All their callers are next to each other; all of them
want the total amount of pages and, possibly, the
offset in the partial final buffer.
Combine into a new helper (pipe_npages()), fix the
bogosity in pipe_space_for_user(), while we are at it.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
We often need to find whether the last buffer is anon or not, and
currently it's rather clumsy:
check if ->iov_offset is non-zero (i.e. that pipe is not empty)
if so, get the corresponding pipe_buffer and check its ->ops
if it's &default_pipe_buf_ops, we have an anon buffer.
Let's replace the use of ->iov_offset (which is nowhere near similar to
its role for other flavours) with signed field (->last_offset), with
the following rules:
empty, no buffers occupied: 0
anon, with bytes up to N-1 filled: N
zero-copy, with bytes up to N-1 filled: -N
That way abs(i->last_offset) is equal to what used to be in i->iov_offset
and empty vs. anon vs. zero-copy can be distinguished by the sign of
i->last_offset.
Checks for "should we extend the last buffer or should we start
a new one?" become easier to follow that way.
Note that most of the operations can only be done in a sane
state - i.e. when the pipe has nothing past the current position of
iterator. About the only thing that could be done outside of that
state is iov_iter_advance(), which transitions to the sane state by
truncating the pipe. There are only two cases where we leave the
sane state:
1) iov_iter_get_pages()/iov_iter_get_pages_alloc(). Will be
dealt with later, when we make get_pages advancing - the callers are
actually happier that way.
2) iov_iter copied, then something is put into the copy. Since
they share the underlying pipe, the original gets behind. When we
decide that we are done with the copy (original is not usable until then)
we advance the original. direct_io used to be done that way; nowadays
it operates on the original and we do iov_iter_revert() to discard
the excessive data. At the moment there's nothing in the kernel that
could do that to ITER_PIPE iterators, so this reason for insane state
is theoretical right now.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Fold pipe_truncate() into it, clean up. We can release buffers
in the same loop where we walk backwards to the iterator beginning
looking for the place where the new position will be.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
instead of setting ->iov_offset for new position and calling
pipe_truncate() to adjust ->len of the last buffer and discard
everything after it, adjust ->len at the same time we set ->iov_offset
and use pipe_discard_from() to deal with buffers past that.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
it's only used to get to the partial buffer we can add to,
and that's always the last one, i.e. pipe->head - 1.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Expand the only remaining call of push_pipe() (in
__pipe_get_pages()), combine it with the page-collecting loop there.
Note that the only reason it's not a loop doing append_pipe() is
that append_pipe() is advancing, while iov_iter_get_pages() is not.
As soon as it switches to saner semantics, this thing will switch
to using append_pipe().
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
New helper: append_pipe(). Extends the last buffer if possible,
allocates a new one otherwise. Returns page and offset in it
on success, NULL on failure. iov_iter is advanced past the
data we've got.
Use that instead of push_pipe() in copy-to-pipe primitives;
they get simpler that way. Handling of short copy (in "mc" one)
is done simply by iov_iter_revert() - iov_iter is in consistent
state after that one, so we can use that.
[Fix for braino caught by Liu Xinpeng <liuxp11@chinatelecom.cn> folded in]
[another braino fix, this time in copy_pipe_to_iter() and pipe_zero();
caught by testcase from Hugh Dickins]
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
There are only two kinds of pipe_buffer in the area used by ITER_PIPE.
1) anonymous - copy_to_iter() et.al. end up creating those and copying
data there. They have zero ->offset, and their ->ops points to
default_pipe_page_ops.
2) zero-copy ones - those come from copy_page_to_iter(), and page
comes from caller. ->offset is also caller-supplied - it might be
non-zero. ->ops points to page_cache_pipe_buf_ops.
Move creation and insertion of those into helpers - push_anon(pipe, size)
and push_page(pipe, page, offset, size) resp., separating them from
the "could we avoid creating a new buffer by merging with the current
head?" logics.
