1
0
Fork 0
mirror of https://git.tukaani.org/xz.git synced 2024-04-04 12:36:23 +02:00

liblzma: LZMA decoder: Optimize loop comparison.

But now it needs one more local variable.
This commit is contained in:
Lasse Collin 2024-02-12 17:09:10 +02:00
parent 88276f9f2c
commit 5e04706b91
2 changed files with 11 additions and 4 deletions

View file

@ -261,7 +261,7 @@ lzma_decode(void *coder_ptr, lzma_dict *restrict dictptr,
const size_t dict_start = dict.pos; const size_t dict_start = dict.pos;
// Range decoder // Range decoder
rc_to_local(coder->rc, *in_pos); rc_to_local(coder->rc, *in_pos, LZMA_IN_REQUIRED);
// State // State
uint32_t state = coder->state; uint32_t state = coder->state;
@ -340,8 +340,7 @@ lzma_decode(void *coder_ptr, lzma_dict *restrict dictptr,
// If there is not enough room for another LZMA symbol // If there is not enough room for another LZMA symbol
// go to Resumable mode. // go to Resumable mode.
if (unlikely(rc_in_end - rc_in_ptr < LZMA_IN_REQUIRED if (unlikely(!rc_is_fast_allowed() || dict.pos == dict.limit))
|| dict.pos == dict.limit))
goto slow; goto slow;
// Decode the first bit from the next LZMA symbol. // Decode the first bit from the next LZMA symbol.

View file

@ -61,13 +61,21 @@ rc_read_init(lzma_range_decoder *rc, const uint8_t *restrict in,
/// Makes local copies of range decoder and *in_pos variables. Doing this /// Makes local copies of range decoder and *in_pos variables. Doing this
/// improves speed significantly. The range decoder macros expect also /// improves speed significantly. The range decoder macros expect also
/// variables 'in' and 'in_size' to be defined. /// variables 'in' and 'in_size' to be defined.
#define rc_to_local(range_decoder, in_pos) \ #define rc_to_local(range_decoder, in_pos, fast_mode_in_required) \
lzma_range_decoder rc = range_decoder; \ lzma_range_decoder rc = range_decoder; \
const uint8_t *rc_in_ptr = in + (in_pos); \ const uint8_t *rc_in_ptr = in + (in_pos); \
const uint8_t *rc_in_end = in + in_size; \ const uint8_t *rc_in_end = in + in_size; \
const uint8_t *rc_in_fast_end \
= (rc_in_end - rc_in_ptr) <= (fast_mode_in_required) \
? rc_in_ptr \
: rc_in_end - (fast_mode_in_required); \
uint32_t rc_bound uint32_t rc_bound
/// Evaluates to true if there is enough input remaining to use fast mode.
#define rc_is_fast_allowed() (rc_in_ptr < rc_in_fast_end)
/// Stores the local copes back to the range decoder structure. /// Stores the local copes back to the range decoder structure.
#define rc_from_local(range_decoder, in_pos) \ #define rc_from_local(range_decoder, in_pos) \
do { \ do { \