Make more fixes and improvements

- Remove PAGESIZE constant
- Fix realloc() documentation
- Fix ttyname_r() error reporting
- Make forking more reliable on Windows
- Make execvp() a few microseconds faster
- Make system() a few microseconds faster
- Tighten up the socket-related magic numbers
- Loosen restrictions on mmap() offset alignment
- Improve GetProgramExecutableName() with getenv("_")
- Use mkstemp() as basis for mktemp(), tmpfile(), tmpfd()
- Fix flakes in pthread_cancel_test, unix_test, fork_test
- Fix recently introduced futex stack overflow regression
- Let sockets be passed as stdio to subprocesses on Windows
- Improve security of bind() on Windows w/ SO_EXCLUSIVEADDRUSE

tool/viz/bf.c

@@ -35,6 +35,10 @@
 #define W 64
 #define C 253
 
+static unsigned long roundup2pow(unsigned long x) {
+  return x > 1 ? 2ul << _bsrl(x - 1) : x ? 1 : 0;
+}
+
 void bf(int fd) {
   ssize_t rc;
   uint64_t w;
@@ -44,7 +48,7 @@ void bf(int fd) {
   char ibuf[W], obuf[12 + 1 + W * 6 + 1 + W * (2 + 11) + 11 + 1];
   o = 0;
   if (fstat(fd, &st) != -1 && S_ISREG(st.st_mode) && st.st_size) {
-    bits = _bsr(_roundup2pow(st.st_size));
+    bits = _bsr(roundup2pow(st.st_size));
     bits = ROUNDUP(bits, 4);
   } else {
     bits = 48;

tool/viz/lib/dither.c

@@ -18,13 +18,12 @@
 ╚─────────────────────────────────────────────────────────────────────────────*/
 #include "dsp/tty/quant.h"
 #include "libc/intrin/hilbert.h"
-#include "libc/intrin/morton.h"
 #include "libc/log/check.h"
 #include "libc/log/log.h"
 #include "libc/macros.internal.h"
 #include "libc/math.h"
-#include "libc/mem/mem.h"
 #include "libc/mem/gc.internal.h"
+#include "libc/mem/mem.h"
 #include "libc/str/str.h"
 #include "tool/viz/lib/graphic.h"
 #include "tool/viz/lib/knobs.h"

tool/viz/lib/ycbcr2rgb3.c

@@ -27,6 +27,7 @@
 #include "dsp/scale/scale.h"
 #include "libc/calls/calls.h"
 #include "libc/calls/struct/sigset.h"
+#include "libc/intrin/bsr.h"
 #include "libc/intrin/pmulhrsw.h"
 #include "libc/log/check.h"
 #include "libc/log/log.h"
@@ -80,6 +81,14 @@ struct YCbCr {
   } luma, chroma;
 };
 
+static unsigned long roundup2pow(unsigned long x) {
+  return x > 1 ? 2ul << _bsrl(x - 1) : x ? 1 : 0;
+}
+
+static unsigned long rounddown2pow(unsigned long x) {
+  return x ? 1ul << _bsrl(x) : 0;
+}
+
 /**
  * Computes magnums for Y′CbCr decoding.
  *
@@ -95,8 +104,8 @@ void YCbCrComputeCoefficients(int swing, double gamma,
   double x;
   double f1[6][3];
   long longs[6][6];
-  long bitlimit = _roundup2pow(swing);
-  long wordoffset = _rounddown2pow((bitlimit - swing) / 2);
+  long bitlimit = roundup2pow(swing);
+  long wordoffset = rounddown2pow((bitlimit - swing) / 2);
   long chromaswing = swing + 2 * (bitlimit / 2. - swing / 2. - wordoffset);
   long lumamin = wordoffset;
   long lumamax = wordoffset + swing;

tool/viz/life.c

@@ -258,6 +258,12 @@ static char16_t statusline16[256];
     RES = (B11 | r0) & r1 & ~r2;                                           \
   } while (0)
 
+static void SwapBoards(void) {
+  uint64_t *t = board;
+  board = board2;
+  board2 = t;
+}
+
 static void Step(void) {
   long y, x, yn, xn;
   yn = byn >> 3;
@@ -275,7 +281,7 @@ static void Step(void) {
            board[(y + 1 < yn ? y + 1 : 0) * xn + (x + 1 < xn ? x + 1 : 0)]);
     }
   }
-  xchg(&board, &board2);
+  SwapBoards();
   ++generation;
 }
 
