See: https://godbolt.org/z/P5nGMohxd Am seeing false positive with -O1 -fanalyzer -fanalyzer-checker=taint on this code: struct raw_ep { /* ...snip... */ int state; /* ...snip... */ }; struct raw_dev { /* ...snip... */ struct raw_ep eps[30]; int eps_num; /* ...snip... */ }; int __attribute__((tainted_args)) simplified_raw_ioctl_ep_disable(struct raw_dev *dev, unsigned long value) { int ret = 0, i = value; if (i < 0 || i >= dev->eps_num) { ret = -16; goto out_unlock; } if (dev->eps[i].state == 0) { ret = -22; goto out_unlock; } out_unlock: return ret; } ../../src/raw_gadget.c: In function ‘simplified_raw_ioctl_ep_disable’: ../../src/raw_gadget.c:23:18: warning: use of attacker-controlled value ‘value’ in array lookup without upper-bounds checking [CWE-129] [-Wanalyzer-tainted-array-index] 23 | if (dev->eps[i].state == 0) { | ~~~~~~~~~~~^~~~~~ ‘simplified_raw_ioctl_ep_disable’: event 1 | | 15 | simplified_raw_ioctl_ep_disable(struct raw_dev *dev, unsigned long value) | | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | | | | | (1) function ‘simplified_raw_ioctl_ep_disable’ marked with ‘__attribute__((tainted_args))’ | +--> ‘simplified_raw_ioctl_ep_disable’: events 2-5 | | 15 | simplified_raw_ioctl_ep_disable(struct raw_dev *dev, unsigned long value) | | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | | | | | (2) entry to ‘simplified_raw_ioctl_ep_disable’ |...... | 19 | if (i < 0 || i >= dev->eps_num) { | | ~ | | | | | (3) following ‘false’ branch... |...... | 23 | if (dev->eps[i].state == 0) { | | ~~~~~~~~~~~~~~~~~ | | | | | (4) ...to here | | (5) use of attacker-controlled value ‘value’ in array lookup without upper-bounds checking | Seems to need at least -O1. Reduced from false positives seen in various ioctl handlers in Linux kernel in drivers/usb/gadget/legacy/raw_gadget.c (with "allyesconfig").
The issue isn't the casts; it's the comparison against non-const, which looks like: _1 = dev_6(D)->eps_num; if (_1 <= i_4(D)) at the gimple level, and where i_4(D) is on the RHS, and sm-taint.cc is only looking at comparisons of the LHS. I'm testing a fix.
The master branch has been updated by David Malcolm <dmalcolm@gcc.gnu.org>: https://gcc.gnu.org/g:5e830693dd335621940368b6d39b23afc2c98545 commit r13-1768-g5e830693dd335621940368b6d39b23afc2c98545 Author: David Malcolm <dmalcolm@redhat.com> Date: Wed Jul 20 17:25:35 2022 -0400 analyzer: update "tainted" state of RHS in comparisons [PR106373] Doing so fixes various false positives from -Wanalyzer-tainted-array-index at -O1 and above (e.g. seen on the Linux kernel) gcc/analyzer/ChangeLog: PR analyzer/106373 * sm-taint.cc (taint_state_machine::on_condition): Potentially update the state of the RHS as well as the LHS. gcc/testsuite/ChangeLog: PR analyzer/106373 * gcc.dg/analyzer/torture/taint-read-index-3.c: New test. Signed-off-by: David Malcolm <dmalcolm@redhat.com>
Should be fixed on trunk by the above patch.