util/nvmutil: optimise cmd_swap() further
don't do xor swap. we know gbe2 is always 4KB higher than gbe in memory, so we can just set gbe2 to the value of gbe, and OR the size in bytes of 4KB into gbe2 this is only a marginal speed boost, negligible even, but it's done for the lulzfsdg20230625
parent
effcb942ce
commit
448ee5105d
|
@ -355,18 +355,16 @@ void
|
|||
cmd_swap(void)
|
||||
{
|
||||
int part0, part1;
|
||||
size_t g1 = (size_t) gbe;
|
||||
size_t g2 = (size_t) gbe2;
|
||||
size_t ptr;
|
||||
|
||||
part0 = validChecksum(0);
|
||||
part1 = validChecksum(1);
|
||||
|
||||
if (part0 || part1) {
|
||||
g1 ^= g2;
|
||||
g2 ^= g1;
|
||||
g1 ^= g2;
|
||||
gbe = (uint8_t *) g1;
|
||||
gbe2 = (uint8_t *) g2;
|
||||
gbe2 = gbe;
|
||||
ptr = (size_t) gbe;
|
||||
ptr |= SIZE_4KB;
|
||||
gbe = (uint8_t *) ptr;
|
||||
|
||||
gbeFileModified = 1;
|
||||
nvmPartModified[0] = 1;
|
||||
|
|
Loading…
Reference in New Issue