search for: is_zero

...The clever approach here was suggested by Eric Blake. See: + * https://www.redhat.com/archives/libguestfs/2017-April/msg00171.html + * https://rusty.ozlabs.org/?p=560 + * + * See also: + * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69908 + */ +static inline bool __attribute__((__nonnull__ (1))) +is_zero (const char *buffer, size_t size) +{ + size_t i; + const size_t limit = size < 16 ? size : 16; + + for (i = 0; i < limit; ++i) + if (buffer[i]) + return false; + if (size != limit) + return ! memcmp (buffer, buffer + 16, size - 16); + + return true; +} + +#endif /* NBDKIT_ISZE...

v1 -> v2: The first patch is the same (the pure refactoring), but in the second patch I implement Eric Blake's suggested version. Rich.

...enamev, int fd) return combined_index; } -/* Return true iff the buffer is all zero bytes. - * - * Note that gcc is smart enough to optimize this properly: - * http://stackoverflow.com/questions/1493936/faster-means-of-checking-for-an-empty-buffer-in-c/1493989#1493989 - */ -static inline int -is_zero (const unsigned char *buffer, size_t size) -{ - size_t i; - - for (i = 0; i < size; ++i) { - if (buffer[i] != 0) - return 0; - } - - return 1; -} - struct global_state { /* Current iterator. Threads update this, but it is protected by a * mutex, and each thread takes a copy...

When sparse_array_zero() is used for a range larger than a page, there's no need to waste time in memset() or is_zero() - we already know the page will be free()d. Signed-off-by: Eric Blake <eblake@redhat.com> --- Here's a fun one :) common/sparse/sparse.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/common/sparse/sparse.c b/common/sparse/sparse.c index cb44743c..5e08576...

...ue iff the buffer is all zero bytes. > > + * > > + * Note that gcc is smart enough to optimize this properly: > > + * http://stackoverflow.com/questions/1493936/faster-means-of-checking-for-an-empty-buffer-in-c/1493989#1493989 > > + */ > > +static inline int > > +is_zero (const char *buffer, size_t size) > > +{ > > + size_t i; > > + > > + for (i = 0; i < size; ++i) { > > + if (buffer[i] != 0) > > + return 0; > > + } > > + > > + return 1; > > +} > > This is still byte-at-a-time. Often...

.... > --- > +/* Return true iff the buffer is all zero bytes. > + * > + * Note that gcc is smart enough to optimize this properly: > + * http://stackoverflow.com/questions/1493936/faster-means-of-checking-for-an-empty-buffer-in-c/1493989#1493989 > + */ > +static inline int > +is_zero (const char *buffer, size_t size) > +{ > + size_t i; > + > + for (i = 0; i < size; ++i) { > + if (buffer[i] != 0) > + return 0; > + } > + > + return 1; > +} This is still byte-at-a-time. Often, you can get even faster results by exploiting libc's o...

On Wednesday, 19 April 2017 18:46:37 CEST Richard W.M. Jones wrote: > v1 -> v2: > > The first patch is the same (the pure refactoring), but in the second > patch I implement Eric Blake's suggested version. LGTM -- can you please add Eric's text to the comment of is_zero, so it is easier to remember why the 16-bytes trick is done? Thanks, -- Pino Toscano

On 1/30/20 8:05 AM, Richard W.M. Jones wrote: > Fast testing whether the AIO buffer (or regions within it) contain all > zeroes, which allows Python code to quickly do sparsification when > copying. > > This includes the iszero.h header from nbdkit which is distributed > under a compatible license. > --- > common/include/Makefile.am | 5 +-- >

On Thu, Jan 30, 2020 at 02:29:21PM +0000, Richard W.M. Jones wrote: >On Thu, Jan 30, 2020 at 08:16:30AM -0600, Eric Blake wrote: >> On 1/30/20 8:05 AM, Richard W.M. Jones wrote: >> >Fast testing whether the AIO buffer (or regions within it) contain all >> >zeroes, which allows Python code to quickly do sparsification when >> >copying. >> > >>

On Thu, Jan 30, 2020 at 04:16:09PM +0100, Martin Kletzander wrote: > I know this probably got answered somewhere, but I've never gotten > any when asking myself. So let me use this opportunity. > > Because I really despise useless processes and duplicated > information I always hated the way all the "mandatory" file headers. > As far as I understand it's

On Wed, Aug 14, 2019 at 09:10:15PM -0500, Eric Blake wrote: > When sparse_array_zero() is used for a range larger than a page, > there's no need to waste time in memset() or is_zero() - we already > know the page will be free()d. > > Signed-off-by: Eric Blake <eblake@redhat.com> > --- > > Here's a fun one :) > > common/sparse/sparse.c | 7 +++++-- > 1 file changed, 5 insertions(+), 2 deletions(-) > > diff --git a/common/sparse/sp...

