dax_get_props - get properties
cc [ flag... ] file... -ldax [ library...] #include <dax.h> dax_status_t dax_get_props(dax_context_t *ctx, dax_props_t *props);
This function returns in props the properties of DAX for the API version used to create the ctx.
The members of props are:
Maximum number of bits in a bit-sized element. DAX supports any number of bits from 1 to max_elem_bits.
Maximum number of bytes in a fixed-width byte-sized element such that DAX supports any number of bytes from 1 to max_elem_bytes.
Maximum number of bytes in a fixed-width byte-sized element for power-of-2 sizes. DAX supports sizes of 2i where i can have any value from 0 to max_log_elem_bytes, inclusive. In some versions of the libdax library, 2max_log_elem_bytes may be larger than max_elem_bytes.
Maximum number of bytes in a variable-width byte-sized element. DAX supports any number of bytes from 1 to max_var_elem_bytes.
Maximum number of bits in the length element of a run-length vector. DAX supports any number of bits from 1 to max_rle_elem_bits by powers of 2.
Maximum number of bits in the width element of a widths vector. DAX supports any number of bits from 1 to max_width_elem_bits by powers of 2.
Maximum number of significant bytes in dax_int_t, used as the scan value and scan range comparison value(s).
Required alignment in bytes for the destination buffer in all DAX operations.
Required length in bytes for rounding and additional padding in all DAX operations. For more information, see the dax_vec_t(3DAX) man page.
Required alignment for a dax_zip_create_contig() codec.
The optimal alignment for a codec in the dax_zip_create_contig() function. This may be larger than zip_align.
Required alignment for a dax_translate() bitmap.
The optimal alignment for a bitmap in the dax_translate() function. This may be larger than trans_bitmap_align.
Required alignment in bytes of byte-sized input elements whose width is greater than max_elem_bytes.
Maximum number of bits in a dax_translate() bitmap that DAX supports natively. DAX can support more bits but with lower efficiency.
Maximum number of symbols in a zip codec.
The maximum buffer size for the dax_zip_get_contig() function.
If the sum of zip symbol lengths is constrained to 14 bytes as described in dax_zip_create(), this is 1, else 0.
If the codec in dax_zip_create_contig() is passed directly to DAX without conversion, this is 1, else 0.
When an index written to the destination buffer for the ONES_INDEX flag is too large and the written value saturates at the maximum value, this is 1. If DAX generates an error for indices that are too large, this is 0.
If the dax_post() functions support the DAX_PIPE_SRC, DAX_PIPE_AUX, and DAX_PIPE_MASK flags, this is 1, else 0.
Maximum number of src elements in a single DAX hardware command. For larger vectors, libdax breaks the vector into multiple DAX commands.
Maximum number of dst vector bytes in a single DAX hardware command. For larger vectors, libdax breaks the vector into multiple DAX commands.
A mask of the dax_set_debug() options supported by this version.
A mask of the dax_set_log() options supported by this version.
The values of the properties for each version of the libdax API are as follows.
|
Operation completed successfully
The calling thread did not create ctx
See attributes(7) for descriptions of the following attributes:
|
dax_scan_value(3DAX), dax_zip_create(3DAX), dax_copy(3DAX), dax_vec_t(3DAX), dax_set_debug(3DAX), dax_set_log(3DAX)