CVE-2026-112894

============================================================================
Quadratic-time DoS in PDF save_all (PdfParser.next_object_id)
Location: src/PIL/PdfParser.py:629  in next_object_id()
Commit:   a707c6b21c313880d0771e54965d142dd42358b6
============================================================================

Pillow's PDF writer takes time proportional to the square of the number of frames when it saves a multi-frame image as a multi-page PDF. Anyone who can get an application to convert their uploaded image this way can tie up a worker at full CPU for minutes with a small file. A 240 KB animated image takes about 70 seconds, a 720 KB one an estimated 10 minutes, and a 125-byte animated PNG that merely declares a large frame count triggers the same slowdown with no upper bound short of 2³¹ frames. Memory stays under 100 MB and the decompression-bomb limit never applies.

The cause is PdfParser.next_object_id() in src/PIL/PdfParser.py, which finds the next free id by computing max(self.xref_table.keys()). XrefTable.keys() rebuilds three sets over every id reserved so far on each call. PdfImagePlugin._save() calls it three times (four for palette images with transparency) per frame, for n_frames frames, before any page is written. That makes the id-reservation loop alone O(N²). Because n_frames is taken from the file before any frame is seeked, APNG's acTL count drives the loop even when the frames don't exist. In that case the save raises OSError only after the full delay.

The precondition is that the application calls im.save(..., "PDF", save_all=True) on the attacker's file; Image.open() and load() alone do not reach this code. The attached reproducer shows 4× the frames costing about 11× the CPU for a valid GIF and about 16× for a fixed 125-byte APNG. The behaviour was introduced in 6207b44ab1ff (2018-01-18) and first shipped in Pillow 5.1.0. A small patch that tracks the highest id in XrefTable makes the save linear.

Details

Saving a multi-frame image as a multi-page PDF (im.save(fp, "PDF", save_all=True)) costs CPU time proportional to the square of the frame count, independent of pixel content, because every object id the PDF writer reserves is found by rebuilding and scanning the set of all ids reserved so far.

A 1×1-pixel animated GIF made of N identical 15-byte frames (built by the reproducer) gives, on a 2-CPU container with the stock build at this commit (absolute timings vary by machine; the growth rate does not):

| frames | GIF size | im.save(BytesIO(), "PDF", save_all=True) | peak RSS | |---|---|---|---| | 2 000 | 30 KB | 1.1 s | | | 8 000 | 120 KB | 14.7 s | | | 16 000 | 240 KB | 70 s | 60 MB | | 32 000 | 480 KB | 262 s | 95 MB | | 48 000 | 720 KB | still running at 120 s (extrapolated ≈ 10 min) | |

Each doubling of the input roughly quadruples the time. cProfile of the 8 000-frame save attributes 13.5 s of 17.6 s to 24 003 calls of PdfParser.next_object_id (8.9 s inside XrefTable.keys, 3.9 s inside max); encoding and writing the 8 000 pages is the remaining ~4 s and scales linearly.

The per-page id reservation loop runs before any page is written:

# src/PIL/PdfImagePlugin.py:247-259  (_save)
    for im in ims:
        im_number_of_pages = 1
        if save_all:
            im_number_of_pages = getattr(im, "n_frames", 1)
        …
        for i in range(im_number_of_pages):
            image_refs.append(existing_pdf.next_object_id(0))
            if im.mode == "P" and "transparency" in im.info:
                image_refs.append(existing_pdf.next_object_id(0))

            page_refs.append(existing_pdf.next_object_id(0))
            contents_refs.append(existing_pdf.next_object_id(0))

n_frames comes straight from the file (for GIF it is the number of image blocks; the GIF plugin counts them by walking block headers, which is cheap). The frames do not even have to exist: the loop runs on n_frames before any…

Python is an interpreted, interactive, object-oriented programming language, which includes modules, classes, exceptions, very high level dynamic data types and dynamic typing. Python supports interfaces to many system calls and libraries, as well as to various windowing systems.

