[PATCH v2 0/3] Convert vmcore to use an iov_iter

[PATCH v2 0/3] Convert vmcore to use an iov_iter

[Matthew Wilcox (Oracle)]

For some reason several people have been sending bad patches to fix
compiler warnings in vmcore recently.  Here's how it should be done.
Compile-tested only on x86.  As noted in the first patch, s390 should
take this conversion a bit further, but I'm not inclined to do that
work myself.

v2:
 - Removed unnecessary kernel-doc
 - Included uio.h to fix compilation problems
 - Made read_from_oldmem_iter static to avoid compile warnings during the
   conversion
 - Use iov_iter_truncate() (Christoph)

Matthew Wilcox (Oracle) (3):
  vmcore: Convert __read_vmcore to use an iov_iter
  vmcore: Convert read_from_oldmem() to take an iov_iter
  iov-kaddr

 arch/x86/kernel/crash_dump_64.c |   7 +-
 fs/9p/vfs_dir.c                 |   5 +-
 fs/9p/xattr.c                   |   6 +-
 fs/proc/vmcore.c                | 119 ++++++++++++--------------------
 include/linux/crash_dump.h      |  10 ++-
 include/linux/uio.h             |   9 +++
 lib/iov_iter.c                  |  32 +++++++++
 7 files changed, 97 insertions(+), 91 deletions(-)

-- 
2.33.0

[PATCH v2 1/3] vmcore: Convert __read_vmcore to use an iov_iter

[Matthew Wilcox (Oracle)]

This gets rid of copy_to() and let us use proc_read_iter() instead
of proc_read().

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 fs/proc/vmcore.c | 81 +++++++++++++++++-------------------------------
 1 file changed, 29 insertions(+), 52 deletions(-)

diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 958cad6476e6..7b25f568d20d 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -252,22 +252,8 @@ ssize_t __weak copy_oldmem_page_encrypted(struct iov_iter *iter,
 	return copy_oldmem_page(iter, pfn, csize, offset);
 }
 
-/*
- * Copy to either kernel or user space
- */
-static int copy_to(void *target, void *src, size_t size, int userbuf)
-{
-	if (userbuf) {
-		if (copy_to_user((char __user *) target, src, size))
-			return -EFAULT;
-	} else {
-		memcpy(target, src, size);
-	}
-	return 0;
-}
-
 #ifdef CONFIG_PROC_VMCORE_DEVICE_DUMP
-static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
+static int vmcoredd_copy_dumps(struct iov_iter *iter, u64 start, size_t size)
 {
 	struct vmcoredd_node *dump;
 	u64 offset = 0;
@@ -280,14 +266,13 @@ static int vmcoredd_copy_dumps(void *dst, u64 start, size_t size, int userbuf)
 		if (start < offset + dump->size) {
 			tsz = min(offset + (u64)dump->size - start, (u64)size);
 			buf = dump->buf + start - offset;
-			if (copy_to(dst, buf, tsz, userbuf)) {
+			if (copy_to_iter(buf, tsz, iter) < tsz) {
 				ret = -EFAULT;
 				goto out_unlock;
 			}
 
 			size -= tsz;
 			start += tsz;
-			dst += tsz;
 
 			/* Leave now if buffer filled already */
 			if (!size)
@@ -343,33 +328,28 @@ static int vmcoredd_mmap_dumps(struct vm_area_struct *vma, unsigned long dst,
 /* Read from the ELF header and then the crash dump. On error, negative value is
  * returned otherwise number of bytes read are returned.
  */
-static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
-			     int userbuf)
+static ssize_t __read_vmcore(struct iov_iter *iter, loff_t *fpos)
 {
 	ssize_t acc = 0, tmp;
 	size_t tsz;
 	u64 start;
 	struct vmcore *m = NULL;
 
-	if (buflen == 0 || *fpos >= vmcore_size)
+	if (iter->count == 0 || *fpos >= vmcore_size)
 		return 0;
 
-	/* trim buflen to not go beyond EOF */
-	if (buflen > vmcore_size - *fpos)
-		buflen = vmcore_size - *fpos;
+	iov_iter_truncate(iter, vmcore_size - *fpos);
 
 	/* Read ELF core header */
 	if (*fpos < elfcorebuf_sz) {
-		tsz = min(elfcorebuf_sz - (size_t)*fpos, buflen);
-		if (copy_to(buffer, elfcorebuf + *fpos, tsz, userbuf))
+		tsz = min(elfcorebuf_sz - (size_t)*fpos, iter->count);
+		if (copy_to_iter(elfcorebuf + *fpos, tsz, iter) < tsz)
 			return -EFAULT;
-		buflen -= tsz;
 		*fpos += tsz;
-		buffer += tsz;
 		acc += tsz;
 
 		/* leave now if filled buffer already */
-		if (buflen == 0)
+		if (iter->count == 0)
 			return acc;
 	}
 
@@ -390,35 +370,31 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
 		/* Read device dumps */
 		if (*fpos < elfcorebuf_sz + vmcoredd_orig_sz) {
 			tsz = min(elfcorebuf_sz + vmcoredd_orig_sz -
-				  (size_t)*fpos, buflen);
+				  (size_t)*fpos, iter->count);
 			start = *fpos - elfcorebuf_sz;
-			if (vmcoredd_copy_dumps(buffer, start, tsz, userbuf))
+			if (vmcoredd_copy_dumps(iter, start, tsz))
 				return -EFAULT;
 
-			buflen -= tsz;
 			*fpos += tsz;
-			buffer += tsz;
 			acc += tsz;
 
 			/* leave now if filled buffer already */
-			if (!buflen)
+			if (!iter->count)
 				return acc;
 		}
 #endif /* CONFIG_PROC_VMCORE_DEVICE_DUMP */
 
 		/* Read remaining elf notes */
-		tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, buflen);
+		tsz = min(elfcorebuf_sz + elfnotes_sz - (size_t)*fpos, iter->count);
 		kaddr = elfnotes_buf + *fpos - elfcorebuf_sz - vmcoredd_orig_sz;
-		if (copy_to(buffer, kaddr, tsz, userbuf))
+		if (copy_to_iter(kaddr, tsz, iter) < tsz)
 			return -EFAULT;
 
-		buflen -= tsz;
 		*fpos += tsz;
-		buffer += tsz;
 		acc += tsz;
 
 		/* leave now if filled buffer already */
-		if (buflen == 0)
+		if (iter->count == 0)
 			return acc;
 	}
 
@@ -426,19 +402,17 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
 		if (*fpos < m->offset + m->size) {
 			tsz = (size_t)min_t(unsigned long long,
 					    m->offset + m->size - *fpos,
-					    buflen);
+					    iter->count);
 			start = m->paddr + *fpos - m->offset;
-			tmp = read_from_oldmem(buffer, tsz, &start,
-					       userbuf, cc_platform_has(CC_ATTR_MEM_ENCRYPT));
+			tmp = read_from_oldmem_iter(iter, tsz, &start,
+					cc_platform_has(CC_ATTR_MEM_ENCRYPT));
 			if (tmp < 0)
 				return tmp;
-			buflen -= tsz;
 			*fpos += tsz;
-			buffer += tsz;
 			acc += tsz;
 
 			/* leave now if filled buffer already */
-			if (buflen == 0)
+			if (iter->count == 0)
 				return acc;
 		}
 	}
@@ -446,15 +420,14 @@ static ssize_t __read_vmcore(char *buffer, size_t buflen, loff_t *fpos,
 	return acc;
 }
 
-static ssize_t read_vmcore(struct file *file, char __user *buffer,
-			   size_t buflen, loff_t *fpos)
+static ssize_t read_vmcore(struct kiocb *iocb, struct iov_iter *iter)
 {
-	return __read_vmcore((__force char *) buffer, buflen, fpos, 1);
+	return __read_vmcore(iter, &iocb->ki_pos);
 }
 