@@ -593,7 +599,7 @@ static int LoadFile(const char *path) {
     if ((c = ReadChar(f)) == -1) goto ReadError;
   }
   if (yn > byn || xn > bxn) goto ReadError;
-  xchg(&board, &board2);
+  SwapBoards();
   bzero(board, (byn * bxn) >> 3);
   yo = byn / 2 - yn / 2;
   xo = bxn / 2 - xn / 2;
@@ -640,7 +646,7 @@ static int LoadFile(const char *path) {
   return 0;
 ReadError:
   fclose(f);
-  xchg(&board, &board2);
+  SwapBoards();
   return -1;
 }
 

tool/viz/memzoom.c

@@ -31,8 +31,8 @@
 #include "libc/fmt/itoa.h"
 #include "libc/intrin/bits.h"
 #include "libc/intrin/bsf.h"
+#include "libc/intrin/bsr.h"
 #include "libc/intrin/hilbert.h"
-#include "libc/intrin/morton.h"
 #include "libc/intrin/safemacros.internal.h"
 #include "libc/log/log.h"
 #include "libc/macros.internal.h"
@@ -188,6 +188,10 @@ static void LeaveScreen(void) {
   Write("\e[H\e[J");
 }
 
+static unsigned long rounddown2pow(unsigned long x) {
+  return x ? 1ul << _bsrl(x) : 0;
+}
+
 static void GetTtySize(void) {
   struct winsize wsize;
   wsize.ws_row = tyn + 1;
@@ -195,8 +199,8 @@ static void GetTtySize(void) {
   tcgetwinsize(out, &wsize);
   tyn = MAX(2, wsize.ws_row) - 1;
   txn = MAX(17, wsize.ws_col) - 16;
-  tyn = _rounddown2pow(tyn);
-  txn = _rounddown2pow(txn);
+  tyn = rounddown2pow(tyn);
+  txn = rounddown2pow(txn);
   tyn = MIN(tyn, txn);
 }
 
@@ -278,11 +282,30 @@ static void SetupCanvas(void) {
   }
   displaysize = ROUNDUP(ROUNDUP((tyn * txn) << zoom, 16), 1ul << zoom);
   canvassize = ROUNDUP(displaysize, FRAMESIZE);
-  buffersize = ROUNDUP(tyn * txn * 16 + PAGESIZE, FRAMESIZE);
+  buffersize = ROUNDUP(tyn * txn * 16 + 4096, FRAMESIZE);
   canvas = Allocate(canvassize);
   buffer = Allocate(buffersize);
 }
 
+/**
+ * Interleaves bits.
+ * @see https://en.wikipedia.org/wiki/Z-order_curve
+ * @see unmorton()
+ */
+static unsigned long morton(unsigned long y, unsigned long x) {
+  x = (x | x << 020) & 0x0000FFFF0000FFFF;
+  x = (x | x << 010) & 0x00FF00FF00FF00FF;
+  x = (x | x << 004) & 0x0F0F0F0F0F0F0F0F;
+  x = (x | x << 002) & 0x3333333333333333;
+  x = (x | x << 001) & 0x5555555555555555;
+  y = (y | y << 020) & 0x0000FFFF0000FFFF;
+  y = (y | y << 010) & 0x00FF00FF00FF00FF;
+  y = (y | y << 004) & 0x0F0F0F0F0F0F0F0F;
+  y = (y | y << 002) & 0x3333333333333333;
+  y = (y | y << 001) & 0x5555555555555555;
+  return x | y << 1;
+}
+
 static long IndexSquare(long y, long x) {
   switch (order) {
     case LINEAR:

tool/viz/printvideo.c

@@ -371,7 +371,7 @@ static bool CloseSpeaker(void) {
 }
 
 static void ResizeVtFrame(struct VtFrame *f, size_t yn, size_t xn) {
-  BALLOC(&f->b, PAGESIZE, 64 + yn * (xn * 32 + 8), __FUNCTION__);
+  BALLOC(&f->b, 4096, 64 + yn * (xn * 32 + 8), __FUNCTION__);
   f->i = f->n = 0;
 }
 
@@ -837,7 +837,8 @@ static void OnVideo(plm_t *mpeg, plm_frame_t *pf, void *user) {
   } else {
     TranscodeVideo(pf);
     if (!f1_->n) {
-      xchg(&f1_, &f2_);
+      struct VtFrame *t = f1_;
+      f1_ = f2_, f2_ = t;
       f1_start_ = decode_start_;
     } else {
       f2_start_ = decode_start_;
@@ -872,7 +873,7 @@ static void OpenVideo(void) {
 static ssize_t WriteVideoCall(void) {
   size_t amt;
   ssize_t rc;
-  amt = min(PAGESIZE * 4, f1_->n - f1_->i);
+  amt = min(4096 * 4, f1_->n - f1_->i);
   if ((rc = write(outfd_, f1_->bytes + f1_->i, amt)) != -1) {
     if ((f1_->i += rc) == f1_->n) {
       if (plm_get_audio_enabled(plm_)) {
@@ -883,7 +884,8 @@ static ssize_t WriteVideoCall(void) {
       }
       f1_start_ = f2_start_;
       f1_->i = f1_->n = 0;
-      xchg(&f1_, &f2_);
+      struct VtFrame *t = f1_;
+      f1_ = f2_, f2_ = t;
       RecordFactThatFrameWasFullyRendered();
     }
   }