[PATCH net-next v3 0/2] net: rds: convert rds to getsockopt_iter

[PATCH net-next v3 0/2] net: rds: convert rds to getsockopt_iter

[Breno Leitao]

This series continues the conversion of the remaining proto_ops getsockopt
callbacks to the new getsockopt_iter callback introduced in commit
67fab22a7adc ("net: add getsockopt_iter callback to proto_ops"), this time
for RDS.

RDS is a little more involved than the protocols converted so far, because
the RDS_INFO_* options snapshot kernel state directly into the destination
buffer: the info producers memcpy into the pages under a spinlock via
kmap_atomic() and so must not fault.

The conversion preserves that model — it obtains the same page array and
starting offset from opt->iter_out with iov_iter_extract_pages(),
preallocating the array so the iterator fills it in place, and leaves
the rds_info_iterator / rds_info_copy machinery and all producer
callbacks unchanged; kernel (ITER_KVEC) buffers remain unsupported on
the RDS_INFO path, as before.

I've vibe-coded a kselftest exercising both the simple options and the
RDS_INFO_* snapshot path, feel free to drop it in case this is not
useful.

Signed-off-by: Breno Leitao <leitao@debian.org>
---
Changes in v3:
- Resolved the conflict with net-next
- Added additional tags (assisted-by and reviewes)
- Link to v2: https://lore.kernel.org/r/20260605-getsock_more-v2-0-80f38cdb8706@debian.org

Changes in v2:
- rds: reject non-user-backed buffers with !user_backed_iter() instead of
  iov_iter_is_kvec() on the RDS_INFO path (Allison Henderson)
- rds: gate the page unpin on iov_iter_extract_will_pin() and comment the
  implicit pin (Allison Henderson)
- selftest: size the snapshot mmap from the probed length instead of a fixed
  two pages (Allison Henderson)
- Add a new patch to fix a concern raised by Sashiko.
- Link to v1:
https://lore.kernel.org/r/20260603-getsock_more-v1-0-43b8d40c8849@debian.org

---
Breno Leitao (2):
      selftests: net: rds: add getsockopt() conversion test
      rds: convert to getsockopt_iter

 net/rds/af_rds.c                             |  36 ++---
 net/rds/info.c                               |  76 +++++-----
 net/rds/info.h                               |   3 +-
 tools/testing/selftests/net/rds/.gitignore   |   1 +
 tools/testing/selftests/net/rds/Makefile     |   4 +
 tools/testing/selftests/net/rds/getsockopt.c | 208 +++++++++++++++++++++++++++
 6 files changed, 278 insertions(+), 50 deletions(-)
---
base-commit: 903db046d5579bef0ea699eae4b279dd6455fc9f
change-id: 20260603-getsock_more-46be8d1c56fd

Best regards,
-- 
Breno Leitao <leitao@debian.org>

[PATCH net-next v3 1/2] selftests: net: rds: add getsockopt() conversion test

[Breno Leitao]

Add a kselftest that exercises the RDS getsockopt() paths converted to
the getsockopt_iter() / sockopt_t callback:

- RDS_RECVERR and SO_RDS_TRANSPORT, which return their int value through
  copy_to_iter() and report the written length in opt->optlen.

- RDS_INFO_*, which obtains the userspace buffer pages with
  iov_iter_extract_pages() (including a non-zero starting page offset)
  and lets the info producers copy the snapshot in under a spinlock.

Assisted-by: Claude:claude-opus-4-8
Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Allison Henderson <achender@kernel.org>
---
 tools/testing/selftests/net/rds/.gitignore   |   1 +
 tools/testing/selftests/net/rds/Makefile     |   4 +
 tools/testing/selftests/net/rds/getsockopt.c | 208 +++++++++++++++++++++++++++
 3 files changed, 213 insertions(+)

diff --git a/tools/testing/selftests/net/rds/.gitignore b/tools/testing/selftests/net/rds/.gitignore
index 1c6f04e2aa11..7ca4b1440f51 100644
--- a/tools/testing/selftests/net/rds/.gitignore
+++ b/tools/testing/selftests/net/rds/.gitignore
@@ -1 +1,2 @@
 include.sh
+getsockopt
diff --git a/tools/testing/selftests/net/rds/Makefile b/tools/testing/selftests/net/rds/Makefile
index ec10ae24e4cf..ab9e92399a6d 100644
--- a/tools/testing/selftests/net/rds/Makefile
+++ b/tools/testing/selftests/net/rds/Makefile
@@ -5,6 +5,8 @@ all:
 