Security Fix(es):

• pillow: Pillow: Denial of Service via unbounded recursion in PDF parser (CVE-2026-112391)

• pillow: pillow: Denial of service via double free during image transformation (CVE-2026-112402)

• pillow: pillow: Denial of Service via zero-size images in rank filters (CVE-2026-112392)

• pillow: pillow: Denial of Service via out-of-bounds memory access on zero-width images (CVE-2026-112397)

• pillow: Pillow: Denial of Service via null pointer dereference in Image.merge() (CVE-2026-112405)

• pillow: pillow: Denial of Service via crafted multi-frame image conversion (CVE-2026-112894)

• pillow: pillow: Denial of service via crafted PDF during append operations (CVE-2026-112895)

• pillow: pillow: Denial of Service via infinite loop during PDF page rewriting (CVE-2026-112406)

• pillow: pillow: Denial of Service via duplicate page references in PDF files (CVE-2026-96640)

• pillow: pillow: Denial of Service via unbounded memory expansion in MSP image decoder (CVE-2026-112892)

For more details about the security issue(s), including the impact, a CVSS score, acknowledgments, and other related information, refer to the CVE page(s) listed in the References section.

Pillow is a Python imaging library. From 5.1.0 until 13.0.0, Pillow takes quadratic time to reserve PDF object identifiers when saving many frames with save_all enabled. PdfParser.next_object_id() repeatedly rebuilds and scans all previously reserved identifiers for every object. An attacker can supply a multi-frame image, including a small APNG with a large declared frame count, to an application that converts it to a multipage PDF. This can consume excessive CPU and deny service. This issue is fixed in version 13.0.0.

Python is an interpreted, interactive, object-oriented programming language, which includes modules, classes, exceptions, very high level dynamic data types and dynamic typing. Python supports interfaces to many system calls and libraries, as well as to various windowing systems.

Security Fix(es):

• pillow: Pillow: Denial of Service via unbounded recursion in PDF parser (CVE-2026-112391)
• pillow: pillow: Denial of service via double free during image transformation (CVE-2026-112402)
• pillow: pillow: Denial of Service via zero-size images in rank filters (CVE-2026-112392)
• pillow: pillow: Denial of Service via out-of-bounds memory access on zero-width images (CVE-2026-112397)
• pillow: Pillow: Denial of Service via null pointer dereference in Image.merge() (CVE-2026-112405)
• pillow: pillow: Denial of Service via crafted multi-frame image conversion (CVE-2026-112894)
• pillow: pillow: Denial of service via crafted PDF during append operations (CVE-2026-112895)
• pillow: pillow: Denial of Service via infinite loop during PDF page rewriting (CVE-2026-112406)
• pillow: pillow: Denial of Service via duplicate page references in PDF files (CVE-2026-96640)
• pillow: pillow: Denial of Service via unbounded memory expansion in MSP image decoder (CVE-2026-112892)

For more details about the security issue(s), including the impact, a CVSS score, acknowledgments, and other related information, refer to the CVE page(s) listed in the References section.

Pillow is a Python imaging library. From 5.1.0 until 13.0.0, Pillow takes quadratic time to reserve PDF object identifiers when saving many frames with save_all enabled. PdfParser.next_object_id() repeatedly rebuilds and scans all previously reserved identifiers for every object. An attacker can supply a multi-frame image, including a small APNG with a large declared frame count, to an application that converts it to a multipage PDF. This can consume excessive CPU and deny service. This issue is fixed in version 13.0.0.

TuxCare rebuilt pillow to address 10 CVEs. Fixed in 11.3.0.post8+tuxcare.

TuxCare rebuilt pillow to address 10 CVEs. Fixed in 11.2.1.post10+tuxcare.

TuxCare rebuilt pillow to address 10 CVEs. Fixed in 10.4.0.post12+tuxcare.