 /*
  * The vmcore fault handler uses the page cache and fills data using the
- * standard __vmcore_read() function.
+ * standard __read_vmcore() function.
  *
  * On s390 the fault handler is used for memory regions that can't be mapped
  * directly with remap_pfn_range().
@@ -464,9 +437,10 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
 #ifdef CONFIG_S390
 	struct address_space *mapping = vmf->vma->vm_file->f_mapping;
 	pgoff_t index = vmf->pgoff;
+	struct iov_iter iter;
+	struct kvec kvec;
 	struct page *page;
 	loff_t offset;
-	char *buf;
 	int rc;
 
 	page = find_or_create_page(mapping, index, GFP_KERNEL);
@@ -474,8 +448,11 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf)
 		return VM_FAULT_OOM;
 	if (!PageUptodate(page)) {
 		offset = (loff_t) index << PAGE_SHIFT;
-		buf = __va((page_to_pfn(page) << PAGE_SHIFT));
-		rc = __read_vmcore(buf, PAGE_SIZE, &offset, 0);
+		kvec.iov_base = page_address(page);
+		kvec.iov_len = PAGE_SIZE;
+		iov_iter_kvec(&iter, READ, &kvec, 1, PAGE_SIZE);
+
+		rc = __read_vmcore(&iter, &offset);
 		if (rc < 0) {
 			unlock_page(page);
 			put_page(page);
@@ -725,7 +702,7 @@ static int mmap_vmcore(struct file *file, struct vm_area_struct *vma)
 
 static const struct proc_ops vmcore_proc_ops = {
 	.proc_open	= open_vmcore,
-	.proc_read	= read_vmcore,
+	.proc_read_iter	= read_vmcore,
 	.proc_lseek	= default_llseek,
 	.proc_mmap	= mmap_vmcore,
 };
-- 
2.33.0

[PATCH v2 2/3] vmcore: Convert read_from_oldmem() to take an iov_iter

[Matthew Wilcox (Oracle)]

Remove the read_from_oldmem() wrapper introduced earlier and convert
all the remaining callers to pass an iov_iter.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 arch/x86/kernel/crash_dump_64.c |  7 +++++-
 fs/proc/vmcore.c                | 40 +++++++++++++--------------------
 include/linux/crash_dump.h      | 10 ++++-----
 3 files changed, 25 insertions(+), 32 deletions(-)

diff --git a/arch/x86/kernel/crash_dump_64.c b/arch/x86/kernel/crash_dump_64.c
index f922d51c9d1f..0fa87648e55c 100644
--- a/arch/x86/kernel/crash_dump_64.c
+++ b/arch/x86/kernel/crash_dump_64.c
@@ -55,6 +55,11 @@ ssize_t copy_oldmem_page_encrypted(struct iov_iter *iter, unsigned long pfn,
 
 ssize_t elfcorehdr_read(char *buf, size_t count, u64 *ppos)
 {
-	return read_from_oldmem(buf, count, ppos, 0,
+	struct kvec kvec = { .iov_base = buf, .iov_len = count };
+	struct iov_iter iter;
+
+	iov_iter_kvec(&iter, READ, &kvec, 1, count);
+
+	return read_from_oldmem(&iter, count, ppos,
 				cc_platform_has(CC_ATTR_GUEST_MEM_ENCRYPT));
 }
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 7b25f568d20d..bee05b40196b 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -133,7 +133,7 @@ static int open_vmcore(struct inode *inode, struct file *file)
 }
 
 /* Reads a page from the oldmem device from given offset. */
-static ssize_t read_from_oldmem_iter(struct iov_iter *iter, size_t count,
+ssize_t read_from_oldmem(struct iov_iter *iter, size_t count,
 			 u64 *ppos, bool encrypted)
 {
 	unsigned long pfn, offset;
@@ -181,27 +181,6 @@ static ssize_t read_from_oldmem_iter(struct iov_iter *iter, size_t count,
 	return read;
 }
 