 TEST_PROGS := rds_run.sh
 
+TEST_GEN_PROGS := getsockopt
+
 TEST_FILES := \
 	include.sh \
 	settings \
@@ -16,4 +18,6 @@ EXTRA_CLEAN := \
 	/tmp/rds_logs \
 # end of EXTRA_CLEAN
 
+CFLAGS += $(KHDR_INCLUDES)
+
 include ../../lib.mk
diff --git a/tools/testing/selftests/net/rds/getsockopt.c b/tools/testing/selftests/net/rds/getsockopt.c
new file mode 100644
index 000000000000..93ff252c69b8
--- /dev/null
+++ b/tools/testing/selftests/net/rds/getsockopt.c
@@ -0,0 +1,208 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Exercise the RDS getsockopt() paths that were converted to the
+ * getsockopt_iter() / sockopt_t callback.
+ *
+ * Three distinct paths are covered:
+ *
+ *   - RDS_RECVERR and SO_RDS_TRANSPORT, which now return their int value
+ *     through copy_to_iter() and report the written length in opt->optlen.
+ *
+ *   - RDS_INFO_*, which pins the userspace buffer with
+ *     iov_iter_extract_pages() (including a non-zero starting page offset)
+ *     and lets the info producers memcpy the snapshot in under a spinlock.
+ *
+ * The kvec (in-kernel buffer) -> -EOPNOTSUPP path of rds_info_getsockopt()
+ * is not reachable from a userspace getsockopt() and so is not tested here.
+ */
+#include <errno.h>
+#include <stdint.h>
+#include <string.h>
+#include <unistd.h>
+#include <sys/mman.h>
+#include <sys/socket.h>
+#include <linux/rds.h>
+
+#include "../../kselftest_harness.h"
+
+#ifndef AF_RDS
+#define AF_RDS 21
+#endif
+
+FIXTURE(rds) {
+	int fd;
+};
+
+FIXTURE_SETUP(rds)
+{
+	self->fd = socket(AF_RDS, SOCK_SEQPACKET, 0);
+	if (self->fd < 0)
+		SKIP(return, "AF_RDS unavailable (errno %d) - load the rds module",
+		     errno);
+}
+
+FIXTURE_TEARDOWN(rds)
+{
+	if (self->fd >= 0)
+		close(self->fd);
+}
+
+/* RDS_RECVERR defaults to 0 and is reported back as a 4-byte int. */
+TEST_F(rds, recverr_default)
+{
+	socklen_t len = sizeof(int);
+	int val = 0xdeadbeef;
+
+	ASSERT_EQ(0, getsockopt(self->fd, SOL_RDS, RDS_RECVERR, &val, &len));
+	EXPECT_EQ(sizeof(int), len);
+	EXPECT_EQ(0, val);
+}
+
+/* A value set via setsockopt() must be readable back unchanged. */
+TEST_F(rds, recverr_set_get)
+{
+	socklen_t len = sizeof(int);
+	int val = 1;
+
+	ASSERT_EQ(0, setsockopt(self->fd, SOL_RDS, RDS_RECVERR, &val, len));
+
+	val = 0;
+	ASSERT_EQ(0, getsockopt(self->fd, SOL_RDS, RDS_RECVERR, &val, &len));
+	EXPECT_EQ(sizeof(int), len);
+	EXPECT_EQ(1, val);
+}
+
+/* A buffer smaller than an int is rejected with EINVAL, not silently. */
+TEST_F(rds, recverr_short_buffer)
+{
+	socklen_t len = sizeof(int) - 1;
+	char buf[sizeof(int)];
+
+	EXPECT_EQ(-1, getsockopt(self->fd, SOL_RDS, RDS_RECVERR, buf, &len));
+	EXPECT_EQ(EINVAL, errno);
+}
+
+/* An unbound socket reports RDS_TRANS_NONE for SO_RDS_TRANSPORT. */
+TEST_F(rds, transport_unbound)
+{
+	socklen_t len = sizeof(int);
+	int val = 0;
+
+	ASSERT_EQ(0, getsockopt(self->fd, SOL_RDS, SO_RDS_TRANSPORT, &val,
+				&len));
+	EXPECT_EQ(sizeof(int), len);
+	EXPECT_EQ(RDS_TRANS_NONE, (unsigned int)val);
+}
+
+TEST_F(rds, transport_short_buffer)
+{
+	socklen_t len = sizeof(int) - 1;
+	char buf[sizeof(int)];
+
+	EXPECT_EQ(-1, getsockopt(self->fd, SOL_RDS, SO_RDS_TRANSPORT, buf,
+				 &len));
