The current overflow checking is broken in the corner case of the strings'
combined length being exactly SIZE_MAX: After the loop, l would be SIZE_MAX,
but we're not testing whether the l+1 expression overflows.
Fix it by simply pre-accounting for the final '\0': initialize l to 1 instead
of 0.
The loops over (x, then all varargs, until a NULL is found) can be written much
simpler with an ordinary for loop. Just initialize the loop variable to x, test
that, and in the increment part, fetch the next va_arg(). That removes a level
of indentation, and avoids doing a separate strlen()/stpcpy() call for x.
While touching this code anyway, change (size_t)-1 to the more readable
SIZE_MAX.
This reverts commit c0443b97b7.
I got various cases wrong:
"usb:v04F3p2B7Cd5912dc00dsc00dp00ic03isc00ip00in00"
"usb:v0627p0001:QEMU USB Tablet"
"input:b0003v0627p0001e0001-e0,1,2,4,k110,111,112,r0,1,8,B,am4,lsfw"
OTOH:
-evdev:name:ETPS/2 Elantech Touchpad:dmi:*svnASUSTeKComputerInc.:pnN53SV:*
+evdev:name:ETPS/2 Elantech Touchpad:dmi:*svnASUSTeKComputerInc.:pnN53SV*
is OK. Other parts follow after 'pn'.
-mouse:*:name:*Trackball*:*
-mouse:*:name:*trackball*:*
-mouse:*:name:*TrackBall*:*
+mouse:*:name:*Trackball*:
+mouse:*:name:*trackball*:
+mouse:*:name:*TrackBall*:
... and anything else with :name should be OK too, because our imports always
include ":" at the end:
IMPORT{builtin}="hwdb 'joystick:$env{ID_BUS}:v$attr{id/vendor}p$attr{id/product}:name:$attr{name}:'"
Including '*' at the end makes the pattern work even if we decide to add
something to the match string later.
Fixes#17499.
This beefs up the READ_FULL_FILE_CONNECT_SOCKET logic of
read_full_file_full() a bit: when used a sender socket name may be
specified. If specified as NULL behaviour is as before: the client
socket name is picked by the kernel. But if specified as non-NULL the
client can pick a socket name to use when connecting. This is useful to
communicate a minimal amount of metainformation from client to server,
outside of the transport payload.
Specifically, these beefs up the service credential logic to pass an
abstract AF_UNIX socket name as client socket name when connecting via
READ_FULL_FILE_CONNECT_SOCKET, that includes the requesting unit name
and the eventual credential name. This allows servers implementing the
trivial credential socket logic to distinguish clients: via a simple
getpeername() it can be determined which unit is requesting a
credential, and which credential specifically.
Example: with this patch in place, in a unit file "waldo.service" a
configuration line like the following:
LoadCredential=foo:/run/quux/creds.sock
will result in a connection to the AF_UNIX socket /run/quux/creds.sock,
originating from an abstract namespace AF_UNIX socket:
@$RANDOM/unit/waldo.service/foo
(The $RANDOM is replaced by some randomized string. This is included in
the socket name order to avoid namespace squatting issues: the abstract
socket namespace is open to unprivileged users after all, and care needs
to be taken not to use guessable names)
The services listening on the /run/quux/creds.sock socket may thus
easily retrieve the name of the unit the credential is requested for
plus the credential name, via a simpler getpeername(), discarding the
random preifx and the /unit/ string.
This logic uses "/" as separator between the fields, since both unit
names and credential names appear in the file system, and thus are
designed to use "/" as outer separators. Given that it's a good safe
choice to use as separators here, too avoid any conflicts.
This is a minimal patch only: the new logic is used only for the unit
file credential logic. For other places where we use
READ_FULL_FILE_CONNECT_SOCKET it is probably a good idea to use this
scheme too, but this should be done carefully in later patches, since
the socket names become API that way, and we should determine the right
amount of info to pass over.
umount emits an error message "no mount point specified" if the
tmpfs isn't mounted yet, which is the normal case.
Suppress that by redirecting stderr.
Script for generating LOTS of SCSI disk and partition devices in
the fake sysfs we use for udev testing.
This script is supposed to be run after sys-script.py. It uses
code from sys-script.py as template to generate additional SCSI disk data
structures. Together with the "generator" code in udev-test.pl
added in the previous patch, it allows to run udev tests with almost
arbitrarily many devices, and thus to do performance scaling tests.
Manually listing all devices in the test definition becomes cumbersome with
lots of devices. Add a function that scans on all block devices in
the test sysfs and generates a list of devices to test.
Add 4 new tests using multiple devices. Number 2-4 use many
devices claiming the same symlink, where only one device has
a higher priority thatn the others. They fail sporadically with
the current code, if a race condition causes the symlink to point
to the wrong device. Test 4 is like test 2 with sleeps in between,
it's much less likely to fail.
the "last_rule" option hasn't been supported for some time.
Therefore this test fails if a "not_exp_links" attribute is added,
as it should be. Mark it appropriately.
Instead of testing the existence or non-exisitence of just a single
symlink, allow testing of several links per device.
Change the test definitions accordingly.
Test if symlinks are created correctly by comparing the symlink
targets to the devnode path. This implies (for the symlink) that
major/minor numbers and permissions are correct, as we have tested
that on the devnode already.
More often than not, the created devnode is the basename of the
sysfs entry. The "devnode" device may be used to override the
auto-detected node name.
Permissions and major/minor number are now verified on the devnode
itself, not on symlinks.
For those tests where exp_name is set to the computed devnode name,
the explicit "exp_name" can be removed. "exp_name" is only required for
symlinks.
This allows separate testing for devnodes and symlinks an a follow-up
patch.
Allow testing cases where multiple devices are added and removed
simultaneously. Tests are started as synchronously as possible using a
semaphore, in order to test possible race conditions. If this isn't desired,
the test parameter "sleep_us" can be set to the number of microseconds to wait
between udev invocations.
Allow testing cases where multiple devices are added and removed.
This implies a change of the data structure: every test allows
for multiple devices to be added, and "exp_name" etc. are now properties
of the device, not of the test.
When selinux is enabled, the call of
manager_rtnl_enumerate_nexthop() fails.
This fix is to facilitate selinux hook handling for enumerating
nexthop.
In manager_rtnl_enumerate_nexthop() there is a check
if "Not supported" is returned by the send_netlink() call.
This check expects that -EOPNOTSUPP is returned,
the selinux hook seems to return -EINVAL instead.
This happens in kernel older than 5.3
(more specificallytorvalds/linux@65ee00a) as it does not support
nexthop handling through netlink.
And if SELinux is enforced in the order kernel, callingRTM_GETNEXTHOP
returns -EINVAL.
Thus adding a call in the manager_rtnl_enumerate_nexthop for the
extra return -EINVAL.
* Existing valid rule files written with KEY="value" are not affected
* Now, KEY=e"value\n" becomes valid. Where `\n` is a newline character
* Escape sequences supported by src/basic/escape.h:cunescape() is
supported
Lennart Poettering