| I have some questions about the "page" and "swap" entries in /proc/stat.
| Here is the relevant code from 2.4.12
| 292 len += sprintf(page + len,
| 293 "page %u %u\n"
| 294 "swap %u %u\n"
| 295 "intr %u",
| 296 kstat.pgpgin >> 1,
| 297 kstat.pgpgout >> 1,
| 298 kstat.pswpin,
| 299 kstat.pswpout,
| 300 sum
| 301 );
Of course the basic answer is something like
grep -r "kstat.p" * | more
Using the latter one:
| 1. Why are kstat.pgpgin and kstat.pgpgout shifted right / halved?
I had wondered that also, so you made me look.
pgpgin and pgpgout are maintained (counted) in units of 512-byte
blocks but displayed in /proc/stat in 1 KB (KiB :) blocks by
shifting right by 1.
| 2. Are the "page" and "swap" numbers mutually exclusive? That is, if a
| page is brought in from swap and counted in kstat.pswpin, is it also
| counted in kstat.pgpgin? I found the places in the code where the counts
| are incremented, but I couldn't tell if the swapin routine calls the
| block driver or not.
No, "page" and "swap" counts are not mutually exclusive.
Both paths call submit_bh().
/* reads or writes a swap page */
which calls submit_bh();
For all low-level block I/O:
/* low-level access to block devices */
count := is number of 512-byte blocks;
kstat.pgpgout += count;
kstat.pgpgin += count;
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to email@example.com More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/