Acked-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
pipe_buffer instances of a pipe are organized as a ring buffer,
with power-of-2 size. Indices are kept *not* reduced modulo ring
size, so the buffer refered to by index N is
pipe->bufs[N & (pipe->ring_size - 1)].
Ring size can change over the lifetime of a pipe, but not while
the pipe is locked. So for any iov_iter primitives it's a constant.
Original conversion of pipes to this layout went overboard trying
to microoptimize that - calculating pipe->ring_size - 1, storing
it in a local variable and using through the function. In some
cases it might be warranted, but most of the times it only
obfuscates what's going on in there.
Introduce a helper (pipe_buf(pipe, N)) that would encapsulate
that and use it in the obvious cases. More will follow...
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Christian Brauner (Microsoft) <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Equivalent of single-segment iovec. Initialized by iov_iter_ubuf(),
checked for by iter_is_ubuf(), otherwise behaves like ITER_IOVEC
ones.
We are going to expose the things like ->write_iter() et.al. to those
in subsequent commits.
New predicate (user_backed_iter()) that is true for ITER_IOVEC and
ITER_UBUF; places like direct-IO handling should use that for
checking that pages we modify after getting them from iov_iter_get_pages()
would need to be dirtied.
DO NOT assume that replacing iter_is_iovec() with user_backed_iter()
will solve all problems - there's code that uses iter_is_iovec() to
decide how to poke around in iov_iter guts and for that the predicate
replacement obviously won't suffice.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
iov_iter work will pretty much overwrite that one, but it's
much harder to backport.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQQqUNBr3gm4hGXdBJlZ7Krx/gZQ6wUCYurHhgAKCRBZ7Krx/gZQ
6+cgAP9XLVFQn0UE1e1JTOitdC/pqjBDMUDQ7SDazrmiVWoMcAEAm+PJSl7GVvMi
fo3fIlaullRT3DjCidAmbbaOOZqkmgA=
=z+N2
-----END PGP SIGNATURE-----
Merge tag 'pull-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Pull copy_to_iter_mc fix from Al Viro:
"Backportable fix for copy_to_iter_mc() - the second part of iov_iter
work will pretty much overwrite this, but would be much harder to
backport"
* tag 'pull-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
fix short copy handling in copy_mc_pipe_to_iter()
One of the goals is to reduce the overhead of using ->read_iter()
and ->write_iter() instead of ->read()/->write(); new_sync_{read,write}()
has a surprising amount of overhead, in particular inside iocb_flags().
That's why the beginning of the series is in this pile; it's not directly
iov_iter-related, but it's a part of the same work...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQQqUNBr3gm4hGXdBJlZ7Krx/gZQ6wUCYurGOQAKCRBZ7Krx/gZQ
6ysyAP91lvBfMRepcxpd9kvtuzWkU8A3rfSziZZteEHANB9Q7QEAiPn2a2OjWkcZ
uAyUWfCkHCNx+dSMkEvUgR5okQ0exAM=
=9UCV
-----END PGP SIGNATURE-----
Merge tag 'pull-work.iov_iter-base' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Pull vfs iov_iter updates from Al Viro:
"Part 1 - isolated cleanups and optimizations.
One of the goals is to reduce the overhead of using ->read_iter() and
->write_iter() instead of ->read()/->write().
new_sync_{read,write}() has a surprising amount of overhead, in
particular inside iocb_flags(). That's the explanation for the
beginning of the series is in this pile; it's not directly
iov_iter-related, but it's a part of the same work..."
