior devpts can just clear this flag during mount. Simplifying the code and making it easier to audit how the code works. While still preserving the invariant that s_iflags is only modified during mount. Acked-by:
-
Eric W. Biederman authored
Replace the implict setting of MNT_NODEV on mounts that happen with just user namespace permissions with an implicit setting of SB_I_NODEV in s_iflags. The visibility of the implicit MNT_NODEV has caused problems in the past. With this change the fragile case where an implicit MNT_NODEV needs to be preserved in do_remount is removed. Using SB_I_NODEV is much less fragile as s_iflags are set during the original mount and never changed. In do_new_mount with the implicit setting of MNT_NODEV gone, the only code that can affect mnt_flags is fs_fully_visible so simplify the if statement and reduce the indentation of the code to make that clear. Acked-by:
-
Eric W. Biederman authored
Verify all filesystems that we check in mount_too_revealing set SB_I_NOEXEC and SB_I_NODEV in sb->s_iflags. That is true for today and it should remain true in the future. Remove the now unnecessary checks from mnt_already_visibile that ensure MNT_LOCK_NOSUID, MNT_LOCK_NOEXEC, and MNT_LOCK_NODEV are preserved. Making the code shorter and easier to read. Relying on SB_I_NOEXEC and SB_I_NODEV instead of the user visible MNT_NOSUID, MNT_NOEXEC, and MNT_NODEV ensures the many current systems where proc and sysfs are mounted with "nosuid, nodev, noexec" and several slightly buggy container applications don't bother to set those flags continue to work. Acked-by:
-
Eric W. Biederman authored
Introduce a function may_open_dev that tests MNT_NODEV and a new superblock flab SB_I_NODEV. Use this new function in all of the places where MNT_NODEV was previously tested. Add the new SB_I_NODEV s_iflag to proc, sysfs, and mqueuefs as those filesystems should never support device nodes, and a simple superblock flags makes that very hard to get wrong. With SB_I_NODEV set if any device nodes somehow manage to show up on on a filesystem those device nodes will be unopenable. Acked-by:
-
Eric W. Biederman authored
Set SB_I_NOEXEC on mqueuefs to ensure small implementation mistakes do not result in executable on mqueuefs by accident. Acked-by:
-
Eric W. Biederman authored
The cgroup filesystem is in the same boat as sysfs. No one ever permits executables of any kind on the cgroup filesystem, and there is no reasonable future case to support executables in the future. Therefore move the setting of SB_I_NOEXEC which makes the code proof against future mistakes of accidentally creating executables from sysfs to kernfs itself. Making the code simpler and covering the sysfs, cgroup, and cgroup2 filesystems. Acked-by:
-
Eric W. Biederman authored
Allowing a filesystem to be mounted by other than root in the initial user namespace is a filesystem property not a mount namespace property and as such should be checked in filesystem specific code. Move the FS_USERNS_MOUNT test into super.c:sget_userns(). Acked-by:
-
Eric W. Biederman authored
Start marking filesystems with a user namespace owner, s_user_ns. In this change this is only used for permission checks of who may mount a filesystem. Ultimately s_user_ns will be used for translating ids and checking capabilities for filesystems mounted from user namespaces. The default policy for setting s_user_ns is implemented in sget(), which arranges for s_user_ns to be set to current_user_ns() and to ensure that the mounter of the filesystem has CAP_SYS_ADMIN in that user_ns. The guts of sget are split out into another function sget_userns(). The function sget_userns calls alloc_super with the specified user namespace or it verifies the existing superblock that was found has the expected user namespace, and fails with EBUSY when it is not. This failing prevents users with the wrong privileges mounting a filesystem. The reason for the split of sget_userns from sget is that in some cases such as mount_ns and kernfs_mount_ns a different policy for permission checking