| From: | David Steele <david(at)pgmasters(dot)net> |
|---|---|
| To: | Andres Freund <andres(at)anarazel(dot)de>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: ubsan |
| Date: | 2022-03-25 15:55:45 |
| Message-ID: | 0befb6fb-d125-95f2-9815-7dbda8086f25@pgmasters.net |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 3/23/22 16:55, Andres Freund wrote:
>
> It's particularly impressive that the cost of running with ASAN is *so* much
> lower than valgrind. On my workstation a check-world with
> -fsanitize=alignment,undefined,address takes 3min17s, vs 1min10s or so without
> -fsanitize. Not something to always use, but certainly better than valgrind.
It also catches things that valgrind does not so that's a bonus.
One thing to note, though. I have noticed that when enabling
-fsanitize=undefined and/or -fsanitize=address in combination with
-fprofile-arcs -ftest-coverage there is a loss in reported coverage, at
least on gcc 9.3. This may not be very obvious unless coverage is
normally at 100%.
Regards,
-David
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Guillaume Lelarge | 2022-03-25 15:59:00 | Re: Probable memory leak with ECPG and AIX |
| Previous Message | Tom Lane | 2022-03-25 15:50:48 | Re: Corruption during WAL replay |