Slot already mounted detection fails with symlinks #388
Labels
Comments
|
A solution to this would be to normalize device paths before comparing them. This could be done by recursively following symlinks, i.e. realpath(), but that does not catch a number of cases. The mounted device could be a file which is loopback mounted (which is already detected by rauc). This file could have multiple hard links to it. That would not be detected by comparing realpaths. There is no rule that only one device node of a given type/major/minor combo can exist. It's perfectly possible to mknod multiple nodes for the same underlying device. This, much like multiple hardlinks to a file, it not detected by comparing realpaths. Here's a solution which allows things that can be mounted to be compared in a way that is immune to the above and not terribly complex. // Something that is, or can be, mounted onto a mount point
typedef struct {
gboolean is_device; // vs being a loop mounted file
dev_t dev; // the device, or for a file, the device the file is on
ino_t inode; // inode of file for a non-device
} MountableObj;
// Take a device (or file) path and normalize it
static MountableObj *normalize_mountable_object(const gchar *devicepath)
{
MountableObj *obj = g_new0(MountableObj, 1);
GStatBuf st;
if (g_stat(devicepath, &st) == -1) {
/* Virtual filesystems like devpts trigger this case */
g_debug("Can't stat '%s', must not be mountable", devicepath);
return NULL;
}
if (S_ISBLK(st.st_mode)) {
obj->is_device = TRUE;
obj->dev = st.st_rdev;
} else if (S_ISREG(st.st_mode)) {
obj->is_device = FALSE;
obj->dev = st.st_dev;
obj->inode = st.st_ino;
} else {
g_debug("Device '%s' is not something which is mountable", devicepath);
g_clear_pointer(&obj, g_free);
}
return obj;
}
/* Compare two MountableObj for equality */
static gboolean is_same_mountable_object(const MountableObj *a, const MountableObj *b)
{
return a->is_device == b->is_device && a->dev == b->dev &&
(a->is_device || a->inode == b->inode);
} |
|
I could send a pull request to fix this in the manner I described above if it appears acceptable. |
xyzzy42
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Feb 8, 2019
This way a mounted slot can be found it the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
xyzzy42
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Feb 11, 2019
This way a mounted slot can be found it the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
xyzzy42
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Feb 11, 2019
This way a mounted slot can be found it the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
xyzzy42
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Feb 11, 2019
This way a mounted slot can be found it the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
xyzzy42
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Feb 25, 2019
This way a mounted slot can be found if the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. This is a fallback after simplying comparing the paths, as if the paths are the same then the devices must the same too. This avoids chaning behavior in cases there the devices do not exist, for instance if rauc status before a device is online. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
xyzzy42
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Feb 25, 2019
This way a mounted slot can be found if the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. This is a fallback after simplying comparing the paths, as if the paths are the same then the devices must the same too. This avoids chaning behavior in cases there the devices do not exist, for instance if rauc status before a device is online. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
xyzzy42
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Mar 20, 2019
This way a mounted slot can be found if the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. This is a fallback after simplying comparing the paths, as if the paths are the same then the devices must the same too. This avoids chaning behavior in cases there the devices do not exist, for instance if rauc status before a device is online. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
xyzzy42
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Mar 21, 2019
This way a mounted slot can be found if the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. This is a fallback after simplying comparing the paths, as if the paths are the same then the devices must the same too. This avoids chaning behavior in cases there the devices do not exist, for instance if rauc status before a device is online. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
xyzzy42
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Mar 21, 2019
This way a mounted slot can be found if the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. This is a fallback after simplying comparing the paths, as if the paths are the same then the devices must the same too. This avoids chaning behavior in cases there the devices do not exist, for instance if rauc status before a device is online. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
jluebbe
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Feb 10, 2020
This way a mounted slot can be found if the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. This is a fallback after simply comparing the paths, as if the paths are the same then the devices must the same too. This avoids changing behavior in cases there the devices do not exist, for instance if rauc status before a device is online. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
jluebbe
pushed a commit
to xyzzy42/rauc
that referenced
this issue
Apr 21, 2020
This way a mounted slot can be found if the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. This is a fallback after simply comparing the paths, as if the paths are the same then the devices must the same too. This avoids changing behavior in cases there the devices do not exist, for instance if rauc status before a device is online. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
ejoerns
pushed a commit
to ejoerns/rauc
that referenced
this issue
Aug 30, 2020
This way a mounted slot can be found if the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. This is a fallback after simply comparing the paths, as if the paths are the same then the devices must the same too. This avoids changing behavior in cases there the devices do not exist, for instance if rauc status before a device is online. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
ejoerns
pushed a commit
to ejoerns/rauc
that referenced
this issue
Apr 12, 2021
This way a mounted slot can be found if the path used to mount the device is not the same as the path in the rauc configuration file. E.g., the configuration file uses a link like /dev/disk/by-partlabel/rootfs0 while the kernel's list of mounted devices uses a name like /dev/mmcblk1p2. This is a fallback after simply comparing the paths, as if the paths are the same then the devices must the same too. This avoids changing behavior in cases there the devices do not exist, for instance if rauc status before a device is online. Not being able to detect that the current rootfs is already mounted causes rauc to try to mount the slot, which will fail. Fixes rauc#388. Signed-off-by: Trent Piepho <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I've configured a slot to have a device path which is a symlink, such as
/dev/disk/by-partlabel/rootfs0. Rauc tries to determine if this slot is already mounted, but it does this by comparing the device name, like above, with a list of mounted device fromg_unix_mounts_get(). But that list will use a device name like/dev/mmcblk1p2, since it's not created by the path which was mounted, but rather reconstructed from the major/minor of the device. A simple string comparison of the two paths fails, even though they are the same device.This results in
determine_slot_states()failing to detect the existing mount and setext_mount_point. Later,load_slot_status_locally()will try to remount the already mounted filesystem and fail.The text was updated successfully, but these errors were encountered: