Allow reading of little-endian frags on big-endian

... and vice-versa. We'll fix up frag header values for our output
parameter from liberasurecode_get_fragment_metadata but otherwise
avoid manipulating the in-memory fragment much.

Change-Id: Idd6833bdea60e27c9a0148ee28b4a2c1070be148

3 files changed, 117 insertions, 22 deletions

diff --git a/include/erasurecode/erasurecode_helpers.h b/include/erasurecode/erasurecode_helpers.h
index f0be41a..0168a55 100644
--- a/include/erasurecode/erasurecode_helpers.h
+++ b/include/erasurecode/erasurecode_helpers.h
@@ -93,5 +93,21 @@ void *get_aligned_buffer16(int size);
 
 /* ==~=*=~==~=*=~==~=*=~==~=*=~==~=*=~==~=*=~==~=*=~==~=*=~==~=*=~==~=*=~== */
 
+#ifndef bswap_32
+static __inline uint32_t __libec_bswap_32(uint32_t __x)
+{
+    return __x>>24 | (__x>>8&0xff00) | (__x<<8&0xff0000) | __x<<24;
+}
+#define bswap_32(x) __libec_bswap_32(x)
+#endif
+
+#ifndef bswap_64
+static __inline uint64_t __libec_bswap_64(uint64_t __x)
+{
+    return (__libec_bswap_32(__x)+0ULL)<<32 | __libec_bswap_32(__x>>32);
+}
+#define bswap_64(x) __libec_bswap_64(x)
+#endif
+
 #endif  // _ERASURECODE_HELPERS_H_
 
diff --git a/src/erasurecode.c b/src/erasurecode.c
index 27542c5..dd59ae5 100644
--- a/src/erasurecode.c
+++ b/src/erasurecode.c
@@ -1055,17 +1055,36 @@ int liberasurecode_get_fragment_metadata(char *fragment,
     memcpy(fragment_metadata, fragment, sizeof(struct fragment_metadata));
     fragment_hdr = (fragment_header_t *) fragment;
     if (LIBERASURECODE_FRAG_HEADER_MAGIC != fragment_hdr->magic) {
-        log_error("Invalid fragment, illegal magic value");
-        ret = -EINVALIDPARAMS;
-        goto out;
+        if (LIBERASURECODE_FRAG_HEADER_MAGIC != bswap_32(fragment_hdr->magic)) {
+            log_error("Invalid fragment, illegal magic value");
+            ret = -EINVALIDPARAMS;
+            goto out;
+        } else {
+            // Must've written this on an opposite-endian architecture.
+            // Fix it in fragment_metadata
+            fragment_metadata->idx = bswap_32(fragment_metadata->idx);
+            fragment_metadata->size = bswap_32(fragment_metadata->size);
+            fragment_metadata->frag_backend_metadata_size =
+                bswap_32(fragment_metadata->frag_backend_metadata_size);
+            fragment_metadata->orig_data_size =
+                bswap_64(fragment_metadata->orig_data_size);
+            fragment_metadata->chksum_type =
+                bswap_32(fragment_metadata->chksum_type);
+            for (int i = 0; i < LIBERASURECODE_MAX_CHECKSUM_LEN; i++) {
+                fragment_metadata->chksum[i] =
+                    bswap_32(fragment_metadata->chksum[i]);
+            }
+            fragment_metadata->backend_version =
+                bswap_32(fragment_metadata->backend_version);
+        }
     }
 
-    switch(fragment_hdr->meta.chksum_type) {
+    switch(fragment_metadata->chksum_type) {
         case CHKSUM_CRC32: {
             uint32_t computed_chksum = 0;
-            uint32_t stored_chksum = fragment_hdr->meta.chksum[0];
+            uint32_t stored_chksum = fragment_metadata->chksum[0];
             char *fragment_data = get_data_ptr_from_fragment(fragment);
-            uint64_t fragment_size = fragment_hdr->meta.size;
+            uint64_t fragment_size = fragment_metadata->size;
             computed_chksum = crc32(0, (unsigned char *) fragment_data, fragment_size);
             if (stored_chksum != computed_chksum) {
                 // Try again with our "alternative" crc32; see
@@ -1095,28 +1114,39 @@ out:
 
 int is_invalid_fragment_header(fragment_header_t *header)
 {
-    uint32_t csum = 0;
+    uint32_t csum = 0, metadata_chksum = 0, libec_version = 0;
     assert (NULL != header);
     if (header->libec_version == 0)
         /* libec_version must be bigger than 0 */
         return 1;
-    if (header->libec_version < _VERSION(1,2,0))
-        /* no metadata checksum support */
-        return 0;
-
+    metadata_chksum = header->metadata_chksum;
+    libec_version = header->libec_version;
     if (header->magic != LIBERASURECODE_FRAG_HEADER_MAGIC) {
-        log_error("Invalid fragment header (get meta chksum)!");
-        return 1;
+        if (bswap_32(header->magic) != LIBERASURECODE_FRAG_HEADER_MAGIC) {
+            log_error("Invalid fragment header (get meta chksum)!");
+            return 1;
+        } else {
+            // Must've written this on an opposite-endian architecture.
+            // Fix our reference checksum and version, but *don't touch
+            // the header data yet*. Leave that for when we're extracting
+            // in liberasurecode_get_fragment_metadata
+            metadata_chksum = bswap_32(metadata_chksum);
+            libec_version = bswap_32(libec_version);
+        }
     }
 
