[RFC PATCH 1/6] mm: Add new system calls mshare, mshare_unlink

[Khalid Aziz]

Add two new system calls to support PTE sharing across processes through
explicit declarations of shared address space.There is almost no
implementation in this patch and it only wires up the system calls for
x86_64 only. mshare() returns a file descriptor which does not support
any operations yet.

Signed-off-by: Khalid Aziz <khalid.aziz@xxxxxxxxxx>
Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
---
 arch/x86/entry/syscalls/syscall_64.tbl |  2 +
 include/uapi/asm-generic/unistd.h      |  7 ++-
 mm/Makefile                            |  2 +-
 mm/mshare.c                            | 60 ++++++++++++++++++++++++++
 4 files changed, 69 insertions(+), 2 deletions(-)
 create mode 100644 mm/mshare.c

diff --git a/arch/x86/entry/syscalls/syscall_64.tbl b/arch/x86/entry/syscalls/syscall_64.tbl
index fe8f8dd157b4..bb403deca1ff 100644
--- a/arch/x86/entry/syscalls/syscall_64.tbl
+++ b/arch/x86/entry/syscalls/syscall_64.tbl
@@ -371,6 +371,8 @@
 447	common	memfd_secret		sys_memfd_secret
 448	common	process_mrelease	sys_process_mrelease
 449	common	futex_waitv		sys_futex_waitv
+450	common	mshare			sys_mshare
+451	common	mshare_unlink		sys_mshare_unlink
 
 #
 # Due to a historical design error, certain syscalls are numbered differently
diff --git a/include/uapi/asm-generic/unistd.h b/include/uapi/asm-generic/unistd.h
index 4557a8b6086f..27349ad579ff 100644
--- a/include/uapi/asm-generic/unistd.h
+++ b/include/uapi/asm-generic/unistd.h
@@ -883,8 +883,13 @@ __SYSCALL(__NR_process_mrelease, sys_process_mrelease)
 #define __NR_futex_waitv 449
 __SYSCALL(__NR_futex_waitv, sys_futex_waitv)
 
+#define __NR_mshare 450
+__SYSCALL(__NR_mshare, sys_mshare)
+#define __NR_mshare_unlink 451
+__SYSCALL(__NR_mshare_unlink, sys_mshare_unlink)
+
 #undef __NR_syscalls
-#define __NR_syscalls 450
+#define __NR_syscalls 452
 
 /*
  * 32 bit systems traditionally used different
diff --git a/mm/Makefile b/mm/Makefile
index d6c0042e3aa0..fca44c0d5e74 100644
--- a/mm/Makefile
+++ b/mm/Makefile
@@ -35,7 +35,7 @@ CFLAGS_init-mm.o += $(call cc-disable-warning, override-init)
 CFLAGS_init-mm.o += $(call cc-disable-warning, initializer-overrides)
 
 mmu-y			:= nommu.o
-mmu-$(CONFIG_MMU)	:= highmem.o memory.o mincore.o \
+mmu-$(CONFIG_MMU)	:= highmem.o memory.o mincore.o mshare.o \
 			   mlock.o mmap.o mmu_gather.o mprotect.o mremap.o \
 			   msync.o page_vma_mapped.o pagewalk.o \
 			   pgtable-generic.o rmap.o vmalloc.o
diff --git a/mm/mshare.c b/mm/mshare.c
new file mode 100644
index 000000000000..c723f8369f06
--- /dev/null
+++ b/mm/mshare.c
@@ -0,0 +1,60 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * mm/mshare.c
+ *
+ * Page table sharing code
+ *
+ *
+ * Copyright (C) 2021 Oracle Corp. All rights reserved.
+ * Authors:	Matthew Wilcox
+ *		Khalid Aziz
+ */
+
+#include <linux/anon_inodes.h>
+#include <linux/fs.h>
+#include <linux/syscalls.h>
+
+static const struct file_operations mshare_fops = {
+};
+
+/*
+ * mshare syscall. Returns a file descriptor
+ */
+SYSCALL_DEFINE5(mshare, const char *, name, unsigned long, addr,
+		unsigned long, len, int, oflag, mode_t, mode)
+{
+	int fd;
+
+	/*
+	 * Address range being shared must be aligned to pgdir
+	 * boundary and its size must be a multiple of pgdir size
+	 */
+	if ((addr | len) & (PGDIR_SIZE - 1))
+		return -EINVAL;
+
+	/*
+	 * Allocate a file descriptor to return
+	 *
+	 * TODO: This code ignores the object name completely. Add
+	 * support for that
+	 */
+	fd = anon_inode_getfd("mshare", &mshare_fops, NULL, O_RDWR);
+
+	return fd;
+}
+
+/*
+ * mshare_unlink syscall. Close and remove the named mshare'd object
+ */
+SYSCALL_DEFINE1(mshare_unlink, const char *, name)
+{
+	int fd;
+
+	/*
+	 * Delete the named object
+	 *
+	 * TODO: Mark mshare'd range for deletion
+	 *
+	 */
+	return 0;
+}
-- 
2.32.0