picnoir
/
Systemd
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

capability: introduce CAP_TO_MASK_CORRECTED() macro replacing CAP_TO_MASK() 

linux/capability.h's CAP_TO_MASK potentially shifts a signed int "1"
(i.e. 32bit wide) left by 31 which means it becomes negative. That's
just weird, and ubsan complains about it. Let's introduce our own macro
CAP_TO_MASK_CORRECTED which doesn't fall into this trap, and make use of
it.

Fixes: #10347
This commit is contained in:
Lennart Poettering 2018-10-10 11:07:54 +02:00
parent ca92fe36e0
commit 5f00c5684f
2 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/basic/capability-util.h
View File

@ -39,3 +39,7 @@ static inline bool cap_test_all(uint64_t caps) {
}
bool ambient_capabilities_supported(void);
/* Identical to linux/capability.h's CAP_TO_MASK(), but uses an unsigned 1U instead of a signed 1 for shifting left, in
* order to avoid complaints about shifting a signed int left by 31 bits, which would make it negative. */
#define CAP_TO_MASK_CORRECTED(x) (1U << ((x) & 31U))

2
src/libsystemd/sd-bus/bus-creds.c
View File

@ -661,7 +661,7 @@ static int has_cap(sd_bus_creds *c, unsigned offset, int capability) {
sz = DIV_ROUND_UP(cap_last_cap(), 32U);
return !!(c->capability[offset * sz + CAP_TO_INDEX(capability)] & CAP_TO_MASK(capability));
return !!(c->capability[offset * sz + CAP_TO_INDEX((uint32_t) capability)] & CAP_TO_MASK_CORRECTED((uint32_t) capability));
}
_public_ int sd_bus_creds_has_effective_cap(sd_bus_creds *c, int capability) {