On Thu, Jan 30, 2020 at 08:16:30AM -0600, Eric Blake wrote: > On 1/30/20 8:05 AM, Richard W.M. Jones wrote: > >Fast testing whether the AIO buffer (or regions within it) contain all > >zeroes, which allows Python code to quickly do sparsification when > >copying. > > > >This includes the iszero.h header from nbdkit which is distributed > >under a compatible

On 2/17/20 9:13 AM, Max Reitz wrote: > Hi, > > It’s my understanding that without some is_zero infrastructure for QEMU, > it’s impossible to implement this flag in qemu’s NBD server. You're right that we may need some more infrastructure before being able to decide when to report this bit in all cases. But for raw files, that infrastructure already exists: does block_status at off...

Hi, It’s my understanding that without some is_zero infrastructure for QEMU, it’s impossible to implement this flag in qemu’s NBD server. At the same time, I still haven’t understood what we need the flag for. As far as I understood in our discussion on your qemu series, there is no case where anyone would need to know whether an image is zero. A...

..._up - 1'; I think we favor the spaces everywhere... > + if (round_down > 0) > + size &= ~(round_down-1); ...in which case the spacing here is also suspect. > + if (count > 0) { > + /* The caller must be writing zeroes, else it's an error. */ > + if (!is_zero (buf, count)) { > + nbdkit_error ("truncate: write beyond end of underlying device"); > + *err = EIO; Would ENOSPC be better here? -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3266 Virtualization: qemu.org | libvirt.org

...const char *fmt, ...) extern void cleanup_free (void *ptr); extern void cleanup_free_string_list (void *ptr); extern void cleanup_unlink_free (void *ptr); +extern void cleanup_close (void *ptr); /*-- in names.c (auto-generated) --*/ extern const char *function_names[]; @@ -405,10 +406,12 @@ is_zero (const char *buffer, size_t size) #define CLEANUP_FREE_STRING_LIST \ __attribute__((cleanup(cleanup_free_string_list))) #define CLEANUP_UNLINK_FREE __attribute__((cleanup(cleanup_unlink_free))) +#define CLEANUP_CLOSE __attribute__((cleanup(cleanup_close))) #else #def...

...touch"; "/copyff2/dest"]; ["truncate_size"; "/copyff2/dest"; string_of_int size]; - ["copy_file_to_file"; "/copyff2/src"; "/copyff2/dest"; ""; ""; ""; "true"]; - ["is_zero"; "/copyff2/dest"]]), [] + ["copy_file_to_file"; "/copyff2/src"; "/copyff2/dest"; ""; ""; ""; "true"; "false"]; + ["is_zero"; "/copyff2/dest"]]), []; + InitScratch...

...I know in earlier review I asked whether we can move the clamping... > + > + /* Work out the type of this extent. */ > + if (p == NULL) > + /* No backing page, so it's a hole. */ > + type = NBDKIT_EXTENT_HOLE | NBDKIT_EXTENT_ZERO; > + else { > + if (is_zero (p, n)) > + /* A backing page and it's all zero, it's a zero extent. */ > + type = NBDKIT_EXTENT_ZERO; > + else > + /* Normal allocated data. */ > + type = 0; > + } > + if (nbdkit_add_extent (extents, offset, n, type) == -1) > +...

...uot;/copyff2/dest"]; > ["truncate_size"; "/copyff2/dest"; string_of_int size]; > - ["copy_file_to_file"; "/copyff2/src"; "/copyff2/dest"; ""; ""; ""; "true"]; > - ["is_zero"; "/copyff2/dest"]]), [] > + ["copy_file_to_file"; "/copyff2/src"; "/copyff2/dest"; ""; ""; ""; "true"; "false"]; > + ["is_zero"; "/copyff2/dest"]]), []; > +...

...we can move the clamping... > > > + > > + /* Work out the type of this extent. */ > > + if (p == NULL) > > + /* No backing page, so it's a hole. */ > > + type = NBDKIT_EXTENT_HOLE | NBDKIT_EXTENT_ZERO; > > + else { > > + if (is_zero (p, n)) > > + /* A backing page and it's all zero, it's a zero extent. */ > > + type = NBDKIT_EXTENT_ZERO; > > + else > > + /* Normal allocated data. */ > > + type = 0; > > + } > > + if (nbdkit_add_extent (exte...