abd_iter_page: don't use compound heads on Linux <4.5

Before 4.5 (specifically, torvalds/linux@ddc58f2), head and tail pages
in a compound page were refcounted separately. This means that using the
head page without taking a reference to it could see it cleaned up later
before we're finished with it. Specifically, bio_add_page() would take a
reference, and drop its reference after the bio completion callback
returns.

If the zio is executed immediately from the completion callback, this is
usually ok, as any data is referenced through the tail page referenced
by the ABD, and so becomes "live" that way. If there's a delay in zio
execution (high load, error injection), then the head page can be freed,
along with any dirty flags or other indicators that the underlying
memory is used. Later, when the zio completes and that memory is
accessed, its either unmapped and an unhandled fault takes down the
entire system, or it is mapped and we end up messing around in someone
else's memory. Both of these are very bad.

The solution on these older kernels is to take a reference to the head
page when we use it, and release it when we're done. There's not really
a sensible way under our current structure to do this; the "best" would
be to keep a list of head page references in the ABD, and release them
when the ABD is freed.

Since this additional overhead is totally unnecessary on 4.5+, where
head and tail pages share refcounts, I've opted to simply not use the
compound head in ABD page iteration there. This is theoretically less
efficient (though cleaning up head page references would add overhead),
but its safe, and we still get the other benefits of not mapping pages
before adding them to a bio and not mis-splitting pages.

There doesn't appear to be an obvious symbol name or config option we
can match on to discover this behaviour in configure (and the mm/page
APIs have changed a lot since then anyway), so I've gone with a simple
version check.

Reviewed-by: Alexander Motin <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Signed-off-by: Rob Norris <[email protected]>
Sponsored-by: Klara, Inc.
Sponsored-by: Wasabi Technology, Inc.
Closes #15533
Closes #15588

1 files changed, 14 insertions, 0 deletions

diff --git a/module/os/linux/zfs/abd_os.c b/module/os/linux/zfs/abd_os.c
index 3fe01c0b7..d3255dcbc 100644
--- a/module/os/linux/zfs/abd_os.c
+++ b/module/os/linux/zfs/abd_os.c
@@ -62,6 +62,7 @@
 #include <linux/kmap_compat.h>
 #include <linux/mm_compat.h>
 #include <linux/scatterlist.h>
+#include <linux/version.h>
 #endif
 
 #ifdef _KERNEL
@@ -1061,6 +1062,7 @@ abd_iter_page(struct abd_iter *aiter)
 	}
 	ASSERT(page);
 
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 5, 0)
 	if (PageTail(page)) {
 		/*
 		 * This page is part of a "compound page", which is a group of
@@ -1082,11 +1084,23 @@ abd_iter_page(struct abd_iter *aiter)
 		 * To do this, we need to adjust the offset to be counted from
 		 * the head page. struct page for compound pages are stored
 		 * contiguously, so we can just adjust by a simple offset.
+		 *
+		 * Before kernel 4.5, compound page heads were refcounted
+		 * separately, such that moving back to the head page would
+		 * require us to take a reference to it and releasing it once
+		 * we're completely finished with it. In practice, that means
+		 * when our caller is done with the ABD, which we have no
+		 * insight into from here. Rather than contort this API to
+		 * track head page references on such ancient kernels, we just
+		 * compile this block out and use the tail pages directly. This
+		 * is slightly less efficient, but makes everything far
+		 * simpler.
 		 */
 		struct page *head = compound_head(page);
 		doff += ((page - head) * PAGESIZE);
 		page = head;
 	}
+#endif
 
 	/* final page and position within it */
 	aiter->iter_page = page;