of mounts and setting s_user_ns is necessary, and the existence of sget_userns() allows those policies to be implemented. The helper mount_ns is expected to be used for filesystems such as proc and mqueuefs which present per namespace information. The function mount_ns is modified to call sget_userns instead of sget to ensure the user namespace owner of the namespace whose information is presented by the filesystem is used on the superblock. For sysfs and cgroup the appropriate permission checks are already in place, and kernfs_mount_ns is modified to call sget_userns so that the init_user_ns is the only user namespace used. For the cgroup filesystem cgroup namespace mounts are bind mounts of a subset of the full cgroup filesystem and as such s_user_ns must be the same for all of them as there is only a single superblock. Mounts of sysfs that vary based on the network namespace could in principle change s_user_ns but it keeps the analysis and implementation of kernfs simpler if that is not supported, and at present there appear to be no benefits from supporting a different s_user_ns on any sysfs mount. Getting the details of setting s_user_ns correct has been a long process. Thanks to Pavel Tikhorirorv who spotted a leak in sget_userns. Thanks to Seth Forshee who has kept the work alive. Thanks-to: Seth Forshee <seth.forshee@canonical.com> Thanks-to: Pavel Tikhomirov <ptikhomirov@virtuozzo.com> Acked-by:
-
Eric W. Biederman authored
Move the call of get_pid_ns, the call of proc_parse_options, and the setting of s_iflags into proc_fill_super so that mount_ns can be used. Convert proc_mount to call mount_ns and remove the now unnecessary code. Acked-by:
Djalal Harouni <tixxdz@gmail.com> Signed-off-by:
-
Eric W. Biederman authored
Today what is normally called data (the mount options) is not passed to fill_super through mount_ns. Pass the mount options and the namespace separately to mount_ns so that filesystems such as proc that have mount options, can use mount_ns. Pass the user namespace to mount_ns so that the standard permission check that verifies the mounter has permissions over the namespace can be performed in mount_ns instead of in each filesystems .mount method. Thus removing the duplication between mqueuefs and proc in terms of permission checks. The extra permission check does not currently affect the rpc_pipefs filesystem and the nfsd filesystem as those filesystems do not currently allow unprivileged mounts. Without unpvileged mounts it is guaranteed that the caller has already passed capable(CAP_SYS_ADMIN) which guarantees extra permission check will pass. Update rpc_pipefs and the nfsd filesystem to ensure that the network namespace reference is always taken in fill_super and always put in kill_sb so that the logic is simpler and so that errors originating inside of fill_super do not cause a network namespace leak. Acked-by:
-
Eric W. Biederman authored
Allow the ipc namespace initialization code to depend on ns->user_ns being set during initialization. In particular this allows mq_init_ns to use ns->user_ns for permission checks and initializating s_user_ns while the the mq filesystem is being mounted. Acked-by:
-
Eric W. Biederman authored
Replace the call of fs_fully_visible in do_new_mount from before the new superblock is allocated with a call of mount_too_revealing after the superblock is allocated. This winds up being a much better location for maintainability of the code. The first change this enables is the replacement of FS_USERNS_VISIBLE with SB_I_USERNS_VISIBLE. Moving the flag from struct filesystem_type to sb_iflags on the superblock. Unfortunately mount_too_revealing fundamentally needs to touch mnt_flags adding several MNT_LOCKED_XXX flags at the appropriate times. If the mnt_flags did not need to be touched the code could be easily moved into the filesystem specific mount code. Acked-by:
-
- 15 Jun, 2016 1 commit
-
-
Eric W. Biederman authored