+	EXPECT_EQ(EINVAL, errno);
+}
+
+/*
+ * RDS_INFO_COUNTERS with a zero-length buffer is the "probe" call: it must
+ * fail with ENOSPC and report the required snapshot size in optlen.
+ */
+TEST_F(rds, info_counters_probe)
+{
+	socklen_t len = 0;
+
+	EXPECT_EQ(-1, getsockopt(self->fd, SOL_RDS, RDS_INFO_COUNTERS, NULL,
+				 &len));
+	EXPECT_EQ(ENOSPC, errno);
+	EXPECT_GT(len, 0);
+	/* The snapshot is an array of fixed-size counter records. */
+	EXPECT_EQ(0, len % (socklen_t)sizeof(struct rds_info_counter));
+}
+
+/*
+ * A real snapshot into an unaligned userspace buffer exercises the
+ * iov_iter_extract_pages() path, including the non-zero offset0 handling
+ * that the patch reworked. Place the buffer at a non-page-aligned address
+ * spanning into the next page to make sure multi-page pinning works too.
+ */
+TEST_F(rds, info_counters_snapshot)
+{
+	struct rds_info_counter *ctr;
+	socklen_t need = 0, len;
+	long pagesz = sysconf(_SC_PAGESIZE);
+	size_t offset, map_len;
+	unsigned int i, n;
+	char *region, *buf;
+	int ret;
+
+	/* Probe for the required size. */
+	getsockopt(self->fd, SOL_RDS, RDS_INFO_COUNTERS, NULL, &need);
+	ASSERT_GT(need, 0);
+
+	/*
+	 * Place the buffer at a non-page-aligned offset that runs past the
+	 * first page boundary, and size the mapping from the probed length so
+	 * the test keeps working if the counter set grows.
+	 */
+	offset = pagesz - 64;
+	map_len = ((offset + need + pagesz - 1) / pagesz) * pagesz;
+
+	region = mmap(NULL, map_len, PROT_READ | PROT_WRITE,
+		      MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+	ASSERT_NE(MAP_FAILED, region);
+
+	buf = region + offset;
+
+	/*
+	 * On success the RDS_INFO path returns the positive per-element size
+	 * (lens.each) rather than 0, and writes the full snapshot length back
+	 * into optlen.
+	 */
+	len = need;
+	ret = getsockopt(self->fd, SOL_RDS, RDS_INFO_COUNTERS, buf, &len);
+	ASSERT_GE(ret, 0) {
+		TH_LOG("getsockopt snapshot failed: errno %d", errno);
+	}
+	EXPECT_EQ(sizeof(struct rds_info_counter), ret);
+	EXPECT_EQ(need, len);
+
+	/* The counter names must be NUL-terminated, non-empty strings. */
+	ctr = (struct rds_info_counter *)buf;
+	n = len / sizeof(*ctr);
+	ASSERT_GT(n, 0);
+	for (i = 0; i < n; i++) {
+		size_t namelen = strnlen((char *)ctr[i].name,
+					 sizeof(ctr[i].name));
+
+		EXPECT_GT(namelen, 0);
+		EXPECT_LT(namelen, sizeof(ctr[i].name));
+	}
+
+	munmap(region, map_len);
+}
+
+/*
+ * A non-zero but too-small buffer must report ENOSPC and the full required
+ * length, without corrupting memory past the buffer.
+ */
+TEST_F(rds, info_counters_short_buffer)
+{
+	socklen_t need = 0, len;
+	char small[sizeof(struct rds_info_counter)];
+
+	getsockopt(self->fd, SOL_RDS, RDS_INFO_COUNTERS, NULL, &need);
+	ASSERT_GT(need, 0);
+
+	/* Ask with a buffer guaranteed smaller than the full snapshot. */
+	if (need <= (socklen_t)sizeof(small))
+		SKIP(return, "snapshot fits in one record; nothing to test");
+
+	len = 1; /* < sizeof(struct rds_info_counter) */
+	EXPECT_EQ(-1, getsockopt(self->fd, SOL_RDS, RDS_INFO_COUNTERS, small,
+				 &len));
+	EXPECT_EQ(ENOSPC, errno);
+	EXPECT_EQ(need, len);
+}
+
+TEST_HARNESS_MAIN

