emo/procps
1
0
Fork 0
Code Issues Pull Requests Packages Releases Activity

top: concede integer overflow risks in procs_refresh()

Browse Source 
This is as far as we need go with respect to the issue
of integer overflow addressed in that reference below.

That patch, of course, was reversed to prepare for us.

Reference(s):
. original qualys patch
0105-top-Prevent-integer-overflows-in-procs_refresh.patch
commit 131e5e2fe6

Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit is contained in:
Jim Warner 2018-05-18 00:00:00 -05:00 committed by Craig Small
parent c9dfcdebdc
commit 59f39da852
1 changed files with 4 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
top/top.c
View File

@ -2797,6 +2797,8 @@ static void procs_hlp (proc_t *this) {
}
if (Frame_maxtask+1 >= HHist_siz) {
/* we're subject to integer overflow if total linux tasks ever approach |
400+ million (but, do you think memory might be the bigger problem?) | */
HHist_siz = HHist_siz * 5 / 4 + 100;
PHist_sav = alloc_r(PHist_sav, sizeof(HST_t) * HHist_siz);
PHist_new = alloc_r(PHist_new, sizeof(HST_t) * HHist_siz);
@ -2859,6 +2861,8 @@ static void procs_refresh (void) {
for (;;) {
if (n_used == n_alloc) {
/* we're subject to integer overflow if total linux tasks ever approach |
400+ million (but, do you think memory might be the bigger problem?) | */
n_alloc = 10 + ((n_alloc * 5) / 4); // grow by over 25%
private_ppt = alloc_r(private_ppt, sizeof(proc_t*) * n_alloc);
// ensure NULL pointers for the additional memory just acquired