Thread (9 messages) 9 messages, 4 authors, 2020-04-15

Re: [PATCH v7] mremap.2: Add information for MREMAP_DONTUNMAP.

From: Brian Geffon <hidden>
Date: 2020-04-15 15:25:00
Also in: linux-man, linux-mm, lkml

Hi Michael,
I'll make those changes and start a new thread.

Thanks for the feedback,
Brian

On Tue, Apr 14, 2020 at 11:41 PM Michael Kerrisk (man-pages)
[off-list ref] wrote:
Hello Brian,

I see that MREMAP_DONTUNMAP has been merged. Thanks for the
patch below.

In addition to Vlastimil's comments, could you please take a look
at my comments below.

On 2/21/20 6:42 PM, Brian Geffon wrote:
quoted
Signed-off-by: Brian Geffon <redacted>
---
 man2/mremap.2 | 30 +++++++++++++++++++++++++++++-
 1 file changed, 29 insertions(+), 1 deletion(-)
diff --git a/man2/mremap.2 b/man2/mremap.2
index d73fb64fa..54ec67b20 100644
--- a/man2/mremap.2
+++ b/man2/mremap.2
@@ -26,7 +26,8 @@
 .\" 1996-04-12 Tom Bjorkholm <tomb@mydata.se>
 .\"            Update for Linux 1.3.87 and later
 .\" 2005-10-11 mtk: Added NOTES for MREMAP_FIXED; revised EINVAL text.
-.\"
+.\" 2020-02-05 Brian Geffon <bgeffon@google.com>
+.\"            Update for MREMAP_DONTUNMAP.
No need to add this piece. This info is maintained
via the Git log these days.
quoted
 .TH MREMAP 2 2019-03-06 "Linux" "Linux Programmer's Manual"
 .SH NAME
 mremap \- remap a virtual memory address
@@ -129,6 +130,13 @@ If
 is specified, then
 .B MREMAP_MAYMOVE
 must also be specified.
+.TP
+.BR MREMAP_DONTUNMAP " (since Linux 5.7)"
+This flag which must be used in conjuction with
s/conjuction/conjunction/
quoted
+.B MREMAP_MAYMOVE
+remaps a mapping to a new address and it does not unmap the mapping at \fIold_address\fP. This flag can only be used with private anonymous mappings. Any access to the range specified at \fIold_address\fP after completion will result in a page fault. If a
Please wrap source lines to no more than about 75 columns.
Also, always start new sentences on new lines ("Semantic newlines").

As a general rule, I prefer formatting to be done like this:

.BR old_address .

rather than:

\fIold_address\fP.

(Yes, I know there's plenty of existing text that goes the other
way, but I try to avoid the \fX...\fP style for new text.

Re the "Any access to the range ... will result in a page fault", I think
it would be helpful to be more explicit. I presume that if we
access the range at old_address the mapping is repopulated with
zero-filled pages, right? It would be good to note that explicitly,
quoted
+.BR userfaultfd (2)
+was registered on the mapping specified by \fIold_address\fP it will continue to watch that mapping for faults.
(See comments above re wrapping and formatting.)

Perhaps it would be nice to have a short paragraph on use cases?
quoted
 .PP
 If the memory segment specified by
 .I old_address
@@ -176,6 +184,8 @@ a value other than
 .B MREMAP_MAYMOVE
 or
 .B MREMAP_FIXED
+or
+.B MREMAP_DONTUNMAP
 was specified in
 .IR flags ;
 .IP *
@@ -197,9 +207,17 @@ and
 .IR old_size ;
 .IP *
 .B MREMAP_FIXED
+or
+.B MREMAP_DONTUNMAP
 was specified without also specifying
 .BR MREMAP_MAYMOVE ;
 .IP *
+.B MREMAP_DONTUNMAP
+was specified with an \fIold_address\fP that was not private anonymous;
+.IP *
+.B MREMAP_DONTUNMAP
+was specified and \fIold_size\fP was not equal to \fInew_size\fP;
+.IP *
 \fIold_size\fP was zero and \fIold_address\fP does not refer to a
 shareable mapping (but see BUGS);
 .IP *
@@ -209,10 +227,20 @@ flag was not specified.
 .RE
 .TP
 .B ENOMEM
+Not enough memory was available to complete the operation.
+Possible causes are:
+.RS
+.IP * 3
 The memory area cannot be expanded at the current virtual address, and the
 .B MREMAP_MAYMOVE
 flag is not set in \fIflags\fP.
 Or, there is not enough (virtual) memory available.
+.IP *
+.B MREMAP_DONTUNMAP
+was used without
+.B MREMAP_FIXED
+causing a new mapping to be created that would exceed the virtual memory available or it would exceed the maximum number of allowed mappings.
(See comments above re wrapping.)
quoted
+.RE
 .SH CONFORMING TO
 This call is Linux-specific, and should not be used in programs
 intended to be portable.

Thanks,

Michael


--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
Linux/UNIX System Programming Training: http://man7.org/training/
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help