-- 
2.53.0-Meta

[PATCH net-next v3 2/2] rds: convert to getsockopt_iter

[Breno Leitao]

Convert RDS socket's getsockopt implementation to use the new
getsockopt_iter callback with sockopt_t.

Key changes:
- Replace (char __user *optval, int __user *optlen) with sockopt_t *opt
- Use opt->optlen for buffer length (input) and returned size (output)
- Use copy_to_iter() instead of put_user()/copy_to_user()

The RDS_INFO_* snapshot path in rds_info_getsockopt() used to pin the
userspace buffer with pin_user_pages_fast() on the raw optval address;
the info producers then memcpy into those pages under a spinlock via
kmap_atomic() and so must not fault. Obtain the same page array and
starting offset from opt->iter_out with iov_iter_extract_pages(), which
pins for write because iter_out is ITER_DEST.

The page array is preallocated here (sized with iov_iter_npages()) and
passed in, so iov_iter_extract_pages() fills it in place rather than
allocating one for us; RDS therefore keeps ownership of the array on
every return path and frees it itself. The rds_info_iterator /
rds_info_copy machinery and all producer callbacks are unchanged.

Kernel buffers (ITER_KVEC) are not page-backed in a way the info
producers can use, so the RDS_INFO path returns -EOPNOTSUPP for them;
this matches the previous behaviour, where a kernel-buffer getsockopt
hit the WARN_ONCE() path in do_sock_getsockopt() and returned
-EOPNOTSUPP. The simple RDS_RECVERR and SO_RDS_TRANSPORT options keep
working for kernel buffers via copy_to_iter().

Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Allison Henderson <achender@kernel.org>
---
 net/rds/af_rds.c | 36 +++++++++++++++------------
 net/rds/info.c   | 76 ++++++++++++++++++++++++++++++++------------------------
 net/rds/info.h   |  3 +--
 3 files changed, 65 insertions(+), 50 deletions(-)

diff --git a/net/rds/af_rds.c b/net/rds/af_rds.c
index 6f4f9cf352bd..d5defe9172e3 100644
--- a/net/rds/af_rds.c
+++ b/net/rds/af_rds.c
@@ -37,6 +37,7 @@
 #include <linux/in.h>
 #include <linux/ipv6.h>
 #include <linux/poll.h>
+#include <linux/uio.h>
 #include <net/sock.h>
 
 #include "rds.h"
@@ -485,35 +486,36 @@ static int rds_setsockopt(struct socket *sock, int level, int optname,
 }
 
 static int rds_getsockopt(struct socket *sock, int level, int optname,
-			  char __user *optval, int __user *optlen)
+			  sockopt_t *opt)
 {
 	struct rds_sock *rs = rds_sk_to_rs(sock->sk);
 	int ret = -ENOPROTOOPT, len;
 	int trans;
+	int val;
 
 	if (level != SOL_RDS)
 		goto out;
 
-	if (get_user(len, optlen)) {
-		ret = -EFAULT;
-		goto out;
-	}
+	len = opt->optlen;
 
 	switch (optname) {
 	case RDS_INFO_FIRST ... RDS_INFO_LAST:
-		ret = rds_info_getsockopt(sock, optname, optval,
-					  optlen);
+		ret = rds_info_getsockopt(sock, optname, opt);
 		break;
 
 	case RDS_RECVERR:
-		if (len < sizeof(int))
+		if (len < sizeof(int)) {
 			ret = -EINVAL;
-		else
-		if (put_user(rs->rs_recverr, (int __user *) optval) ||
-		    put_user(sizeof(int), optlen))
+			break;
+		}
+		val = rs->rs_recverr;
+		if (copy_to_iter(&val, sizeof(int), &opt->iter_out) !=
+		    sizeof(int)) {
 			ret = -EFAULT;
-		else
+		} else {
+			opt->optlen = sizeof(int);
 			ret = 0;
+		}
 		break;
 	case SO_RDS_TRANSPORT:
 		if (len < sizeof(int)) {
@@ -522,11 +524,13 @@ static int rds_getsockopt(struct socket *sock, int level, int optname,
 		}
 		trans = (rs->rs_transport ? rs->rs_transport->t_type :
 			 RDS_TRANS_NONE); /* unbound */
-		if (put_user(trans, (int __user *)optval) ||
-		    put_user(sizeof(int), optlen))
+		if (copy_to_iter(&trans, sizeof(int), &opt->iter_out) !=
+		    sizeof(int)) {
 			ret = -EFAULT;
-		else
+		} else {
+			opt->optlen = sizeof(int);
 			ret = 0;
+		}
 		break;
 	default:
 		break;
@@ -653,7 +657,7 @@ static const struct proto_ops rds_proto_ops = {
 	.listen =	sock_no_listen,
 	.shutdown =	sock_no_shutdown,
 	.setsockopt =	rds_setsockopt,
-	.getsockopt =	rds_getsockopt,
+	.getsockopt_iter =	rds_getsockopt,
 	.sendmsg =	rds_sendmsg,
 	.recvmsg =	rds_recvmsg,
 	.mmap =		sock_no_mmap,
diff --git a/net/rds/info.c b/net/rds/info.c
index f1b29994934a..499b3774860e 100644
--- a/net/rds/info.c
+++ b/net/rds/info.c
@@ -35,6 +35,7 @@
 #include <linux/slab.h>
 #include <linux/proc_fs.h>
 #include <linux/export.h>
+#include <linux/uio.h>
 
