diff --git a/tools/sbl_patch.py b/tools/sbl_patch.py
index 5951879..d1df582 100644
--- a/tools/sbl_patch.py
+++ b/tools/sbl_patch.py
@@ -2694,6 +2694,156 @@ def main():
 
     save(d, 'img0_b40.elf')
 
+    # b41 = b40 + desert-call retarget (0x8007790 bl EB90 -> EB0A):
+    # mid-function entry skipped the 48B prologue push; epilogue pop.w
+    # then over-popped 48B per S20 visit (100B/Kinsn creep, S20-rerun trap).
+    # Via prologue the frame balances and S20-pop returns to the caller.
+    # b40 = b39 + S12-wrapper pop restore (0x8007B66 bx ip -> pop):
+    # b25 chained the wrapper pop to bx ip; with S12 preserving sp the
+    # frame is intact, while bx ip orphaned 8B per call (105M sp-smash).
+    # Content below = b39 body with the 0x8007B66 hunk reverted to pop.
+    # naked clock leaf never sets r3 (stale stacked return addr,
+    # observed 0x8007975); the store writes 0xffff over live code,
+    # faulting the next fetch there. Peripheral write, nothing in emu
+    # reads it back (polls use scaffolded [0x73A100]). Same rationale
+    # as the b17 stm NOP. 2B->2B, no eating.
+    d = bytearray(base)
+
+    for va, (a, b) in BLX2BL.items():
+
+        set4(d, va, b)
+
+    for va, exp, new in [
+
+        (0x802F664, 'e4f77efc', 'c046c046'),
+
+        (0x802F672, 'e6f7bbfa', 'c046c046'),
+
+        (0x802F678, 'e9f7fef9', 'c046c046'),
+
+        (0x801B71E, '9847fee7', 'c046c046'),
+
+        (0x8008396, 'fef780f9', 'c046c046'),
+
+        (0x8016184, 'fdf790fc', 'c046c046'),
+
+        (0x801B76C, '280b0508', '00000708'),
+
+        (0x8017AFC, 'bde8fc87', 'eef74fb8'),
+
+        (0x802F690, '9847fee7', 'c046c046'),
+
+        (0x801671C, '9847fee7', 'c046c046'),
+
+        (0x8013BAC, 'f8bd0a4a', 'f3f730b8'),
+
+        (0x801B6BE, 'bde8f081', 'ebf7c3ba'),
+
+        (0x8016720, '10bd0000', 'f0f7b2ba'),
+
+        (0x8034810, 'bde8fc87', 'd2f76eba'),
+
+        (0x8007798, 'bde8f08f', 'fff7c3ba'),
+
+        (0x8016334, 'f0f7a4ff', 'c046c046'),
+
+        (0x8016342, 'f0f79eef', 'f0f79dbf'),
+
+        (0x802F9A0, 'ffe70648', 'd7f7c4b9'),
+
+        (0x802F9B2, 'fee78817', 'd7f7bbb9'),
+
+        (0x802F9A2, 'c4b943f2', 'd7f7c3b9'),
+
+        (0x800EB96, 'bde8fe8f', 'f8f7cfb8'),
+
+        (0x8007916, 'bde8f09f', 'fff720ba'),
+
+    ]:
+
+        assert hx(d, va) == exp, (hex(va), hx(d, va))
+
+        set4(d, va, new)
+
+    # Full bl idiom at 0x8006682 (bl + pop, 6B) -> NOPs (see b18).
+    o = off(0x8006682)
+
+    assert d[o:o + 6].hex() == '00f001f807bd', hx(d, 0x8006682, 6)
+
+    d[o:o + 6] = bytes.fromhex('c046c046c046')
+
+    o = off(0x8013AA8)
+
+    assert d[o:o + 4].hex() == '07f026be', hx(d, 0x8013AA8)
+
+    d[o:o + 4] = bytes.fromhex('c046c046')
+
+    o = off(0x8013BA6)
+
+    assert d[o:o + 2].hex() == 'b888', (hex(0x8013BA6), hx(d, 0x8013BA6, 2))
+
+    d[o:o + 2] = bytes.fromhex('ff20')
+
+    # b40: S12-wrapper pop RESTORED (revert b25 chain at this site).
+    # b25 sent pop to bx ip because the pop read scratch; with S12
+    # preserving sp the frame is intact, while bx ip orphaned 8B per
+    # call and smashed sp into text at 105M. S16 keeps serving B66.
+    o = off(0x8007B66)
+
+    assert d[o:o + 2].hex() == '10bd', hx(d, 0x8007B66, 2)
+
+    # Clock S11 (see b24).
+    o = off(0x800E988)
+
+    assert d[o:o + 2].hex() == '10bd', hx(d, 0x800E988, 2)
+
+    d[o:o + 2] = bytes.fromhex('7047')
+
+    # Poll re-init skip (see b27).
+    o = off(0x8007748)
+
+    assert d[o:o + 2].hex() == '96b1', hx(d, 0x8007748, 2)
+
+    d[o:o + 2] = bytes.fromhex('14e0')
+
+    # Poll-epilogue add NOP (see b28).
+    o = off(0x8007794)
+
+    assert d[o:o + 2].hex() == '07b0', hx(d, 0x8007794, 2)
+
+    d[o:o + 2] = bytes.fromhex('c046')
+
+    # Partition wrap-check (see b33).
+    o = off(0x801844E)
+
+    assert d[o:o + 2].hex() == '04d3', hx(d, 0x801844E, 2)
+
+    d[o:o + 2] = bytes.fromhex('04e0')
+
+    # Hang-site neutralize (see b34).
+    o = off(0x8018458)
+
+    assert d[o:o + 2].hex() == 'fee7', hx(d, 0x8018458, 2)
+
+    d[o:o + 2] = bytes.fromhex('ffe7')
+
+    # Clock-store NOP: str r1,[r3] (2B 1960) -> mov r8,r8 (2B c046).
+    o = off(0x800E986)
+
+    assert d[o:o + 2].hex() == '1960', hx(d, 0x800E986, 2)
+
+    d[o:o + 2] = bytes.fromhex('c046')
+
+    # b41 retarget.
+    o = off(0x8007790)
+
+    assert d[o:o + 4].hex() == '07f0fef9', hx(d, 0x8007790)
+
+    d[o:o + 4] = bytes.fromhex('07f0bbfb')
+
+    save(d, 'img0_b41.elf')
+
+