TuxCare rebuilt pillow to address 9 CVEs. Fixed in 9.5.0.post10+tuxcare.

TuxCare rebuilt pillow to address 9 CVEs. Fixed in 9.4.0.post11+tuxcare.

TuxCare rebuilt pillow to address 9 CVEs. Fixed in 8.4.0.post11+tuxcare.

These are all security issues fixed in the python313-Pillow-13.0.0-1.1 package on the GA media of openSUSE Tumbleweed.

Root has patched CVE-2026-112894 in the pillow package for Root:PyPI. Multiple fixed versions available.

This update for python-Pillow fixes the following issues:

• CVE-2026-96636: decompression-bomb protection bypass via multi-frame seek() in MPO and DCX (bsc#1286802).
• CVE-2026-96637: decompression bomb check bypass in SpiderImagePlugin.seek() leads to unbounded allocation DoS (bsc#1286801).
• CVE-2026-96640: PDF append accepts malformed shared page-tree DAG, which can cause resource exhaustion (bsc#1286798).
• CVE-2026-96641: reachable undefined behavior in native integer operations can lead to denial of service or logic integrity compromise (bsc#1286797).
• CVE-2026-96642: TarIO negative size field causes infinite CPU-spin loop (bsc#1286796).
• CVE-2026-112390: Denial of Service via infinite loop during crafted MIC image parsing (bsc#1286256).
• CVE-2026-112391: Denial of Service via unbounded recursion in PDF parser (bsc#1286273).
• CVE-2026-112392: Denial of Service via zero-size images in rank filters (bsc#1286287).
• CVE-2026-112393: heap buffer overflow and stale-heap disclosure in _tiffWriteProc when saving compressed TIFF data (bsc#1286325).
• CVE-2026-112394: P;2L and P;4L decoders may read 1 to 3 bytes past the passed buffer size (bsc#1286311).
• CVE-2026-112397: out-of-bounds access in _imagingmorph on a zero-width image (bsc#1286314).
• CVE-2026-112398: heap buffer overflow in FPX JPEG decoding can lead to code execution (bsc#1286406).
• CVE-2026-112399: Denial of service via crafted APNG image (bsc#1286368).
• CVE-2026-112400: Denial of Service via memory exhaustion from crafted GIF image (bsc#1286370).
• CVE-2026-112401: Denial of Service via crafted GIF comments (bsc#1286372).
• CVE-2026-112402: Denial of service via double free during image transformation (bsc#1286381).
• CVE-2026-112403: heap buffer overflow in ImagingJpegDecode allows for code execution (bsc#1286405).
• CVE-2026-112405: NULL pointer dereference in Image.merge via an unvalidated band 0 (bsc#1286404).
• CVE-2026-112406: infinite loop in PdfParser.rewrite_pages on page tree Parent cycle (bsc#1286403).
• CVE-2026-112891: quadratic GIF comment copying during save (bsc#1286402).
• CVE-2026-112892: MSP RLE decoder expands a tiny file into unbounded memory (bsc#1286401).
• CVE-2026-112894: Denial of Service via crafted multi-frame image conversion (bsc#1286408).
• CVE-2026-112895: Denial of service via crafted PDF during append operations (bsc#1286411).
• CVE-2026-112896: TIFF encoder ColorMap plane-size mismatch causes a heap out-of-bounds read (bsc#1286400).
• CVE-2026-112897: TiffDecode returns uninitialized memory as pixels for short JPEG strips (bsc#1286398).
• CVE-2026-112899: EPS load(scale) lets Ghostscript render past the decompression bomb limit (bsc#1286489).
• CVE-2026-112900: argument Injection in Ghostscript call can lead to remote code execution (bsc#1286397).
• CVE-2026-112901: ReDoS in XbmImagePlugin can be triggered by 512-byte file (bsc#1286396).
• CVE-2026-112902: memory disclosure and segfault in libtiff encoder via crafted TIFF tags (bsc#1286804).