* tag 'pull-work.iov_iter-base' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
first_iovec_segment(): just return address
iov_iter: massage calling conventions for first_{iovec,bvec}_segment()
iov_iter: first_{iovec,bvec}_segment() - simplify a bit
iov_iter: lift dealing with maxpages out of first_{iovec,bvec}_segment()
iov_iter_get_pages{,_alloc}(): cap the maxsize with MAX_RW_COUNT
iov_iter_bvec_advance(): don't bother with bvec_iter
copy_page_{to,from}_iter(): switch iovec variants to generic
keep iocb_flags() result cached in struct file
iocb: delay evaluation of IS_SYNC(...) until we want to check IOCB_DSYNC
struct file: use anonymous union member for rcuhead and llist
btrfs: use IOMAP_DIO_NOSYNC
teach iomap_dio_rw() to suppress dsync
No need of likely/unlikely on calls of check_copy_size()
We return length + offset in page via *size. Don't bother - the caller
can do that arithmetics just as well; just report the length to it.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
All callers can and should handle iov_iter_get_pages() returning
fewer pages than requested. All in-kernel ones do. And it makes
the arithmetical overflow analysis much simpler...
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
do what we do for iovec/kvec; that ends up generating better code,
AFAICS.
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Christian Brauner (Microsoft) <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Unlike other copying operations on ITER_PIPE, copy_mc_to_iter() can
result in a short copy. In that case we need to trim the unused
buffers, as well as the length of partially filled one - it's not
enough to set ->head, ->iov_offset and ->count to reflect how
much had we copied. Not hard to fix, fortunately...
I'd put a helper (pipe_discard_from(pipe, head)) into pipe_fs_i.h,
rather than iov_iter.c - it has nothing to do with iov_iter and
having it will allow us to avoid an ugly kludge in fs/splice.c.
We could put it into lib/iov_iter.c for now and move it later,
but I don't see the point going that way...
Cc: stable@kernel.org # 4.19+
Fixes: ca146f6f09 "lib/iov_iter: Fix pipe handling in _copy_to_iter_mcsafe()"
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Christian Brauner (Microsoft) <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
we can do copyin/copyout under kmap_local_page(); it shouldn't overflow
the kmap stack - the maximal footprint increase only by one here.
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Christian Brauner (Microsoft) <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
The existing iov_iter_alignment() function returns the logical OR of
address and length. For cases where address and length need to be
considered separately, introduce a helper function that a caller can
specificy length and address masks that indicate if the iov is
unaligned.
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/20220610195830.3574005-9-kbusch@fb.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Commit 6c77676645 ("iov_iter: Fix iter_xarray_get_pages{,_alloc}()")
introduced a problem on some 32-bit architectures (at least arm, xtensa,
csky,sparc and mips), that have a 'size_t' that is 'unsigned int'.
The reason is that we now do
min(nr * PAGE_SIZE - offset, maxsize);
where 'nr' and 'offset' and both 'unsigned int', and PAGE_SIZE is
'unsigned long'. As a result, the normal C type rules means that the
first argument to 'min()' ends up being 'unsigned long'.
In contrast, 'maxsize' is of type 'size_t'.
Now, 'size_t' and 'unsigned long' are always the same physical type in
the kernel, so you'd think this doesn't matter, and from an actual
arithmetic standpoint it doesn't.
But on 32-bit architectures 'size_t' is commonly 'unsigned int', even if
it could also be 'unsigned long'. In that situation, both are unsigned
32-bit types, but they are not the *same* type.
And as a result 'min()' will complain about the distinct types (ignore
the "pointer types" part of the error message: that's an artifact of the
way we have made 'min()' check types for being the same):
lib/iov_iter.c: In function 'iter_xarray_get_pages':
include/linux/minmax.h:20:35: error: comparison of distinct pointer types lacks a cast [-Werror]
20 | (!!(sizeof((typeof(x) *)1 == (typeof(y) *)1)))
| ^~
lib/iov_iter.c:1464:16: note: in expansion of macro 'min'
1464 | return min(nr * PAGE_SIZE - offset, maxsize);
| ^~~
This was not visible on 64-bit architectures (where we always define
'size_t' to be 'unsigned long').
Force these cases to use 'min_t(size_t, x, y)' to make the type explicit
and avoid the issue.