 #include "rds.h"
 
@@ -144,60 +145,68 @@ void rds_info_copy(struct rds_info_iterator *iter, void *data,
 EXPORT_SYMBOL_GPL(rds_info_copy);
 
 /*
- * @optval points to the userspace buffer that the information snapshot
- * will be copied into.
- *
- * @optlen on input is the size of the buffer in userspace.  @optlen
- * on output is the size of the requested snapshot in bytes.
+ * @opt->iter_out describes the buffer that the information snapshot will be
+ * copied into, and @opt->optlen is the size of that buffer on input.  On
+ * output @opt->optlen is set to the size of the requested snapshot in bytes.
  *
  * This function returns -errno if there is a failure, particularly -ENOSPC
- * if the given userspace buffer was not large enough to fit the snapshot.
- * On success it returns the positive number of bytes of each array element
- * in the snapshot.
+ * if the given buffer was not large enough to fit the snapshot.  On success
+ * it returns the positive number of bytes of each array element in the
+ * snapshot.
  */
-int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
-			int __user *optlen)
+int rds_info_getsockopt(struct socket *sock, int optname, sockopt_t *opt)
 {
 	struct rds_info_iterator iter;
 	struct rds_info_lengths lens;
 	unsigned long nr_pages = 0;
-	unsigned long start;
 	rds_info_func func;
 	struct page **pages = NULL;
+	size_t offset0 = 0;
+	int npages = 0;
 	int ret;
 	int len;
 	int total;
 
-	if (get_user(len, optlen)) {
-		ret = -EFAULT;
-		goto out;
-	}
+	len = opt->optlen;
 
 	/* check for all kinds of wrapping and the like */
-	start = (unsigned long)optval;
-	if (len < 0 || len > INT_MAX - PAGE_SIZE + 1 || start + len < start) {
+	if (len < 0 || len > INT_MAX - PAGE_SIZE + 1) {
 		ret = -EINVAL;
 		goto out;
 	}
 
+	/* The info producers write into the pages with kmap_atomic() while
+	 * holding a spinlock, so they need a genuine page-backed user buffer.
+	 */
+	if (!user_backed_iter(&opt->iter_out)) {
+		ret = -EOPNOTSUPP;
+		goto out;
+	}
+
 	/* a 0 len call is just trying to probe its length */
 	if (len == 0)
 		goto call_func;
 
-	nr_pages = (PAGE_ALIGN(start + len) - (start & PAGE_MASK))
-			>> PAGE_SHIFT;
-
-	pages = kmalloc_objs(struct page *, nr_pages);
+	/*
+	 * Preallocate the page array and pass it in so that
+	 * iov_iter_extract_pages() fills it in place rather than allocating
+	 * one for us.  Handing it a non-NULL array keeps ownership of the
+	 * array with us on every return path, instead of depending on the
+	 * iterator code to allocate and hand it back.
+	 */
+	npages = iov_iter_npages(&opt->iter_out, INT_MAX);
+	pages = kvmalloc_array(npages, sizeof(*pages), GFP_KERNEL);
 	if (!pages) {
 		ret = -ENOMEM;
 		goto out;
 	}
-	ret = pin_user_pages_fast(start, nr_pages, FOLL_WRITE, pages);
-	if (ret != nr_pages) {
-		if (ret > 0)
-			nr_pages = ret;
-		else
-			nr_pages = 0;
+
+	ret = iov_iter_extract_pages(&opt->iter_out, &pages, len, npages,
+				     0, &offset0);
+	if (ret < 0)
+		goto out;
+	nr_pages = DIV_ROUND_UP(offset0 + ret, PAGE_SIZE);
+	if (ret != len) {
 		ret = -EAGAIN; /* XXX ? */
 		goto out;
 	}
@@ -213,7 +222,7 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
 
 	iter.pages = pages;
 	iter.addr = NULL;
-	iter.offset = start & (PAGE_SIZE - 1);
+	iter.offset = offset0;
 
 	func(sock, len, &iter, &lens);
 	BUG_ON(lens.each == 0);
@@ -230,13 +239,16 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
 		ret = lens.each;
 	}
 