-ssize_t read_from_oldmem(char *buf, size_t count,
-			 u64 *ppos, int userbuf,
-			 bool encrypted)
-{
-	struct iov_iter iter;
-	struct iovec iov;
-	struct kvec kvec;
-
-	if (userbuf) {
-		iov.iov_base = (__force void __user *)buf;
-		iov.iov_len = count;
-		iov_iter_init(&iter, READ, &iov, 1, count);
-	} else {
-		kvec.iov_base = buf;
-		kvec.iov_len = count;
-		iov_iter_kvec(&iter, READ, &kvec, 1, count);
-	}
-
-	return read_from_oldmem_iter(&iter, count, ppos, encrypted);
-}
-
 /*
  * Architectures may override this function to allocate ELF header in 2nd kernel
  */
@@ -221,7 +200,12 @@ void __weak elfcorehdr_free(unsigned long long addr)
  */
 ssize_t __weak elfcorehdr_read(char *buf, size_t count, u64 *ppos)
 {
-	return read_from_oldmem(buf, count, ppos, 0, false);
+	struct kvec kvec = { .iov_base = buf, .iov_len = count };
+	struct iov_iter iter;
+
+	iov_iter_kvec(&iter, READ, &kvec, 1, count);
+
+	return read_from_oldmem(&iter, count, ppos, false);
 }
 
 /*
@@ -229,7 +213,13 @@ ssize_t __weak elfcorehdr_read(char *buf, size_t count, u64 *ppos)
  */
 ssize_t __weak elfcorehdr_read_notes(char *buf, size_t count, u64 *ppos)
 {
-	return read_from_oldmem(buf, count, ppos, 0, cc_platform_has(CC_ATTR_MEM_ENCRYPT));
+	struct kvec kvec = { .iov_base = buf, .iov_len = count };
+	struct iov_iter iter;
+
+	iov_iter_kvec(&iter, READ, &kvec, 1, count);
+
+	return read_from_oldmem(&iter, count, ppos,
+			cc_platform_has(CC_ATTR_MEM_ENCRYPT));
 }
 
 /*
@@ -404,7 +394,7 @@ static ssize_t __read_vmcore(struct iov_iter *iter, loff_t *fpos)
 					    m->offset + m->size - *fpos,
 					    iter->count);
 			start = m->paddr + *fpos - m->offset;
-			tmp = read_from_oldmem_iter(iter, tsz, &start,
+			tmp = read_from_oldmem(iter, tsz, &start,
 					cc_platform_has(CC_ATTR_MEM_ENCRYPT));
 			if (tmp < 0)
 				return tmp;
diff --git a/include/linux/crash_dump.h b/include/linux/crash_dump.h
index a1cf7d5c03c7..0f3a656293b0 100644
--- a/include/linux/crash_dump.h
+++ b/include/linux/crash_dump.h
@@ -134,13 +134,11 @@ static inline int vmcore_add_device_dump(struct vmcoredd_data *data)
 #endif /* CONFIG_PROC_VMCORE_DEVICE_DUMP */
 
 #ifdef CONFIG_PROC_VMCORE
-ssize_t read_from_oldmem(char *buf, size_t count,
-			 u64 *ppos, int userbuf,
-			 bool encrypted);
+ssize_t read_from_oldmem(struct iov_iter *iter, size_t count,
+			 u64 *ppos, bool encrypted);
 #else
-static inline ssize_t read_from_oldmem(char *buf, size_t count,
-				       u64 *ppos, int userbuf,
-				       bool encrypted)
+static inline ssize_t read_from_oldmem(struct iov_iter *iter, size_t count,
+				       u64 *ppos, bool encrypted)
 {
 	return -EOPNOTSUPP;
 }
-- 
2.33.0

[PATCH v2 3/3] iov-kaddr

[Matthew Wilcox (Oracle)]