[ Nit-picky note: technically 'size_t' doesn't have to match 'unsigned
long' arithmetically. We've certainly historically seen environments
with 16-bit address spaces and 32-bit 'unsigned long'.
Similarly, even in 64-bit modern environments, 'size_t' could be its
own type distinct from 'unsigned long', even if it were arithmetically
identical.
So the above type commentary is only really descriptive of the kernel
environment, not some kind of universal truth for the kinds of wild
and crazy situations that are allowed by the C standard ]
Reported-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Link: https://lore.kernel.org/all/YqRyL2sIqQNDfky2@debian/
Cc: Jeff Layton <jlayton@kernel.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The maths at the end of iter_xarray_get_pages() to calculate the actual
size doesn't work under some circumstances, such as when it's been asked to
extract a partial single page. Various terms of the equation cancel out
and you end up with actual == offset. The same issue exists in
iter_xarray_get_pages_alloc().
Fix these to just use min() to select the lesser amount from between the
amount of page content transcribed into the buffer, minus the offset, and
the size limit specified.
This doesn't appear to have caused a problem yet upstream because network
filesystems aren't getting the pages from an xarray iterator, but rather
passing it directly to the socket, which just iterates over it. Cachefiles
*does* do DIO from one to/from ext4/xfs/btrfs/etc. but it always asks for
whole pages to be written or read.
Fixes: 7ff5062079 ("iov_iter: Add ITER_XARRAY")
Reported-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Alexander Viro <viro@zeniv.linux.org.uk>
cc: Dominique Martinet <asmadeus@codewreck.org>
cc: Mike Marshall <hubcap@omnibond.com>
cc: Gao Xiang <xiang@kernel.org>
cc: linux-afs@lists.infradead.org
cc: v9fs-developer@lists.sourceforge.net
cc: devel@lists.orangefs.org
cc: linux-erofs@lists.ozlabs.org
cc: linux-cachefs@redhat.com
cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
The functions copy_page_to_iter_pipe() and push_pipe() can both
allocate a new pipe_buffer, but the "flags" member initializer is
missing.
Fixes: 241699cd72 ("new iov_iter flavour: pipe-backed")
To: Alexander Viro <viro@zeniv.linux.org.uk>
To: linux-fsdevel@vger.kernel.org
To: linux-kernel@vger.kernel.org
Cc: stable@vger.kernel.org
Signed-off-by: Max Kellermann <max.kellermann@ionos.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Take advantage of how kmap_local_folio() works to simplify the loop.
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Introduce a new nofault flag to indicate to iov_iter_get_pages not to
fault in user pages.
This is implemented by passing the FOLL_NOFAULT flag to get_user_pages,
which causes get_user_pages to fail when it would otherwise fault in a
page. We'll use the ->nofault flag to prevent iomap_dio_rw from faulting
in pages when page faults are not allowed.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Introduce a new fault_in_iov_iter_writeable helper for safely faulting
in an iterator for writing. Uses get_user_pages() to fault in the pages
without actually writing to them, which would be destructive.
We'll use fault_in_iov_iter_writeable in gfs2 once we've determined that
the iterator passed to .read_iter isn't in memory.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Turn iov_iter_fault_in_readable into a function that returns the number
of bytes not faulted in, similar to copy_to_user, instead of returning a
non-zero value when any of the requested pages couldn't be faulted in.
This supports the existing users that require all pages to be faulted in
as well as new users that are happy if any pages can be faulted in.
Rename iov_iter_fault_in_readable to fault_in_iov_iter_readable to make
sure this change doesn't silently break things.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Turn fault_in_pages_{readable,writeable} into versions that return the
number of bytes not faulted in, similar to copy_to_user, instead of
returning a non-zero value when any of the requested pages couldn't be
faulted in. This supports the existing users that require all pages to
be faulted in as well as new users that are happy if any pages can be
faulted in.
Rename the functions to fault_in_{readable,writeable} to make sure
this change doesn't silently break things.
