diff options
author | Adam Borowski <kilobyte@angband.pl> | 2016-09-15 16:47:09 +0200 |
---|---|---|
committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2016-09-22 11:41:54 +0200 |
commit | 669e0a51b1b50052b1615683cde64e1c28ae895f (patch) | |
tree | 63c34a3da13db5cb27ae308aa462d8a1275f1464 /drivers/tty | |
parent | d2ec3f77de8e67b7a3dab3ec827467e0fd797c86 (diff) |
vt: Fix a read-past-array in vc_t416_color().
This makes it show up on UBSAN:
perl -e 'for (0..15) {my @x=("0")x$_;push @x,qw(38 2 64 128 192 4);printf
"\e[%smAfter %d zeroes.\e[0m\n", join(";",@x[0..($_+5<15?$_+5:15)]), $_}'
Seems harmless: if you can programmatically read attributes of a vt
character (/dev/vcsa*), multiple probes can obtain parts of vt_mode then
lowest byte (5th on 64-bit big-endian) of a pointer.
Signed-off-by: Adam Borowski <kilobyte@angband.pl>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Diffstat (limited to 'drivers/tty')
-rw-r--r-- | drivers/tty/vt/vt.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/drivers/tty/vt/vt.c b/drivers/tty/vt/vt.c index 2705ca960e92..b51586fea4e8 100644 --- a/drivers/tty/vt/vt.c +++ b/drivers/tty/vt/vt.c @@ -1316,7 +1316,7 @@ static int vc_t416_color(struct vc_data *vc, int i, /* 256 colours -- ubiquitous */ i++; rgb_from_256(vc->vc_par[i], &c); - } else if (vc->vc_par[i] == 2 && i <= vc->vc_npar + 3) { + } else if (vc->vc_par[i] == 2 && i + 3 <= vc->vc_npar) { /* 24 bit -- extremely rare */ c.r = vc->vc_par[i + 1]; c.g = vc->vc_par[i + 2]; |