-	if (put_user(len, optlen))
-		ret = -EFAULT;
+	opt->optlen = len;
 
 out:
-	if (pages)
+	/*
+	 * iov_iter_extract_pages() pins only user-backed (ubuf) iters;
+	 * iov_iter_extract_will_pin() reports whether an unpin is owed here.
+	 */
+	if (pages && iov_iter_extract_will_pin(&opt->iter_out))
 		unpin_user_pages(pages, nr_pages);
-	kfree(pages);
+	kvfree(pages);
 
 	return ret;
 }
diff --git a/net/rds/info.h b/net/rds/info.h
index a069b51c4679..1aab62ab6d00 100644
--- a/net/rds/info.h
+++ b/net/rds/info.h
@@ -21,8 +21,7 @@ typedef void (*rds_info_func)(struct socket *sock, unsigned int len,
 
 void rds_info_register_func(int optname, rds_info_func func);
 void rds_info_deregister_func(int optname, rds_info_func func);
-int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
-			int __user *optlen);
+int rds_info_getsockopt(struct socket *sock, int optname, sockopt_t *opt);
 void rds_info_copy(struct rds_info_iterator *iter, void *data,
 		   unsigned long bytes);
 void rds_info_iter_unmap(struct rds_info_iterator *iter);

-- 
2.53.0-Meta

Re: [PATCH net-next v3 0/2] net: rds: convert rds to getsockopt_iter

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (main)
by Paolo Abeni <pabeni@redhat.com>:

On Mon, 08 Jun 2026 02:44:56 -0700 you wrote:
> This series continues the conversion of the remaining proto_ops getsockopt
> callbacks to the new getsockopt_iter callback introduced in commit
> 67fab22a7adc ("net: add getsockopt_iter callback to proto_ops"), this time
> for RDS.
> 
> RDS is a little more involved than the protocols converted so far, because
> the RDS_INFO_* options snapshot kernel state directly into the destination
> buffer: the info producers memcpy into the pages under a spinlock via
> kmap_atomic() and so must not fault.
> 
> [...]

Here is the summary with links:
  - [net-next,v3,1/2] selftests: net: rds: add getsockopt() conversion test
    https://git.kernel.org/netdev/net-next/c/b74360369e13
  - [net-next,v3,2/2] rds: convert to getsockopt_iter
    https://git.kernel.org/netdev/net-next/c/6e94eeb2a2a6

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

Re: [PATCH net-next v3 2/2] rds: convert to getsockopt_iter: manual merge

[Matthieu Baerts]

[-- Attachment #1: Type: text/plain, Size: 3346 bytes --]

Hi Breno, Allison,

On 08/06/2026 11:44, Breno Leitao wrote:
> Convert RDS socket's getsockopt implementation to use the new
> getsockopt_iter callback with sockopt_t.
> 
> Key changes:
> - Replace (char __user *optval, int __user *optlen) with sockopt_t *opt
> - Use opt->optlen for buffer length (input) and returned size (output)
> - Use copy_to_iter() instead of put_user()/copy_to_user()
> 
> The RDS_INFO_* snapshot path in rds_info_getsockopt() used to pin the
> userspace buffer with pin_user_pages_fast() on the raw optval address;
> the info producers then memcpy into those pages under a spinlock via
> kmap_atomic() and so must not fault. Obtain the same page array and
> starting offset from opt->iter_out with iov_iter_extract_pages(), which
> pins for write because iter_out is ITER_DEST.
> 
> The page array is preallocated here (sized with iov_iter_npages()) and
> passed in, so iov_iter_extract_pages() fills it in place rather than
> allocating one for us; RDS therefore keeps ownership of the array on
> every return path and frees it itself. The rds_info_iterator /
> rds_info_copy machinery and all producer callbacks are unchanged.
> 
> Kernel buffers (ITER_KVEC) are not page-backed in a way the info
> producers can use, so the RDS_INFO path returns -EOPNOTSUPP for them;
> this matches the previous behaviour, where a kernel-buffer getsockopt
> hit the WARN_ONCE() path in do_sock_getsockopt() and returned
> -EOPNOTSUPP. The simple RDS_RECVERR and SO_RDS_TRANSPORT options keep
> working for kernel buffers via copy_to_iter().

(...)

> diff --git a/net/rds/info.c b/net/rds/info.c
> index f1b29994934a..499b3774860e 100644
> --- a/net/rds/info.c
> +++ b/net/rds/info.c

(...)

> @@ -230,13 +239,16 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
>  		ret = lens.each;
>  	}
>  
> -	if (put_user(len, optlen))
> -		ret = -EFAULT;
> +	opt->optlen = len;
>  
>  out:
> -	if (pages)
> +	/*
> +	 * iov_iter_extract_pages() pins only user-backed (ubuf) iters;
> +	 * iov_iter_extract_will_pin() reports whether an unpin is owed here.
> +	 */
> +	if (pages && iov_iter_extract_will_pin(&opt->iter_out))
>  		unpin_user_pages(pages, nr_pages);

FYI, we got a small conflict when merging 'net' in 'net-next' in the
MPTCP tree due to this patch applied in 'net':

  f512db8267b73 ("rds: mark snapshot pages dirty in rds_info_getsockopt()")

and this one from 'net-next':

  6e94eeb2a2a6 ("rds: convert to getsockopt_iter")

----- Generic Message -----
The best is to avoid conflicts between 'net' and 'net-next' trees but if
they cannot be avoided when preparing patches, a note about how to fix
them is much appreciated.

