lkml.org 
[lkml]   [2020]   [Mar]   [31]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[RFC PATCH 08/12] fs/io_uring: Use iovec_import() not import_iovec().
Date

This is a mechanical change to this horrid code.
I think it is correct.

Signed-off-by: David Laight <david.laight@aculab.com>
---
fs/io_uring.c | 165 +++++++++++++++++++++++++++++++---------------------------
1 file changed, 87 insertions(+), 78 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index d8dc2e2..27d66cf 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -77,6 +77,10 @@
#include <linux/eventpoll.h>
#include <linux/fs_struct.h>

+/* Temporary for commit bisection */
+#define sendmsg_copy_msghdr(a, b, c, d) sendmsg_copy_msghdr(a, b, c, (void *)d)
+#define recvmsg_copy_msghdr(a, b, c, d, e) recvmsg_copy_msghdr(a, b, c, d, (void *)e)
+
#define CREATE_TRACE_POINTS
#include <trace/events/io_uring.h>

@@ -435,7 +439,7 @@ struct io_async_connect {
};

struct io_async_msghdr {
- struct iovec fast_iov[UIO_FASTIOV];
+ struct iovec_cache fast_iov;
struct iovec *iov;
struct sockaddr __user *uaddr;
struct msghdr msg;
@@ -443,7 +447,7 @@ struct io_async_msghdr {
};

struct io_async_rw {
- struct iovec fast_iov[UIO_FASTIOV];
+ struct iovec_cache fast_iov;
struct iovec *iov;
ssize_t nr_segs;
ssize_t size;
@@ -2052,47 +2056,39 @@ static ssize_t io_import_fixed(struct io_kiocb *req, int rw,
return len;
}

-static ssize_t io_import_iovec(int rw, struct io_kiocb *req,
- struct iovec **iovec, struct iov_iter *iter)
+static struct iovec *io_import_iovec(int rw, struct io_kiocb *req,
+ struct iovec_cache *cache, struct iov_iter *iter)
{
void __user *buf = u64_to_user_ptr(req->rw.addr);
size_t sqe_len = req->rw.len;
u8 opcode;

opcode = req->opcode;
- if (opcode == IORING_OP_READ_FIXED || opcode == IORING_OP_WRITE_FIXED) {
- *iovec = NULL;
- return io_import_fixed(req, rw, iter);
- }
+ if (opcode == IORING_OP_READ_FIXED || opcode == IORING_OP_WRITE_FIXED)
+ return ERR_PTR(io_import_fixed(req, rw, iter));

/* buffer index only valid with fixed read/write */
if (req->rw.kiocb.private)
- return -EINVAL;
+ return ERR_PTR(-EINVAL);

- if (opcode == IORING_OP_READ || opcode == IORING_OP_WRITE) {
- ssize_t ret;
- ret = import_single_range(rw, buf, sqe_len, *iovec, iter);
- *iovec = NULL;
- return ret < 0 ? ret : sqe_len;
- }
+ if (opcode == IORING_OP_READ || opcode == IORING_OP_WRITE)
+ return ERR_PTR(import_single_range(rw, buf, sqe_len, cache->iov, iter));

if (req->io) {
struct io_async_rw *iorw = &req->io->rw;

- *iovec = iorw->iov;
- iov_iter_init(iter, rw, *iovec, iorw->nr_segs, iorw->size);
- if (iorw->iov == iorw->fast_iov)
- *iovec = NULL;
- return iorw->size;
+ iov_iter_init(iter, rw, iorw->iov, iorw->nr_segs, iorw->size);
+ if (iorw->iov != iorw->fast_iov.iov)
+ return iorw->iov;
+ return NULL;
}

#ifdef CONFIG_COMPAT
if (req->ctx->compat)
- return compat_import_iovec(rw, buf, sqe_len, UIO_FASTIOV,
- iovec, iter);
+ return compat_iovec_import(rw, buf, sqe_len, cache, iter);
#endif

- return import_iovec(rw, buf, sqe_len, UIO_FASTIOV, iovec, iter);
+ return iovec_import(rw, buf, sqe_len, cache, iter);
}

