Archive-only list for patches
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: stable@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	patches@lists.linux.dev, Jens Axboe <axboe@kernel.dk>
Subject: [PATCH 5.10 02/63] iov_iter: add helper to save iov_iter state
Date: Tue,  3 Jan 2023 09:13:32 +0100	[thread overview]
Message-ID: <20230103081308.699550936@linuxfoundation.org> (raw)
In-Reply-To: <20230103081308.548338576@linuxfoundation.org>

From: Jens Axboe <axboe@kernel.dk>

[ Upstream commit 8fb0f47a9d7acf620d0fd97831b69da9bc5e22ed ]

In an ideal world, when someone is passed an iov_iter and returns X bytes,
then X bytes would have been consumed/advanced from the iov_iter. But we
have use cases that always consume the entire iterator, a few examples
of that are iomap and bdev O_DIRECT. This means we cannot rely on the
state of the iov_iter once we've called ->read_iter() or ->write_iter().

This would be easier if we didn't always have to deal with truncate of
the iov_iter, as rewinding would be trivial without that. We recently
added a commit to track the truncate state, but that grew the iov_iter
by 8 bytes and wasn't the best solution.

Implement a helper to save enough of the iov_iter state to sanely restore
it after we've called the read/write iterator helpers. This currently
only works for IOVEC/BVEC/KVEC as that's all we need, support for other
iterator types are left as an exercise for the reader.

Link: https://lore.kernel.org/linux-fsdevel/CAHk-=wiacKV4Gh-MYjteU0LwNBSGpWrK-Ov25HdqB1ewinrFPg@mail.gmail.com/
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 include/linux/uio.h |   15 +++++++++++++++
 lib/iov_iter.c      |   52 +++++++++++++++++++++++++++++++++-------------------
 2 files changed, 48 insertions(+), 19 deletions(-)

--- a/include/linux/uio.h
+++ b/include/linux/uio.h
@@ -26,6 +26,12 @@ enum iter_type {
 	ITER_DISCARD = 64,
 };
 
+struct iov_iter_state {
+	size_t iov_offset;
+	size_t count;
+	unsigned long nr_segs;
+};
+
 struct iov_iter {
 	/*
 	 * Bit 0 is the read/write bit, set if we're writing.
@@ -55,6 +61,14 @@ static inline enum iter_type iov_iter_ty
 	return i->type & ~(READ | WRITE);
 }
 
+static inline void iov_iter_save_state(struct iov_iter *iter,
+				       struct iov_iter_state *state)
+{
+	state->iov_offset = iter->iov_offset;
+	state->count = iter->count;
+	state->nr_segs = iter->nr_segs;
+}
+
 static inline bool iter_is_iovec(const struct iov_iter *i)
 {
 	return iov_iter_type(i) == ITER_IOVEC;
@@ -226,6 +240,7 @@ ssize_t iov_iter_get_pages(struct iov_it
 ssize_t iov_iter_get_pages_alloc(struct iov_iter *i, struct page ***pages,
 			size_t maxsize, size_t *start);
 int iov_iter_npages(const struct iov_iter *i, int maxpages);
+void iov_iter_restore(struct iov_iter *i, struct iov_iter_state *state);
 
 const void *dup_iter(struct iov_iter *new, struct iov_iter *old, gfp_t flags);
 
--- a/lib/iov_iter.c
+++ b/lib/iov_iter.c
@@ -1836,24 +1836,38 @@ int import_single_range(int rw, void __u
 }
 EXPORT_SYMBOL(import_single_range);
 
-int iov_iter_for_each_range(struct iov_iter *i, size_t bytes,
-			    int (*f)(struct kvec *vec, void *context),
-			    void *context)
+/**
+ * iov_iter_restore() - Restore a &struct iov_iter to the same state as when
+ *     iov_iter_save_state() was called.
+ *
+ * @i: &struct iov_iter to restore
+ * @state: state to restore from
+ *
+ * Used after iov_iter_save_state() to bring restore @i, if operations may
+ * have advanced it.
+ *
+ * Note: only works on ITER_IOVEC, ITER_BVEC, and ITER_KVEC
+ */
+void iov_iter_restore(struct iov_iter *i, struct iov_iter_state *state)
 {
-	struct kvec w;
-	int err = -EINVAL;
-	if (!bytes)
-		return 0;
-
-	iterate_all_kinds(i, bytes, v, -EINVAL, ({
-		w.iov_base = kmap(v.bv_page) + v.bv_offset;
-		w.iov_len = v.bv_len;
-		err = f(&w, context);
-		kunmap(v.bv_page);
-		err;}), ({
-		w = v;
-		err = f(&w, context);})
-	)
-	return err;
+	if (WARN_ON_ONCE(!iov_iter_is_bvec(i) && !iter_is_iovec(i)) &&
+			 !iov_iter_is_kvec(i))
+		return;
+	i->iov_offset = state->iov_offset;
+	i->count = state->count;
+	/*
+	 * For the *vec iters, nr_segs + iov is constant - if we increment
+	 * the vec, then we also decrement the nr_segs count. Hence we don't
+	 * need to track both of these, just one is enough and we can deduct
+	 * the other from that. ITER_KVEC and ITER_IOVEC are the same struct
+	 * size, so we can just increment the iov pointer as they are unionzed.
+	 * ITER_BVEC _may_ be the same size on some archs, but on others it is
+	 * not. Be safe and handle it separately.
+	 */
+	BUILD_BUG_ON(sizeof(struct iovec) != sizeof(struct kvec));
+	if (iov_iter_is_bvec(i))
+		i->bvec -= state->nr_segs - i->nr_segs;
+	else
+		i->iov -= state->nr_segs - i->nr_segs;
+	i->nr_segs = state->nr_segs;
 }
-EXPORT_SYMBOL(iov_iter_for_each_range);



  parent reply	other threads:[~2023-01-03  8:14 UTC|newest]