The conflict has been resolved on our side [1] and the resolution we
suggest is attached to this email. Please report any issues linked to
this conflict resolution as it might be used by others. If you worked on
the mentioned patches, don't hesitate to ACK this conflict resolution.
---------------------------

Regarding this conflict, I took the modification from net-next, but
using unpin_user_pages_dirty_lock() from net.

Rerere cache is available in [2].

Cheers,
Matt

1: https://github.com/multipath-tcp/mptcp_net-next/commit/a8d41e018cc6
2: https://github.com/multipath-tcp/mptcp-upstream-rr-cache/commit/88eeb

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.

[-- Attachment #2: a8d41e018cc69a6546ef6acc4a97bcbeb3e75d43.patch --]
[-- Type: text/x-patch, Size: 643 bytes --]

diff --cc net/rds/info.c
index 499b3774860e,17061f6ff74e..21b32eb16559
--- a/net/rds/info.c
+++ b/net/rds/info.c
@@@ -239,16 -230,13 +239,16 @@@ call_func
  		ret = lens.each;
  	}
  
 -	if (put_user(len, optlen))
 -		ret = -EFAULT;
 +	opt->optlen = len;
  
  out:
 -	if (pages)
 +	/*
 +	 * iov_iter_extract_pages() pins only user-backed (ubuf) iters;
 +	 * iov_iter_extract_will_pin() reports whether an unpin is owed here.
 +	 */
 +	if (pages && iov_iter_extract_will_pin(&opt->iter_out))
- 		unpin_user_pages(pages, nr_pages);
+ 		unpin_user_pages_dirty_lock(pages, nr_pages, true);
 -	kfree(pages);
 +	kvfree(pages);
  
  	return ret;
  }

Re: [PATCH net-next v3 2/2] rds: convert to getsockopt_iter: manual merge

[Allison Henderson]

On Thu, 2026-06-11 at 14:52 +0200, Matthieu Baerts wrote:
> Hi Breno, Allison,
> 
> On 08/06/2026 11:44, Breno Leitao wrote:
> > Convert RDS socket's getsockopt implementation to use the new
> > getsockopt_iter callback with sockopt_t.
> > 
> > Key changes:
> > - Replace (char __user *optval, int __user *optlen) with sockopt_t *opt
> > - Use opt->optlen for buffer length (input) and returned size (output)
> > - Use copy_to_iter() instead of put_user()/copy_to_user()
> > 
> > The RDS_INFO_* snapshot path in rds_info_getsockopt() used to pin the
> > userspace buffer with pin_user_pages_fast() on the raw optval address;
> > the info producers then memcpy into those pages under a spinlock via
> > kmap_atomic() and so must not fault. Obtain the same page array and
> > starting offset from opt->iter_out with iov_iter_extract_pages(), which
> > pins for write because iter_out is ITER_DEST.
> > 
> > The page array is preallocated here (sized with iov_iter_npages()) and
> > passed in, so iov_iter_extract_pages() fills it in place rather than
> > allocating one for us; RDS therefore keeps ownership of the array on
> > every return path and frees it itself. The rds_info_iterator /
> > rds_info_copy machinery and all producer callbacks are unchanged.
> > 
> > Kernel buffers (ITER_KVEC) are not page-backed in a way the info
> > producers can use, so the RDS_INFO path returns -EOPNOTSUPP for them;
> > this matches the previous behaviour, where a kernel-buffer getsockopt
> > hit the WARN_ONCE() path in do_sock_getsockopt() and returned
> > -EOPNOTSUPP. The simple RDS_RECVERR and SO_RDS_TRANSPORT options keep
> > working for kernel buffers via copy_to_iter().
> 
> (...)
> 
> > diff --git a/net/rds/info.c b/net/rds/info.c
> > index f1b29994934a..499b3774860e 100644
> > --- a/net/rds/info.c
> > +++ b/net/rds/info.c
> 
> (...)
> 
> > @@ -230,13 +239,16 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
> >  		ret = lens.each;
> >  	}
> >  
> > -	if (put_user(len, optlen))
> > -		ret = -EFAULT;
> > +	opt->optlen = len;
> >  
> >  out:
> > -	if (pages)
> > +	/*
> > +	 * iov_iter_extract_pages() pins only user-backed (ubuf) iters;
> > +	 * iov_iter_extract_will_pin() reports whether an unpin is owed here.
> > +	 */
> > +	if (pages && iov_iter_extract_will_pin(&opt->iter_out))
> >  		unpin_user_pages(pages, nr_pages);
> 
> FYI, we got a small conflict when merging 'net' in 'net-next' in the
> MPTCP tree due to this patch applied in 'net':
> 
>   f512db8267b73 ("rds: mark snapshot pages dirty in rds_info_getsockopt()")
> 
> and this one from 'net-next':
> 
>   6e94eeb2a2a6 ("rds: convert to getsockopt_iter")
> 
> ----- Generic Message -----
> The best is to avoid conflicts between 'net' and 'net-next' trees but if
> they cannot be avoided when preparing patches, a note about how to fix
> them is much appreciated.
> 
> The conflict has been resolved on our side [1] and the resolution we
> suggest is attached to this email. Please report any issues linked to
> this conflict resolution as it might be used by others. If you worked on
> the mentioned patches, don't hesitate to ACK this conflict resolution.
> ---------------------------
> 
> Regarding this conflict, I took the modification from net-next, but
> using unpin_user_pages_dirty_lock() from net.