/*
@@ -2154,13 +2150,13 @@ static ssize_t loop_rw_iter(int rw, struct file *file, struct kiocb *kiocb,
}

static void io_req_map_rw(struct io_kiocb *req, struct iovec *iovec,
- struct iovec *fast_iov, struct iov_iter *iter)
+ struct iovec_cache *fast_iov, struct iov_iter *iter)
{
req->io->rw.nr_segs = iter->nr_segs;
req->io->rw.size = iter->count;
req->io->rw.iov = iovec;
if (!req->io->rw.iov) {
- req->io->rw.iov = req->io->rw.fast_iov;
+ req->io->rw.iov = req->io->rw.fast_iov.iov;
memcpy(req->io->rw.iov, fast_iov,
sizeof(struct iovec) * iter->nr_segs);
} else {
@@ -2177,7 +2173,7 @@ static int io_alloc_async_ctx(struct io_kiocb *req)
}

static int io_setup_async_rw(struct io_kiocb *req, struct iovec *iovec,
- struct iovec *fast_iov, struct iov_iter *iter)
+ struct iovec_cache *fast_iov, struct iov_iter *iter)
{
if (!io_op_defs[req->opcode].async_ctx)
return 0;
@@ -2195,6 +2191,7 @@ static int io_read_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe,
{
struct io_async_ctx *io;
struct iov_iter iter;
+ struct iovec *iov;
ssize_t ret;

ret = io_prep_rw(req, sqe, force_nonblock);
@@ -2209,29 +2206,30 @@ static int io_read_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe,
return 0;

io = req->io;
- io->rw.iov = io->rw.fast_iov;
req->io = NULL;
- ret = io_import_iovec(READ, req, &io->rw.iov, &iter);
+ iov = io_import_iovec(READ, req, &io->rw.fast_iov, &iter);
req->io = io;
- if (ret < 0)
- return ret;
+ if (IS_ERR(iov))
+ return PTR_ERR(iov);
+ io->rw.iov = iov;

- io_req_map_rw(req, io->rw.iov, io->rw.fast_iov, &iter);
+ io_req_map_rw(req, io->rw.iov, &io->rw.fast_iov, &iter);
return 0;
}

static int io_read(struct io_kiocb *req, struct io_kiocb **nxt,
bool force_nonblock)
{
- struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
+ struct iovec_cache cache;
+ struct iovec *iovec;
struct kiocb *kiocb = &req->rw.kiocb;
struct iov_iter iter;
size_t iov_count;
ssize_t ret;

- ret = io_import_iovec(READ, req, &iovec, &iter);
- if (ret < 0)
- return ret;
+ iovec = io_import_iovec(READ, req, &cache, &iter);
+ if (IS_ERR(iovec))
+ return PTR_ERR(iovec);

/* Ensure we clear previously set non-block flag */
if (!force_nonblock)
@@ -2265,7 +2263,7 @@ static int io_read(struct io_kiocb *req, struct io_kiocb **nxt,
kiocb_done(kiocb, ret2, nxt, req->in_async);
} else {
copy_iov:
- ret = io_setup_async_rw(req, iovec, inline_vecs, &iter);
+ ret = io_setup_async_rw(req, iovec, &cache, &iter);
if (ret)
goto out_free;
return -EAGAIN;
@@ -2282,6 +2280,7 @@ static int io_write_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe,
{
struct io_async_ctx *io;
struct iov_iter iter;
+ struct iovec *iov;
ssize_t ret;

ret = io_prep_rw(req, sqe, force_nonblock);
@@ -2296,29 +2295,30 @@ static int io_write_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe,
return 0;

io = req->io;
- io->rw.iov = io->rw.fast_iov;
req->io = NULL;
- ret = io_import_iovec(WRITE, req, &io->rw.iov, &iter);
+ iov = io_import_iovec(WRITE, req, &io->rw.fast_iov, &iter);
req->io = io;
- if (ret < 0)
- return ret;
+ if (IS_ERR(iov))
+ return PTR_ERR(iov);
+ io->rw.iov = iov;

- io_req_map_rw(req, io->rw.iov, io->rw.fast_iov, &iter);
+ io_req_map_rw(req, io->rw.iov, &io->rw.fast_iov, &iter);
return 0;
}

static int io_write(struct io_kiocb *req, struct io_kiocb **nxt,
bool force_nonblock)
{
- struct iovec inline_vecs[UIO_FASTIOV], *iovec = inline_vecs;
+ struct iovec_cache cache;
+ struct iovec *iovec;
struct kiocb *kiocb = &req->rw.kiocb;
struct iov_iter iter;
size_t iov_count;
ssize_t ret;