In rare cases it is possible for s_flags & MS_RDONLY to be set but MNT_READONLY to be clear. This starting combination can cause fs_fully_visible to fail to ensure that the new mount is readonly. Therefore force MNT_LOCK_READONLY in the new mount if MS_RDONLY is set on the source filesystem of the mount. In general both MS_RDONLY and MNT_READONLY are set at the same for mounts so I don't expect any programs to care. Nor do I expect MS_RDONLY to be set on proc or sysfs in the initial user namespace, which further decreases the likelyhood of problems. Which means this change should only affect system configurations by paranoid sysadmins who should welcome the additional protection as it keeps people from wriggling out of their policies. Cc: stable@vger.kernel.org Fixes: 8c6cf9cc ("mnt: Modify fs_fully_visible to deal with locked ro nodev and atime") Signed-off-by:
-
- 07 Jun, 2016 2 commits
-
-
Eric W. Biederman authored
MNT_LOCKED implies on a child mount implies the child is locked to the parent. So while looping through the children the children should be tested (not their parent). Typically an unshare of a mount namespace locks all mounts together making both the parent and the slave as locked but there are a few corner cases where other things work. Cc: stable@vger.kernel.org Fixes: ceeb0e5d ("vfs: Ignore unlocked mounts in fs_fully_visible") Reported-by:
-
Eric W. Biederman authored
Add this trivial missing error handling. Cc: stable@vger.kernel.org Fixes: 1b852bce ("mnt: Refactor the logic for mounting sysfs and proc in a user namespace") Signed-off-by:
-
- 05 Jun, 2016 7 commits
-
-
Linus Torvalds authored
-
git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linuxLinus Torvalds authored
Pull parisc fixes from Helge Deller: - Fix printk time stamps on SMP systems which got wrong due to a patch which was added during the merge window - Fix two bugs in the stack backtrace code: Races in module unloading and possible invalid accesses to memory due to wrong instruction decoding (Mikulas Patocka) - Fix userspace crash when syscalls access invalid unaligned userspace addresses. Those syscalls will now return EFAULT as expected. (tagged for stable kernel series) * 'parisc-4.7-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux: parisc: Move die_if_kernel() prototype into traps.h header parisc: Fix pagefault crash in unaligned __get_user() call parisc: Fix printk time during boot parisc: Fix backtrace on PA-RISC
-
git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-securityLinus Torvalds authored
Pull key handling update from James Morris: "This alters a new keyctl function added in the current merge window to allow for a future extension planned for the next merge window" * 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security: KEYS: Add placeholder for KDF usage with DH
-
Eric W. Biederman authored
The /dev/ptmx device node is changed to lookup the directory entry "pts" in the same directory as the /dev/ptmx device node was opened in. If there is a "pts" entry and that entry is a devpts filesystem /dev/ptmx uses that filesystem. Otherwise the open of /dev/ptmx fails. The DEVPTS_MULTIPLE_INSTANCES configuration option is removed, so that userspace can now safely depend on each mount of devpts creating a new instance of the filesystem. Each mount of devpts is now a separate and equal filesystem. Reserved ttys are now available to all instances of devpts where the mounter is in the initial mount namespace. A new vfs helper path_pts is introduced that finds a directory entry named "pts" in the directory of the passed in path, and changes the passed in path to point to it. The helper path_pts uses a function path_parent_directory that was factored out of follow_dotdot. In the implementation of devpts: - devpts_mnt is killed as it is no longer meaningful if all mounts of devpts are equal. - pts_sb_from_inode is replaced by just inode->i_sb as all cached inodes in the tty layer are now from the devpts filesystem. - devpts_add_ref is rolled into the new function devpts_ptmx. And the unnecessary inode hold is removed. - devpts_del_ref is renamed devpts_release and reduced to just a deacrivate_super. - The