Neither of these functions is entirely trivial and it doesn't seem
useful to inline them, so move them to mm/gup.c.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Both iov_iter_get_pages and iov_iter_get_pages_alloc return the number
of bytes of the iovec they could get the pages for. When they cannot
get any pages, they're supposed to return 0, but when the start of the
iovec isn't page aligned, the calculation goes wrong and they return a
negative value. Fix both functions.
In addition, change iov_iter_get_pages_alloc to return NULL in that case
to prevent resource leaks.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
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>
Fix all kernel-doc warnings in lib/iov_iter.c:
lib/iov_iter.c:695: warning: Function parameter or member 'i' not described in '_copy_mc_to_iter'
lib/iov_iter.c:695: warning: Excess function parameter 'iter' description in '_copy_mc_to_iter'
lib/iov_iter.c:695: warning: No description found for return value of '_copy_mc_to_iter'
lib/iov_iter.c:758: warning: Function parameter or member 'i' not described in '_copy_from_iter_flushcache'
lib/iov_iter.c:758: warning: Excess function parameter 'iter' description in '_copy_from_iter_flushcache'
lib/iov_iter.c:758: warning: No description found for return value of '_copy_from_iter_flushcache'
Link: https://lkml.kernel.org/r/20210809051053.6531-1-rdunlap@infradead.org
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This warning was there to catch any architectures that still use
CONFIG_SET_FS, and that would mis-use iov_iter_init() for anything that
wasn't a proper user space pointer. So that
WARN_ON_ONCE(uaccess_kernel());
makes perfect conceptual sense: you really shouldn't use a kernel
pointer with set_fs(KERNEL_DS) and then pass it to iov_iter_init().
HOWEVER.
Guenter Roeck reports that this warning actually triggers in no-mmu
configurations of both ARM and m68k. And the reason isn't that they
pass in a kernel pointer under set_fs(KERNEL_DS) at all: the reason is
that in those configurations, "uaccess_kernel()" is simply not reliable.
Those no-mmu setups set USER_DS and KERNEL_DS to the same values, so you
can't test for the difference.
In particular, the no-mmu case for ARM does
#define USER_DS KERNEL_DS
#define uaccess_kernel() (true)
so USER_DS and KERNEL_DS have the same value, and uaccess_kernel() is
always trivially true.
The m68k case is slightly different and not quite as obvious. It does
(spread out over multiple header files just to be extra exciting:
asm/processor.h, asm/segment.h and asm-generic/uaccess.h):
#define TASK_SIZE (0xFFFFFFFFUL)
#define USER_DS MAKE_MM_SEG(TASK_SIZE)
#define KERNEL_DS MAKE_MM_SEG(~0UL)
#define get_fs() (current_thread_info()->addr_limit)
#define uaccess_kernel() (get_fs().seg == KERNEL_DS.seg)
but the end result is the same: uaccess_kernel() will always be true,
because USER_DS and KERNEL_DS end up having the same value, even if that
value is defined differently.
This is very arguably a misfeature in those implementations, but in the
end we don't really care. All modern architectures have gotten rid of
set_fs() already, and generic kernel code never uses it. And while the
sanity check was a nice idea, an architecture would have to go the extra
mile to actually break this.
So this well-intentioned warning isn't really all that likely to find
anything but these known false positives, and as such just isn't worth
maintaining.
Reported-by: Guenter Roeck <linux@roeck-us.net>
Fixes: 8cd54c1c84 ("iov_iter: separate direction from flavour")
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1) kmap_atomic() is not needed here, kmap_local_page() is enough.
2) No need to make sum = csum_block_add(sum, next, off); conditional
upon next != 0 - adding 0 is a no-op as far as csum_block_add()
is concerned.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
kmap_local_page() is enough there. Moreover, we can use _copy_to_iter()
for actual copying in those cases - no useful extra checks on the
address we are copying from in that call.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
recalculating offset on each iteration is pointless - on all subsequent
passes through the loop it will be zero anyway.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Iterator macros used to provide the arguments for step callbacks in
a structure matching the flavour - iovec for ITER_IOVEC, kvec for
ITER_KVEC and bio_vec for ITER_BVEC. That already broke down for
ITER_XARRAY (bio_vec there); now that we are using kvec callback
for bvec and xarray cases, we are always passing a pointer + length
(void __user * + size_t for ITER_IOVEC callback, void * + size_t
for everything else).
