On Fri, Dec 27, 2024 at 11:33:10AM -0800, Joanne Koong wrote:
> Add support for reads/writes from buffers backed by hugepages.
> This can be enabled through the '-h' flag. This flag should only be used
> on systems where THP capabilities are enabled.
>
> This is motivated by a recent bug that was due to faulty handling of
> userspace buffers backed by hugepages. This patch is a mitigation
> against problems like this in the future.
>
> Signed-off-by: Joanne Koong <joannelkoong@xxxxxxxxx>
> ---
> ltp/fsx.c | 108 ++++++++++++++++++++++++++++++++++++++++++++++++------
> 1 file changed, 97 insertions(+), 11 deletions(-)
>
[snip]
> +static void *
> +init_hugepages_buf(unsigned len, int hugepage_size, int alignment)
> +{
> + void *buf;
> + long buf_size = roundup(len, hugepage_size) + alignment;
> +
> + if (posix_memalign(&buf, hugepage_size, buf_size)) {
> + prterr("posix_memalign for buf");
> + return NULL;
> + }
> + memset(buf, '\0', buf_size);
> + if (madvise(buf, buf_size, MADV_COLLAPSE)) {
Hi Joanne,
Sorry I have to drop this patchset from the "upcoming" release v2025.01.12. Due to
it cause a regression build error on older system, e.g. RHEL-9:
[CC] fsx
fsx.c: In function 'init_hugepages_buf':
fsx.c:2935:36: error: 'MADV_COLLAPSE' undeclared (first use in this function); did you mean 'MADV_COLD'?
2935 | if (madvise(buf, buf_size, MADV_COLLAPSE)) {
| ^~~~~~~~~~~~~
| MADV_COLD
fsx.c:2935:36: note: each undeclared identifier is reported only once for each function it appears in
gmake[4]: *** [Makefile:51: fsx] Error 1
gmake[4]: *** Waiting for unfinished jobs....
gmake[3]: *** [include/buildrules:30: ltp] Error 2
It might cause xfstests totally can't be used on downstream systems, so it can't
catch up the release of this weekend. Sorry about that, let's try to have it
in next release :)
Thanks,
Zorro
> + prterr("madvise collapse for buf");
> + free(buf);
> + return NULL;
> + }
> +
> + return buf;
> +}
> +
> +static void
> +init_buffers(void)
> +{
> + int i;
> +
> + original_buf = (char *) malloc(maxfilelen);
> + for (i = 0; i < maxfilelen; i++)
> + original_buf[i] = random() % 256;
> + if (hugepages) {
> + long hugepage_size = get_hugepage_size();
> + if (hugepage_size == -1) {
> + prterr("get_hugepage_size()");
> + exit(100);
> + }
> + good_buf = init_hugepages_buf(maxfilelen, hugepage_size, writebdy);
> + if (!good_buf) {
> + prterr("init_hugepages_buf failed for good_buf");
> + exit(101);
> + }
> +
> + temp_buf = init_hugepages_buf(maxoplen, hugepage_size, readbdy);
> + if (!temp_buf) {
> + prterr("init_hugepages_buf failed for temp_buf");
> + exit(101);
> + }
> + } else {
> + unsigned long good_buf_len = maxfilelen + writebdy;
> + unsigned long temp_buf_len = maxoplen + readbdy;
> +
> + good_buf = (char *) malloc(good_buf_len);
> + memset(good_buf, '\0', good_buf_len);
> + temp_buf = (char *) malloc(temp_buf_len);
> + memset(temp_buf, '\0', temp_buf_len);
> + }
> + good_buf = round_ptr_up(good_buf, writebdy, 0);
> + temp_buf = round_ptr_up(temp_buf, readbdy, 0);
> +}
> +
> static struct option longopts[] = {
> {"replay-ops", required_argument, 0, 256},
> {"record-ops", optional_argument, 0, 255},
> @@ -2883,7 +2974,7 @@ main(int argc, char **argv)
> setvbuf(stdout, (char *)0, _IOLBF, 0); /* line buffered stdout */
>
> while ((ch = getopt_long(argc, argv,
> - "0b:c:de:fg:i:j:kl:m:no:p:qr:s:t:uw:xyABD:EFJKHzCILN:OP:RS:UWXZ",
> + "0b:c:de:fg:hi:j:kl:m:no:p:qr:s:t:uw:xyABD:EFJKHzCILN:OP:RS:UWXZ",
> longopts, NULL)) != EOF)
> switch (ch) {
> case 'b':
> @@ -2916,6 +3007,9 @@ main(int argc, char **argv)
> case 'g':
> filldata = *optarg;
> break;
> + case 'h':
> + hugepages = 1;
> + break;
> case 'i':
> integrity = 1;
> logdev = strdup(optarg);
> @@ -3229,15 +3323,7 @@ main(int argc, char **argv)
> exit(95);
> }
> }
> - original_buf = (char *) malloc(maxfilelen);
> - for (i = 0; i < maxfilelen; i++)
> - original_buf[i] = random() % 256;
> - good_buf = (char *) malloc(maxfilelen + writebdy);
> - good_buf = round_ptr_up(good_buf, writebdy, 0);
> - memset(good_buf, '\0', maxfilelen);
> - temp_buf = (char *) malloc(maxoplen + readbdy);
> - temp_buf = round_ptr_up(temp_buf, readbdy, 0);
> - memset(temp_buf, '\0', maxoplen);
> + init_buffers();
> if (lite) { /* zero entire existing file */
> ssize_t written;
>
> --
> 2.47.1
>
>