Thread overview: 74+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-03  8:13 [PATCH 5.10 00/63] 5.10.162-rc1 review Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 01/63] kernel: provide create_io_thread() helper Greg Kroah-Hartman
2023-01-03  8:13 ` Greg Kroah-Hartman [this message]
2023-01-03  8:13 ` [PATCH 5.10 03/63] saner calling conventions for unlazy_child() Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 04/63] fs: add support for LOOKUP_CACHED Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 05/63] fix handling of nd->depth on LOOKUP_CACHED failures in try_to_unlazy* Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 06/63] Make sure nd->path.mnt and nd->path.dentry are always valid pointers Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 07/63] fs: expose LOOKUP_CACHED through openat2() RESOLVE_CACHED Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 08/63] tools headers UAPI: Sync openat2.h with the kernel sources Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 09/63] net: provide __sys_shutdown_sock() that takes a socket Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 10/63] net: add accept helper not installing fd Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 11/63] signal: Add task_sigpending() helper Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 12/63] fs: make do_renameat2() take struct filename Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 13/63] file: Rename __close_fd_get_file close_fd_get_file Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 14/63] fs: provide locked helper variant of close_fd_get_file() Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 15/63] entry: Add support for TIF_NOTIFY_SIGNAL Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 16/63] task_work: Use TIF_NOTIFY_SIGNAL if available Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 17/63] x86: Wire up TIF_NOTIFY_SIGNAL Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 18/63] arc: add support for TIF_NOTIFY_SIGNAL Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 19/63] arm64: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 20/63] m68k: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 21/63] nios32: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 22/63] parisc: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 23/63] powerpc: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 24/63] mips: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 25/63] s390: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 26/63] um: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 27/63] sh: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 28/63] openrisc: " Greg Kroah-Hartman
2023-01-03  8:13 ` [PATCH 5.10 29/63] csky: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 30/63] hexagon: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 31/63] microblaze: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 32/63] arm: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 33/63] xtensa: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 34/63] alpha: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 35/63] c6x: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 36/63] h8300: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 37/63] ia64: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 38/63] nds32: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 39/63] riscv: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 40/63] sparc: " Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 41/63] ia64: dont call handle_signal() unless theres actually a signal queued Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 42/63] ARC: unbork 5.11 bootup: fix snafu in _TIF_NOTIFY_SIGNAL handling Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 43/63] alpha: fix TIF_NOTIFY_SIGNAL handling Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 44/63] task_work: remove legacy TWA_SIGNAL path Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 45/63] kernel: remove checking for TIF_NOTIFY_SIGNAL Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 46/63] coredump: Limit what can interrupt coredumps Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 47/63] kernel: allow fork with TIF_NOTIFY_SIGNAL pending Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 48/63] entry/kvm: Exit to user mode when TIF_NOTIFY_SIGNAL is set Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 49/63] arch: setup PF_IO_WORKER threads like PF_KTHREAD Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 50/63] arch: ensure parisc/powerpc handle PF_IO_WORKER in copy_thread() Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 51/63] x86/process: setup io_threads more like normal user space threads Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 52/63] kernel: stop masking signals in create_io_thread() Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 53/63] kernel: dont call do_exit() for PF_IO_WORKER threads Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 54/63] task_work: add helper for more targeted task_work canceling Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 55/63] io_uring: import 5.15-stable io_uring Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 56/63] signal: kill JOBCTL_TASK_WORK Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 57/63] task_work: unconditionally run task_work from get_signal() Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 58/63] net: remove cmsg restriction from io_uring based send/recvmsg calls Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 59/63] Revert "proc: dont allow async path resolution of /proc/thread-self components" Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 60/63] Revert "proc: dont allow async path resolution of /proc/self components" Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 61/63] eventpoll: add EPOLL_URING_WAKE poll wakeup flag Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 62/63] eventfd: provide a eventfd_signal_mask() helper Greg Kroah-Hartman
2023-01-03  8:14 ` [PATCH 5.10 63/63] io_uring: pass in EPOLL_URING_WAKE for eventfd signaling and wakeups Greg Kroah-Hartman
2023-01-03 13:44 ` [PATCH 5.10 00/63] 5.10.162-rc1 review Pavel Machek
2023-01-03 14:51 ` Guenter Roeck
2023-01-03 15:32 ` Jon Hunter
2023-01-03 16:16 ` Joel Fernandes
2023-01-04  5:29   ` Greg Kroah-Hartman
2023-01-04 21:56     ` Joel Fernandes
2023-01-05 11:43       ` Greg Kroah-Hartman
2023-01-03 18:16 ` Naresh Kamboju
2023-01-03 18:27 ` Florian Fainelli
2023-01-03 18:59 ` Allen Pais

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230103081308.699550936@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=axboe@kernel.dk \
    --cc=patches@lists.linux.dev \
    --cc=stable@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox