On Fri, Jun 07, 2019 at 02:10:15PM +0800, Pingfan Liu wrote:
> On Fri, Jun 7, 2019 at 5:17 AM John Hubbard <jhubb...@nvidia.com> wrote:
> >
> > On 6/5/19 7:19 PM, Pingfan Liu wrote:
> > > On Thu, Jun 6, 2019 at 5:49 AM Andrew Morton <a...@linux-foundation.org>
> > > wrote:
> > ...
> > >>> --- a/mm/gup.c
> > >>> +++ b/mm/gup.c
> > >>> @@ -2196,6 +2196,26 @@ static int __gup_longterm_unlocked(unsigned long
> > >>> start, int nr_pages,
> > >>> return ret;
> > >>> }
> > >>>
> > >>> +#ifdef CONFIG_CMA
> > >>> +static inline int reject_cma_pages(int nr_pinned, struct page **pages)
> > >>> +{
> > >>> + int i;
> > >>> +
> > >>> + for (i = 0; i < nr_pinned; i++)
> > >>> + if (is_migrate_cma_page(pages[i])) {
> > >>> + put_user_pages(pages + i, nr_pinned - i);
> > >>> + return i;
> > >>> + }
> > >>> +
> > >>> + return nr_pinned;
> > >>> +}
> > >>
> > >> There's no point in inlining this.
> > > OK, will drop it in V4.
> > >
> > >>
> > >> The code seems inefficient. If it encounters a single CMA page it can
> > >> end up discarding a possibly significant number of non-CMA pages. I
> > > The trick is the page is not be discarded, in fact, they are still be
> > > referrenced by pte. We just leave the slow path to pick up the non-CMA
> > > pages again.
> > >
> > >> guess that doesn't matter much, as get_user_pages(FOLL_LONGTERM) is
> > >> rare. But could we avoid this (and the second pass across pages[]) by
> > >> checking for a CMA page within gup_pte_range()?
> > > It will spread the same logic to hugetlb pte and normal pte. And no
> > > improvement in performance due to slow path. So I think maybe it is
> > > not worth.
> > >
> > >>
> >
> > I think the concern is: for the successful gup_fast case with no CMA
> > pages, this patch is adding another complete loop through all the
> > pages. In the fast case.
> >
> > If the check were instead done as part of the gup_pte_range(), then
> > it would be a little more efficient for that case.
> >
> > As for whether it's worth it, *probably* this is too small an effect to
> > measure.
> > But in order to attempt a measurement: running fio
> > (https://github.com/axboe/fio)
> > with O_DIRECT on an NVMe drive, might shed some light. Here's an fio.conf
> > file
> > that Jan Kara and Tom Talpey helped me come up with, for related testing:
> >
> > [reader]
> > direct=1
> > ioengine=libaio
> > blocksize=4096
> > size=1g
> > numjobs=1
> > rw=read
> > iodepth=64
> >
Unable to get a NVME device to have a test. And when testing fio on the
tranditional disk, I got the error "fio: engine libaio not loadable
fio: failed to load engine
fio: file:ioengines.c:89, func=dlopen, error=libaio: cannot open shared object
file: No such file or directory"
But I found a test case which can be slightly adjusted to met the aim.
It is tools/testing/selftests/vm/gup_benchmark.c
Test enviroment:
MemTotal: 264079324 kB
MemFree: 262306788 kB
CmaTotal: 0 kB
CmaFree: 0 kB
on AMD EPYC 7601
Test command:
gup_benchmark -r 100 -n 64
gup_benchmark -r 100 -n 64 -l
where -r stands for repeat times, -n is nr_pages param for
get_user_pages_fast(), -l is a new option to test FOLL_LONGTERM in fast
path, see a patch at the tail.
Test result:
w/o 477.800000
w/o-l 481.070000
a 481.800000
a-l 640.410000
b 466.240000 (question a: b outperforms w/o ?)
b-l 529.740000
Where w/o is baseline without any patch using v5.2-rc2, a is this series, b
does the check in gup_pte_range(). '-l' means FOLL_LONGTERM.
I am suprised that b-l has about 17% improvement than a. (640.41 -529.74)/640.41
As for "question a: b outperforms w/o ?", I can not figure out why, maybe it
can be
considered as variance.
Based on the above result, I think it is better to do the check inside
gup_pte_range().
Any comment?
Thanks,
> Yeah, agreed. Data is more persuasive. Thanks for your suggestion. I
> will try to bring out the result.
>
> Thanks,
> Pingfan
>
---
Patch to do check inside gup_pte_range()
diff --git a/mm/gup.c b/mm/gup.c
index 2ce3091..ba213a0 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1757,6 +1757,10 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr,
unsigned long end,
VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
page = pte_page(pte);
+ if (unlikely(flags & FOLL_LONGTERM) &&
+ is_migrate_cma_page(page))
+ goto pte_unmap;
+
head = try_get_compound_head(page, 1);
if (!head)
goto pte_unmap;
@@ -1900,6 +1904,12 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp,
unsigned long addr,
refs++;
} while (addr += PAGE_SIZE, addr != end);
+ if (unlikely(flags & FOLL_LONGTERM) &&
+ is_migrate_cma_page(page)) {
+ *nr -= refs;
+ return 0;
+ }
+
head = try_get_compound_head(pmd_page(orig), refs);
if (!head) {
*nr -= refs;
@@ -1941,6 +1951,12 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp,
unsigned long addr,
refs++;
} while (addr += PAGE_SIZE, addr != end);
+ if (unlikely(flags & FOLL_LONGTERM) &&
+ is_migrate_cma_page(page)) {
+ *nr -= refs;
+ return 0;
+ }
+
head = try_get_compound_head(pud_page(orig), refs);
if (!head) {
*nr -= refs;
@@ -1978,6 +1994,12 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp,
unsigned long addr,
refs++;
} while (addr += PAGE_SIZE, addr != end);
+ if (unlikely(flags & FOLL_LONGTERM) &&
+ is_migrate_cma_page(page)) {
+ *nr -= refs;
+ return 0;
+ }
+
head = try_get_compound_head(pgd_page(orig), refs);
if (!head) {
*nr -= refs;
---
Patch for testing
diff --git a/mm/gup_benchmark.c b/mm/gup_benchmark.c
index 7dd602d..61dec5f 100644
--- a/mm/gup_benchmark.c
+++ b/mm/gup_benchmark.c
@@ -6,8 +6,9 @@
#include <linux/debugfs.h>
#define GUP_FAST_BENCHMARK _IOWR('g', 1, struct gup_benchmark)
-#define GUP_LONGTERM_BENCHMARK _IOWR('g', 2, struct gup_benchmark)
-#define GUP_BENCHMARK _IOWR('g', 3, struct gup_benchmark)
+#define GUP_FAST_LONGTERM_BENCHMARK _IOWR('g', 2, struct gup_benchmark)
+#define GUP_LONGTERM_BENCHMARK _IOWR('g', 3, struct gup_benchmark)
+#define GUP_BENCHMARK _IOWR('g', 4, struct gup_benchmark)
struct gup_benchmark {
__u64 get_delta_usec;
@@ -53,6 +54,11 @@ static int __gup_benchmark_ioctl(unsigned int cmd,
nr = get_user_pages_fast(addr, nr, gup->flags & 1,
pages + i);
break;
+ case GUP_FAST_LONGTERM_BENCHMARK:
+ nr = get_user_pages_fast(addr, nr,
+ (gup->flags & 1) |
FOLL_LONGTERM,
+ pages + i);
+ break;
case GUP_LONGTERM_BENCHMARK:
nr = get_user_pages(addr, nr,
(gup->flags & 1) | FOLL_LONGTERM,
@@ -96,6 +102,7 @@ static long gup_benchmark_ioctl(struct file *filep, unsigned
int cmd,
switch (cmd) {
case GUP_FAST_BENCHMARK:
+ case GUP_FAST_LONGTERM_BENCHMARK:
case GUP_LONGTERM_BENCHMARK:
case GUP_BENCHMARK:
break;
diff --git a/tools/testing/selftests/vm/gup_benchmark.c
b/tools/testing/selftests/vm/gup_benchmark.c
index c0534e2..ade8acb 100644
--- a/tools/testing/selftests/vm/gup_benchmark.c
+++ b/tools/testing/selftests/vm/gup_benchmark.c
@@ -15,8 +15,9 @@
#define PAGE_SIZE sysconf(_SC_PAGESIZE)
#define GUP_FAST_BENCHMARK _IOWR('g', 1, struct gup_benchmark)
-#define GUP_LONGTERM_BENCHMARK _IOWR('g', 2, struct gup_benchmark)
-#define GUP_BENCHMARK _IOWR('g', 3, struct gup_benchmark)
+#define GUP_FAST_LONGTERM_BENCHMARK _IOWR('g', 2, struct gup_benchmark)
+#define GUP_LONGTERM_BENCHMARK _IOWR('g', 3, struct gup_benchmark)
+#define GUP_BENCHMARK _IOWR('g', 4, struct gup_benchmark)
struct gup_benchmark {
__u64 get_delta_usec;
@@ -37,7 +38,7 @@ int main(int argc, char **argv)
char *file = "/dev/zero";
char *p;
- while ((opt = getopt(argc, argv, "m:r:n:f:tTLUSH")) != -1) {
+ while ((opt = getopt(argc, argv, "m:r:n:f:tTlLUSH")) != -1) {
switch (opt) {
case 'm':
size = atoi(optarg) * MB;
@@ -54,6 +55,9 @@ int main(int argc, char **argv)
case 'T':
thp = 0;
break;
+ case 'l':
+ cmd = GUP_FAST_LONGTERM_BENCHMARK;
+ break;
case 'L':
cmd = GUP_LONGTERM_BENCHMARK;
break;
--
2.7.5