---
 fs/9p/vfs_dir.c     |  5 +----
 fs/9p/xattr.c       |  6 ++----
 include/linux/uio.h |  9 +++++++++
 lib/iov_iter.c      | 32 ++++++++++++++++++++++++++++++++
 4 files changed, 44 insertions(+), 8 deletions(-)

diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c
index 8c854d8cb0cd..cad6c24f9f0d 100644
--- a/fs/9p/vfs_dir.c
+++ b/fs/9p/vfs_dir.c
@@ -90,7 +90,6 @@ static int v9fs_dir_readdir(struct file *file, struct dir_context *ctx)
 	struct p9_fid *fid;
 	int buflen;
 	struct p9_rdir *rdir;
-	struct kvec kvec;
 
 	p9_debug(P9_DEBUG_VFS, "name %pD\n", file);
 	fid = file->private_data;
@@ -100,15 +99,13 @@ static int v9fs_dir_readdir(struct file *file, struct dir_context *ctx)
 	rdir = v9fs_alloc_rdir_buf(file, buflen);
 	if (!rdir)
 		return -ENOMEM;
-	kvec.iov_base = rdir->buf;
-	kvec.iov_len = buflen;
 
 	while (1) {
 		if (rdir->tail == rdir->head) {
 			struct iov_iter to;
 			int n;
 
-			iov_iter_kvec(&to, READ, &kvec, 1, buflen);
+			iov_iter_kaddr(&to, READ, rdir->buf, buflen);
 			n = p9_client_read(file->private_data, ctx->pos, &to,
 					   &err);
 			if (err)
diff --git a/fs/9p/xattr.c b/fs/9p/xattr.c
index a824441b95a2..6fb0e7640605 100644
--- a/fs/9p/xattr.c
+++ b/fs/9p/xattr.c
@@ -20,11 +20,10 @@ ssize_t v9fs_fid_xattr_get(struct p9_fid *fid, const char *name,
 	ssize_t retval;
 	u64 attr_size;
 	struct p9_fid *attr_fid;
-	struct kvec kvec = {.iov_base = buffer, .iov_len = buffer_size};
 	struct iov_iter to;
 	int err;
 
-	iov_iter_kvec(&to, READ, &kvec, 1, buffer_size);
+	iov_iter_kaddr(&to, READ, buffer, buffer_size);
 
 	attr_fid = p9_client_xattrwalk(fid, name, &attr_size);
 	if (IS_ERR(attr_fid)) {
@@ -105,11 +104,10 @@ int v9fs_xattr_set(struct dentry *dentry, const char *name,
 int v9fs_fid_xattr_set(struct p9_fid *fid, const char *name,
 		   const void *value, size_t value_len, int flags)
 {
-	struct kvec kvec = {.iov_base = (void *)value, .iov_len = value_len};
 	struct iov_iter from;
 	int retval, err;
 
-	iov_iter_kvec(&from, WRITE, &kvec, 1, value_len);
+	iov_iter_kaddr(&from, WRITE, (void *)value, value_len);
 
 	p9_debug(P9_DEBUG_VFS, "name = %s value_len = %zu flags = %d\n",
 		 name, value_len, flags);
diff --git a/include/linux/uio.h b/include/linux/uio.h
index 6350354f97e9..cedbc1ca5c15 100644
--- a/include/linux/uio.h
+++ b/include/linux/uio.h
@@ -23,6 +23,7 @@ enum iter_type {
 	ITER_KVEC,
 	ITER_BVEC,
 	ITER_PIPE,
+	ITER_KADDR,
 	ITER_XARRAY,
 	ITER_DISCARD,
 };
@@ -43,6 +44,7 @@ struct iov_iter {
 		const struct iovec *iov;
 		const struct kvec *kvec;
 		const struct bio_vec *bvec;
+		void *kaddr;
 		struct xarray *xarray;
 		struct pipe_inode_info *pipe;
 	};
@@ -79,6 +81,11 @@ static inline bool iov_iter_is_kvec(const struct iov_iter *i)
 	return iov_iter_type(i) == ITER_KVEC;
 }
 
