fpga_compress: pad a single bitstream to the interleave boundary

zlib_decompress() walks its output in whole FPGA_INTERLEAVE_SIZE chunks:

  for (long k = 0; k < *outsize / (FPGA_INTERLEAVE_SIZE * num_outfiles); k++)

so a stream that is not a whole number of chunks loses its trailing partial one.
With two or more inputs the read loop zero-pads each stream past EOF and the
total lands on a boundary, but the padding was guarded by 'num_infiles > 1', so
the single input case was left ragged. 42172 bytes of fpga_pm3_hf.bit is 146.43
chunks, and -d handed back 39788 - a clean looking prefix, short by 2384 bytes,
22 of them real bitstream data.

Gate the padding on single_block instead. It must not be 'always pad': -s is the
.data section, and start.c's uncompress_data_section() sizes the decompression
with __data_end__ - __data_start__. Rounding .data from 14944 up to 14976 makes
LZ4_decompress_safe() return an error, and that path is the LED panic loop, so
the firmware would never reach AppMain().

  1 bitstream   42172 in -> 42336 packed, -d round trip byte identical,
                archive 28730 -> 28731
  4 bitstreams  archive byte identical to 43fe6c3eb, round trip exact
  -s .data      byte identical to 43fe6c3eb on the same input, unpadded

The 164 padding bytes never reach the FPGA: DownloadFPGA() shifts out only
bitstream_length bytes, taken from the .bit 'e' section header.

Also simulated the ARM decoder over the new archive - LZ4_decompress_safe_continue()
block by block into a FPGA_RING_BUFFER_BYTES buffer - 3 blocks of 16384/16384/9568,
none over the ring buffer.

Co-Authored-By: Claude Opus 5 (1M context)
This commit is contained in:
iceman1001andClaude Opus 5 (1M context) committed 2026-09-13 08:27:53 +02:00
1 parent 36b16b943f
commit f0b569b507
2 files changed
+12 -1

No files matched your search

+1
View File
@@ -3,6 +3,7 @@ All notable changes to this project will be documented in this file.
This project uses the changelog in accordance with [keepchangelog](http://keepachangelog.com/). Please use this to write notable changes, which is not the same as git commit log...
## [unreleased][unreleased]
- Fixed `fpga_compress -d` - a single bitstream archive was unpacked 2384 bytes short, the trailing partial interleave chunk was dropped (@iceman1001)
- Fixed `fpga_compress` - a build with only one FPGA bitstream (`SKIP_LF`, `SKIP_FELICA`, `SKIP_ISO15693`) packed it as a single 1 MB block the ARM could not decompress, `inflate returned: -13247` (@iceman1001)
- Fixed `FpgaDownloadAndGo` - a failed FPGA decompress leaked the ring buffer, leaving BigBuf 16 kB short for the rest of the session (@iceman1001)
- Added `hf mf dump --sk` - save extracted keys to binary file (@iceman1001)
+11 -1
View File
@@ -71,13 +71,23 @@ static int zlib_compress(FILE *infile[], uint8_t num_infiles, FILE *outfile, boo
return (EXIT_FAILURE);
}
// Pad a bitstream that ran out mid chunk with zeroes, so total_size stays a
// whole number of FPGA_INTERLEAVE_SIZE chunks. zlib_decompress() walks the
// output in whole chunks and drops a trailing partial one, so an unpadded
// stream comes back short - 42172 bytes of fpga_pm3_hf.bit returned as 39788.
//
// Not for "-s". That is the .data section, and start.c sizes the
// decompression with __data_end__ - __data_start__. Rounding .data up to a
// chunk boundary makes LZ4_decompress_safe() overrun that and return an
// error, which drops the firmware into the LED panic loop in
// uncompress_data_section().
for (uint16_t j = 0; j < num_infiles; j++) {
for (uint16_t k = 0; k < FPGA_INTERLEAVE_SIZE; k++) {
uint8_t c = (uint8_t)fgetc(infile[j]);
if (!feof(infile[j])) {
fpga_config[total_size++] = c;
} else if (num_infiles > 1) {
} else if (single_block == false) {
fpga_config[total_size++] = '\0';
}
}