Note that the original reason for bio_vec (page + offset + len) in
case of ITER_BVEC used to be that we did *not* want to kmap a
page when all we wanted was e.g. to find the alignment of its
subrange. Now all such users are gone and the ones that are left
want the page mapped anyway for actually copying the data.
So in all cases we have pointer + length, and there's no good
reason for keeping those in struct iovec or struct kvec - we
can just pass them to callback separately.
Again, less boilerplate in callbacks...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Making iterator macros keep track of the amount of data copied is pretty
easy and it has several benefits:
1) we no longer need the mess like (from += v.iov_len) - v.iov_len
in the callbacks - initial value + total amount copied so far would do
just fine.
2) less obviously, we no longer need to remember the initial amount
of data we wanted to copy; the loops in iterator macros are along the lines
of
wanted = bytes;
while (bytes) {
copy some
bytes -= copied
if short copy
break
}
bytes = wanted - bytes;
Replacement is
offs = 0;
while (bytes) {
copy some
offs += copied
bytes -= copied
if short copy
break
}
bytes = offs;
That wouldn't be a win per se, but unlike the initial value of bytes, the amount
copied so far *is* useful in callbacks.
3) in some cases (csum_and_copy_..._iter()) we already had offs manually
maintained by the callbacks. With that change we can drop that.
Less boilerplate and more readable code...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
After the previous commit we have
* xarray and bvec callbacks idential in all cases
* both equivalent to kvec callback wrapped into
kmap_local_page()/kunmap_local() pair.
So we can pass only two (iovec and kvec) callbacks to
iterate_and_advance() and let iterate_{bvec,xarray} wrap
it into kmap_local_page()/kunmap_local_page().
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
... and now we finally can sort out the mess in _copy_mc_to_iter().
Provide a variant of iterate_and_advance() that does *NOT* ignore
the return values of bvec, xarray and kvec callbacks, use that in
_copy_mc_to_iter(). That gets rid of magic in those callbacks -
we used to need it so we'd get at least the right return value in
case of failure halfway through.
As a bonus, now iterator is advanced by the amount actually copied
for all flavours. That's what the callers expect and it used to do that
correctly in iovec and xarray cases. However, in kvec and bvec cases
the iterator had not been advanced on such failures, breaking the users.
Fixed now...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
... incidentally, using pointer instead of index in an array
(the only change here) trims half-kilobyte of .text...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
The differences between iterate_iovec and iterate_kvec are minor:
* kvec callback is treated as if it returned 0
* initialization of __p is with i->iov and i->kvec resp.
which is trivially dealt with.
No code generation changes - compiler is quite capable of turning
left = ((void)(STEP), 0);
__v.iov_len -= left;
(with no accesses to left downstream) and
(void)(STEP);
into the same code.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Premature optimization is the root of all evil... Trying
to unroll the first pass through the loop makes it harder
to follow and not just for readers - compiler ends up
generating worse code than it would on a "non-optimized"
loop.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
iov_iter_advance() needs to do some non-trivial work when it's given
0 as argument (skip all empty iovecs, mostly). We used to implement
it via iterate_and_advance(); we no longer do so and for all other
users of iterate_and_advance() zero length is a no-op.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Namely, have off counted starting from 0 rather than from csstate->off.
To compensate we need to shift the initial value (csstate->sum) (rotate
by 8 bits, as usual for csum) and do the same after we are finished adding
the pieces up.
What we get out of that is a bit more redundancy in our variables - from
is always equal to addr + off, which will be useful several commits down
the road.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