newinstance mount option continues to be accepted but is now ignored. In devpts_fs.h definitions for when !CONFIG_UNIX98_PTYS are removed as they are never used. Documentation/filesystems/devices.txt is updated to describe the current situation. This has been verified to work properly on openwrt-15.05, centos5, centos6, centos7, debian-6.0.2, debian-7.9, debian-8.2, ubuntu-14.04.3, ubuntu-15.10, fedora23, magia-5, mint-17.3, opensuse-42.1, slackware-14.1, gentoo-20151225 (13.0?), archlinux-2015-12-01. With the caveat that on centos6 and on slackware-14.1 that there wind up being two instances of the devpts filesystem mounted on /dev/pts, the lower copy does not end up getting used. Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Helge Deller authored
Signed-off-by:Helge Deller <deller@gmx.de>
-
Helge Deller authored
One of the debian buildd servers had this crash in the syslog without any other information: Unaligned handler failed, ret = -2 clock_adjtime (pid 22578): Unaligned data reference (code 28) CPU: 1 PID: 22578 Comm: clock_adjtime Tainted: G E 4.5.0-2-parisc64-smp #1 Debian 4.5.4-1 task: 000000007d9960f8 ti: 00000001bde7c000 task.ti: 00000001bde7c000 YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI PSW: 00001000000001001111100000001111 Tainted: G E r00-03 000000ff0804f80f 00000001bde7c2b0 00000000402d2be8 00000001bde7c2b0 r04-07 00000000409e1fd0 00000000fa6f7fff 00000001bde7c148 00000000fa6f7fff r08-11 0000000000000000 00000000ffffffff 00000000fac9bb7b 000000000002b4d4 r12-15 000000000015241c 000000000015242c 000000000000002d 00000000fac9bb7b r16-19 0000000000028800 0000000000000001 0000000000000070 00000001bde7c218 r20-23 0000000000000000 00000001bde7c210 0000000000000002 0000000000000000 r24-27 0000000000000000 0000000000000000 00000001bde7c148 00000000409e1fd0 r28-31 0000000000000001 00000001bde7c320 00000001bde7c350 00000001bde7c218 sr00-03 0000000001200000 0000000001200000 0000000000000000 0000000001200000 sr04-07 0000000000000000 0000000000000000 0000000000000000 0000000000000000 IASQ: 0000000000000000 0000000000000000 IAOQ: 00000000402d2e84 00000000402d2e88 IIR: 0ca0d089 ISR: 0000000001200000 IOR: 00000000fa6f7fff CPU: 1 CR30: 00000001bde7c000 CR31: ffffffffffffffff ORIG_R28: 00000002369fe628 IAOQ[0]: compat_get_timex+0x2dc/0x3c0 IAOQ[1]: compat_get_timex+0x2e0/0x3c0 RP(r2): compat_get_timex+0x40/0x3c0 Backtrace: [<00000000402d4608>] compat_SyS_clock_adjtime+0x40/0xc0 [<0000000040205024>] syscall_exit+0x0/0x14 This means the userspace program clock_adjtime called the clock_adjtime() syscall and then crashed inside the compat_get_timex() function. Syscalls should never crash programs, but instead return EFAULT. The IIR register contains the executed instruction, which disassebles into "ldw 0(sr3,r5),r9". This load-word instruction is part of __get_user() which tried to read the word at %r5/IOR (0xfa6f7fff). This means the unaligned handler jumped in. The unaligned handler is able to emulate all ldw instructions, but it fails if it fails to read the source e.g. because of page fault. The following program reproduces the problem: #define _GNU_SOURCE #include <unistd.h> #include <sys/syscall.h> #include <sys/mman.h> int main(void) { /* allocate 8k */ char *ptr = mmap(NULL, 2*4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0); /* free second half (upper 4k) and make it invalid. */ munmap(ptr+4096, 4096); /* syscall where first int is unaligned and clobbers into invalid memory region */ /* syscall should return EFAULT */ return syscall(__NR_clock_adjtime, 0, ptr+4095); } To fix this issue we simply need to check if the faulting instruction address is in the exception fixup table when the unaligned handler failed. If it is, call the fixup routine instead of crashing. While looking at the unaligned handler I found another issue as well: The target register should not be modified if the handler was unsuccessful. Signed-off-by: -
Helge Deller authored
Avoid showing invalid printk time stamps during boot. Signed-off-by:
Aaro Koskinen <aaro.koskinen@iki.fi>
-