- ret = io_import_iovec(WRITE, req, &iovec, &iter);
- if (ret < 0)
- return ret;
+ iovec = io_import_iovec(WRITE, req, &cache, &iter);
+ if (IS_ERR(iovec))
+ return PTR_ERR(iovec);

/* Ensure we clear previously set non-block flag */
if (!force_nonblock)
@@ -2376,7 +2376,7 @@ static int io_write(struct io_kiocb *req, struct io_kiocb **nxt,
kiocb_done(kiocb, ret2, nxt, req->in_async);
} else {
copy_iov:
- ret = io_setup_async_rw(req, iovec, inline_vecs, &iter);
+ ret = io_setup_async_rw(req, iovec, &cache, &iter);
if (ret)
goto out_free;
return -EAGAIN;
@@ -2994,7 +2994,7 @@ static int io_sendmsg_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
#if defined(CONFIG_NET)
struct io_sr_msg *sr = &req->sr_msg;
struct io_async_ctx *io = req->io;
- int ret;
+ struct iovec *iov;

sr->msg_flags = READ_ONCE(sqe->msg_flags);
sr->msg = u64_to_user_ptr(READ_ONCE(sqe->addr));
@@ -3011,12 +3011,14 @@ static int io_sendmsg_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
if (req->flags & REQ_F_NEED_CLEANUP)
return 0;

- io->msg.iov = io->msg.fast_iov;
- ret = sendmsg_copy_msghdr(&io->msg.msg, sr->msg, sr->msg_flags,
- &io->msg.iov);
- if (!ret)
- req->flags |= REQ_F_NEED_CLEANUP;
- return ret;
+ iov = sendmsg_copy_msghdr(&io->msg.msg, sr->msg, sr->msg_flags,
+ &io->msg.fast_iov);
+ if (IS_ERR(iov))
+ return PTR_ERR(iov);
+
+ io->msg.iov = iov;
+ req->flags |= REQ_F_NEED_CLEANUP;
+ return 0;
#else
return -EOPNOTSUPP;
#endif
@@ -3043,19 +3045,21 @@ static int io_sendmsg(struct io_kiocb *req, struct io_kiocb **nxt,
kmsg->msg.msg_name = &req->io->msg.addr;
/* if iov is set, it's allocated already */
if (!kmsg->iov)
- kmsg->iov = kmsg->fast_iov;
+ kmsg->iov = kmsg->fast_iov.iov;
kmsg->msg.msg_iter.iov = kmsg->iov;
} else {
struct io_sr_msg *sr = &req->sr_msg;
+ struct iovec *iov;

kmsg = &io.msg;
kmsg->msg.msg_name = &io.msg.addr;

- io.msg.iov = io.msg.fast_iov;
- ret = sendmsg_copy_msghdr(&io.msg.msg, sr->msg,
- sr->msg_flags, &io.msg.iov);
- if (ret)
- return ret;
+ iov = sendmsg_copy_msghdr(&io.msg.msg, sr->msg,
+ sr->msg_flags, &io.msg.fast_iov);
+ if (IS_ERR(iov))
+ return PTR_ERR(iov);
+
+ io.msg.iov = iov;
}

flags = req->sr_msg.msg_flags;
@@ -3069,7 +3073,7 @@ static int io_sendmsg(struct io_kiocb *req, struct io_kiocb **nxt,
if (req->io)
return -EAGAIN;
if (io_alloc_async_ctx(req)) {
- if (kmsg->iov != kmsg->fast_iov)
+ if (kmsg->iov != kmsg->fast_iov.iov)
kfree(kmsg->iov);
return -ENOMEM;
}
@@ -3081,7 +3085,7 @@ static int io_sendmsg(struct io_kiocb *req, struct io_kiocb **nxt,
ret = -EINTR;
}

