Thread (5 messages) 5 messages, 2 authors, 2025-07-29

Re: [PATCH v2 1/2] man/man2/mremap.2: describe multiple mapping move

From: Lorenzo Stoakes <hidden>
Date: 2025-07-29 12:10:36
Also in: linux-man, linux-mm, lkml

On Mon, Jul 28, 2025 at 10:34:07PM +0200, Jann Horn wrote:
On Mon, Jul 28, 2025 at 4:05 PM Lorenzo Stoakes
[off-list ref] wrote:
quoted
Document the new behaviour introduced in Linux 6.17 whereby it is now
possible to move multiple mappings in a single operation, as long as the
operation is purely a move, that is old_size is equal to new_size and
MREMAP_FIXED is specified.

To make things clearer, also describe this 'pure move' operation, before
expanding upon it to describe the newly introduced behaviour.

This change also explains the limitations of of this method and the
possibility of partial failure.

Finally, we pluralise language where it makes sense to so the documentation
does not contradict either this new capability nor the pre-existing edge
case.

Signed-off-by: Lorenzo Stoakes <redacted>
---
 man/man2/mremap.2 | 78 ++++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 67 insertions(+), 11 deletions(-)
diff --git a/man/man2/mremap.2 b/man/man2/mremap.2
index 2168ca728..cb3412591 100644
--- a/man/man2/mremap.2
+++ b/man/man2/mremap.2
@@ -25,18 +25,41 @@ moving it at the same time (controlled by the
 argument and
 the available virtual address space).
 .P
+Mappings can simply be moved by specifying equal
(Bikeshedding: This "simply" sounds weird to me. If you're trying to
define a "simple move" with this, the rest of this block is not very
specific about what exactly that is supposed to be. In my opinion,
"pure" would also be a nicer word than "simple" if you're looking for
an expression that means "a move that doesn't do other things".)
Will rephrase, the intent is that I'm saying we can 'simply' perform the
'only move'.
quoted
+.I old_size
+and
+.I new_size
+and specifying
+.IR new_address ,
+see the description of
+.B MREMAP_FIXED
+below.
+Since Linux 6.17,
+while
+.I old_address
+must reside within a mapping,
+.I old_size
+may span multiple mappings
+which do not have to be
+adjacent to one another.
+.P
+If the operation is not a simple move
+then
+.I old_size
+must span only a single mapping.
I'm reading between the lines that "simple move" is supposed to mean
"the size is not changing and MREMAP_DONTUNMAP is not set", which then
implies that in order to actually make anything happen, MREMAP_FIXED
must be specified?
No, MREMAP_DONTUNMAP can be set, but MREMAP_FIXED must always be set for this to
happen.

Let me rephrase for clarity.
quoted
+.P
 .I old_address
-is the old address of the virtual memory block that you
-want to expand (or shrink).
+is the old address of the first virtual memory block that you
+want to expand, shrink, and/or move.
 Note that
 .I old_address
 has to be page aligned.
 .I old_size
-is the old size of the
-virtual memory block.
+is the size of the range containing
+virtual memory blocks to be manipulated.
 .I new_size
 is the requested size of the
-virtual memory block after the resize.
+virtual memory blocks after the resize.
 An optional fifth argument,
 .IR new_address ,
 may be provided; see the description of
@@ -105,13 +128,43 @@ If
 is specified, then
 .B MREMAP_MAYMOVE
 must also be specified.
+.IP
+Since Linux 6.17,
+if
+.I old_size
+is equal to
+.I new_size
+and
+.B MREMAP_FIXED
+is specified, then
+.I old_size
+may span beyond the mapping in which
+.I old_address
+resides.
+In this case,
+gaps between mappings in the original range
+are maintained in the new range.
+The whole operation is performed atomically
+unless an error arises,
+in which case the operation may be partially
+completed,
+that is,
+some mappings may be moved and others not.
This is much clearer to me.
Thanks
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help