Hi Matt,

Yes, the conflict resolution looks correct to me.  
Thank you for fixing this.

Allison

> 
> Rerere cache is available in [2].
> 
> Cheers,
> Matt
> 
> 1: https://github.com/multipath-tcp/mptcp_net-next/commit/a8d41e018cc6
> 2: https://github.com/multipath-tcp/mptcp-upstream-rr-cache/commit/88eeb
> 
> Cheers,
> Matt

Re: [PATCH net-next v3 2/2] rds: convert to getsockopt_iter: manual merge

[Breno Leitao]

Hello Matthieu,

On Thu, Jun 11, 2026 at 02:52:36PM +0200, Matthieu Baerts wrote:
> > +	/*
> > +	 * iov_iter_extract_pages() pins only user-backed (ubuf) iters;
> > +	 * iov_iter_extract_will_pin() reports whether an unpin is owed here.
> > +	 */
> > +	if (pages && iov_iter_extract_will_pin(&opt->iter_out))
> >  		unpin_user_pages(pages, nr_pages);
>
> FYI, we got a small conflict when merging 'net' in 'net-next' in the
> MPTCP tree due to this patch applied in 'net':
>
>   f512db8267b73 ("rds: mark snapshot pages dirty in rds_info_getsockopt()")
>
> and this one from 'net-next':
>
>   6e94eeb2a2a6 ("rds: convert to getsockopt_iter")

I was aware of the conflict but didn't realize a note would be helpful
for the merge. I should have included one.

Could you point me to an example commit/patch that contains such a note so I
can understand the expected format and procedure?

Apologies for the omission.
--breno

Re: [PATCH net-next v3 2/2] rds: convert to getsockopt_iter: manual merge

[Matthieu Baerts]

Hi Breno,

On 12/06/2026 10:19, Breno Leitao wrote:
> Hello Matthieu,
> 
> On Thu, Jun 11, 2026 at 02:52:36PM +0200, Matthieu Baerts wrote:
>>> +	/*
>>> +	 * iov_iter_extract_pages() pins only user-backed (ubuf) iters;
>>> +	 * iov_iter_extract_will_pin() reports whether an unpin is owed here.
>>> +	 */
>>> +	if (pages && iov_iter_extract_will_pin(&opt->iter_out))
>>>  		unpin_user_pages(pages, nr_pages);
>>
>> FYI, we got a small conflict when merging 'net' in 'net-next' in the
>> MPTCP tree due to this patch applied in 'net':
>>
>>   f512db8267b73 ("rds: mark snapshot pages dirty in rds_info_getsockopt()")
>>
>> and this one from 'net-next':
>>
>>   6e94eeb2a2a6 ("rds: convert to getsockopt_iter")
> 
> I was aware of the conflict but didn't realize a note would be helpful
> for the merge. I should have included one.
> 
> Could you point me to an example commit/patch that contains such a note so I
> can understand the expected format and procedure?

In this particular example, I think it would have been easier to have
waited for the fix to land in net-next -- after the weekly sync with net
-- and then send the net-next patches.

When this cannot be avoided, then you can mention the conflict, and
ideally share a diff of the resolution, plus a description, especially
when it is not obvious, when simply saying "take the version from X" is
helpful, when extra modifications are needed, etc. e.g. [1]. Something
similar to what Mark is usually doing on the linux-next ML, or what I
did here.

[1]
https://lore.kernel.org/98386125-c0bb-495e-b2ba-2765aaed19d8@oss.qualcomm.com

> Apologies for the omission.
No problem, it happens fairly regularly ;)

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.