- if (kmsg && kmsg->iov != kmsg->fast_iov)
+ if (kmsg && kmsg->iov != kmsg->fast_iov.iov)
kfree(kmsg->iov);
req->flags &= ~REQ_F_NEED_CLEANUP;
io_cqring_add_event(req, ret);
@@ -3151,7 +3155,7 @@ static int io_recvmsg_prep(struct io_kiocb *req,
#if defined(CONFIG_NET)
struct io_sr_msg *sr = &req->sr_msg;
struct io_async_ctx *io = req->io;
- int ret;
+ struct iovec *iov;

sr->msg_flags = READ_ONCE(sqe->msg_flags);
sr->msg = u64_to_user_ptr(READ_ONCE(sqe->addr));
@@ -3168,12 +3172,14 @@ static int io_recvmsg_prep(struct io_kiocb *req,
if (req->flags & REQ_F_NEED_CLEANUP)
return 0;

- io->msg.iov = io->msg.fast_iov;
- ret = recvmsg_copy_msghdr(&io->msg.msg, sr->msg, sr->msg_flags,
- &io->msg.uaddr, &io->msg.iov);
- if (!ret)
- req->flags |= REQ_F_NEED_CLEANUP;
- return ret;
+ iov = recvmsg_copy_msghdr(&io->msg.msg, sr->msg, sr->msg_flags,
+ &io->msg.uaddr, &io->msg.fast_iov);
+ if (IS_ERR(iov))
+ return PTR_ERR(iov);
+
+ io->msg.iov = iov;
+ req->flags |= REQ_F_NEED_CLEANUP;
+ return 0;
#else
return -EOPNOTSUPP;
#endif
@@ -3200,20 +3206,23 @@ static int io_recvmsg(struct io_kiocb *req, struct io_kiocb **nxt,
kmsg->msg.msg_name = &req->io->msg.addr;
/* if iov is set, it's allocated already */
if (!kmsg->iov)
- kmsg->iov = kmsg->fast_iov;
+ kmsg->iov = kmsg->fast_iov.iov;
kmsg->msg.msg_iter.iov = kmsg->iov;
} else {
struct io_sr_msg *sr = &req->sr_msg;
+ struct iovec *iov;

kmsg = &io.msg;
kmsg->msg.msg_name = &io.msg.addr;

- io.msg.iov = io.msg.fast_iov;
- ret = recvmsg_copy_msghdr(&io.msg.msg, sr->msg,
+ io.msg.iov = io.msg.fast_iov.iov;
+ iov = recvmsg_copy_msghdr(&io.msg.msg, sr->msg,
sr->msg_flags, &io.msg.uaddr,
- &io.msg.iov);
- if (ret)
- return ret;
+ &io.msg.fast_iov);
+ if (IS_ERR(iov))
+ return PTR_ERR(iov);
+
+ io.msg.iov = iov;
}

flags = req->sr_msg.msg_flags;
@@ -3228,7 +3237,7 @@ static int io_recvmsg(struct io_kiocb *req, struct io_kiocb **nxt,
if (req->io)
return -EAGAIN;
if (io_alloc_async_ctx(req)) {
- if (kmsg->iov != kmsg->fast_iov)
+ if (kmsg->iov != kmsg->fast_iov.iov)
kfree(kmsg->iov);
return -ENOMEM;
}
@@ -3240,7 +3249,7 @@ static int io_recvmsg(struct io_kiocb *req, struct io_kiocb **nxt,
ret = -EINTR;
}

- if (kmsg && kmsg->iov != kmsg->fast_iov)
+ if (kmsg && kmsg->iov != kmsg->fast_iov.iov)
kfree(kmsg->iov);
req->flags &= ~REQ_F_NEED_CLEANUP;
io_cqring_add_event(req, ret);
@@ -4269,12 +4278,12 @@ static void io_cleanup_req(struct io_kiocb *req)
case IORING_OP_WRITEV:
case IORING_OP_WRITE_FIXED:
case IORING_OP_WRITE:
- if (io->rw.iov != io->rw.fast_iov)
+ if (io->rw.iov != io->rw.fast_iov.iov)
kfree(io->rw.iov);
break;
case IORING_OP_SENDMSG:
case IORING_OP_RECVMSG:
- if (io->msg.iov != io->msg.fast_iov)
+ if (io->msg.iov != io->msg.fast_iov.iov)
kfree(io->msg.iov);
break;
case IORING_OP_OPENAT:
--
1.8.1.2
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)

\
 
 \ /
  Last update: 2020-03-31 15:53    [W:0.052 / U:0.036 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site