+static inline bool iov_iter_is_kaddr(const struct iov_iter *i)
+{
+	return iov_iter_type(i) == ITER_KVEC;
+}
+
 static inline bool iov_iter_is_bvec(const struct iov_iter *i)
 {
 	return iov_iter_type(i) == ITER_BVEC;
@@ -236,6 +243,8 @@ void iov_iter_init(struct iov_iter *i, unsigned int direction, const struct iove
 			unsigned long nr_segs, size_t count);
 void iov_iter_kvec(struct iov_iter *i, unsigned int direction, const struct kvec *kvec,
 			unsigned long nr_segs, size_t count);
+void iov_iter_kaddr(struct iov_iter *i, unsigned int direction, void *kaddr,
+			size_t count);
 void iov_iter_bvec(struct iov_iter *i, unsigned int direction, const struct bio_vec *bvec,
 			unsigned long nr_segs, size_t count);
 void iov_iter_pipe(struct iov_iter *i, unsigned int direction, struct pipe_inode_info *pipe,
diff --git a/lib/iov_iter.c b/lib/iov_iter.c
index 66a740e6e153..64e6bc33176a 100644
--- a/lib/iov_iter.c
+++ b/lib/iov_iter.c
@@ -38,6 +38,22 @@
 	n = off;						\
 }
 
+#define iterate_kaddr(i, n, base, len, off, STEP) {		\
+	void *base;						\
+	size_t len;						\
+	size_t off = 0;						\
+	size_t skip = i->iov_offset;				\
+	len = min(n, i->count - skip);				\
+	if (likely(len)) {					\
+		base = i->kaddr + skip;				\
+		len -= (STEP);					\
+		off += len;					\
+		skip += len;					\
+	}							\
+	i->iov_offset = skip;					\
+	n = off;						\
+}
+
 #define iterate_bvec(i, n, base, len, off, p, STEP) {		\
 	size_t off = 0;						\
 	unsigned skip = i->iov_offset;				\
@@ -136,6 +152,8 @@ __out:								\
 						kvec, (K))	\
 			i->nr_segs -= kvec - i->kvec;		\
 			i->kvec = kvec;				\
+		} else if (iov_iter_is_kaddr(i)) {		\
+			iterate_kaddr(i, n, base, len, off, (K)) \
 		} else if (iov_iter_is_xarray(i)) {		\
 			void *base;				\
 			size_t len;				\
@@ -1185,6 +1203,20 @@ void iov_iter_kvec(struct iov_iter *i, unsigned int direction,
 }
 EXPORT_SYMBOL(iov_iter_kvec);
 
+void iov_iter_kaddr(struct iov_iter *i, unsigned int direction,
+			void *kaddr, size_t count)
+{
+	WARN_ON(direction & ~(READ | WRITE));
+	*i = (struct iov_iter){
+		.iter_type = ITER_KADDR,
+		.data_source = direction,
+		.kaddr = kaddr,
+		.iov_offset = 0,
+		.count = count
+	};
+}
+EXPORT_SYMBOL(iov_iter_kaddr);
+
 void iov_iter_bvec(struct iov_iter *i, unsigned int direction,
 			const struct bio_vec *bvec, unsigned long nr_segs,
 			size_t count)
-- 
2.33.0

Re: [PATCH v2 3/3] iov-kaddr

[Matthew Wilcox]

On Mon, Dec 13, 2021 at 02:19:07PM +0000, Matthew Wilcox (Oracle) wrote:
> ---
>  fs/9p/vfs_dir.c     |  5 +----
>  fs/9p/xattr.c       |  6 ++----
>  include/linux/uio.h |  9 +++++++++
>  lib/iov_iter.c      | 32 ++++++++++++++++++++++++++++++++
>  4 files changed, 44 insertions(+), 8 deletions(-)

Ugh, this was NOT supposed to go out.  Sorry, I'll do a v3.