+    if (libec_version < _VERSION(1,2,0))
+        /* no metadata checksum support */
+        return 0;
+
     csum = crc32(0, (unsigned char *) &header->meta, sizeof(fragment_metadata_t));
-    if (header->metadata_chksum == csum) {
+    if (metadata_chksum == csum) {
         return 0;
     }
     // Else, try again with our "alternative" crc32; see
     // https://bugs.launchpad.net/liberasurecode/+bug/1666320
     csum = liberasurecode_crc32_alt(0, &header->meta, sizeof(fragment_metadata_t));
-    return (header->metadata_chksum != csum);
+    return (metadata_chksum != csum);
 }
 
 int liberasurecode_verify_fragment_metadata(ec_backend_t be,
diff --git a/test/liberasurecode_test.c b/test/liberasurecode_test.c
index a7bb612..d3ca200 100644
--- a/test/liberasurecode_test.c
+++ b/test/liberasurecode_test.c
@@ -1762,29 +1762,77 @@ static void test_verify_stripe_metadata_frag_idx_invalid(
     verify_fragment_metadata_mismatch_impl(be_id, args, FRAGIDX_OUT_OF_RANGE);
 }
 
-static void test_metadata_crcs()
+static void test_metadata_crcs_le()
 {
     // We've observed headers like this in the wild, using our busted crc32
-    char header[] =
+    char orig_header[] =
         "\x03\x00\x00\x00\x00\x00\x04\x00\x00\x00\x00\x00\x00\x00\x10\x00"
         "\x00\x00\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
         "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
         "\x00\x00\x00\x00\x00\x00\x07\x01\x0e\x02\x00\xcc\x5e\x0c\x0b\x00"
         "\x04\x01\x00\x22\xee\x45\xb9\x00\x00\x00\x00\x00\x00\x00\x00\x00";
+    char header[sizeof(orig_header)];
+    memcpy(header, orig_header, sizeof(orig_header));
 
     fragment_metadata_t res;
 
     assert(liberasurecode_get_fragment_metadata(header, &res) == 0);
+    assert(memcmp(header, orig_header, sizeof(orig_header)) == 0);
+    assert(res.backend_version == _VERSION(2, 14, 1));
     assert(is_invalid_fragment_header((fragment_header_t *) header) == 0);
+    assert(memcmp(header, orig_header, sizeof(orig_header)) == 0);
 
     // Switch it to zlib's implementation
-    header[70] = '\x18';
-    header[69] = '\x73';
-    header[68] = '\xf8';
-    header[67] = '\xec';
+    orig_header[70] = '\x18';
+    orig_header[69] = '\x73';
+    orig_header[68] = '\xf8';
+    orig_header[67] = '\xec';
+    memcpy(header, orig_header, sizeof(orig_header));
 
     assert(liberasurecode_get_fragment_metadata(header, &res) == 0);
+    assert(memcmp(header, orig_header, sizeof(orig_header)) == 0);
+    assert(res.backend_version == _VERSION(2, 14, 1));
     assert(is_invalid_fragment_header((fragment_header_t *) header) == 0);
+    assert(memcmp(header, orig_header, sizeof(orig_header)) == 0);
+
+    // Write down the wrong thing
+    header[70] = '\xff';
+    assert(liberasurecode_get_fragment_metadata(header, &res) == -EBADHEADER);
+    assert(is_invalid_fragment_header((fragment_header_t *) header) == 1);
+}
+
+static void test_metadata_crcs_be()
+{
+    // Like above, but big-endian
+    char orig_header[] =
+        "\x00\x00\x00\x03\x00\x04\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
+        "\x00\x10\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
+        "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
+        "\x00\x00\x00\x00\x00\x00\x07\x00\x02\x0e\x01\x0b\x0c\x5e\xcc\x00"
+        "\x01\x04\x00\xfa\x85\x40\x70\x00\x00\x00\x00\x00\x00\x00\x00\x00";
+    char header[sizeof(orig_header)];
+    memcpy(header, orig_header, sizeof(orig_header));
+
+    fragment_metadata_t res;
+
+    assert(liberasurecode_get_fragment_metadata(header, &res) == 0);
+    assert(memcmp(header, orig_header, sizeof(orig_header)) == 0);
+    assert(res.backend_version == _VERSION(2, 14, 1));
+    assert(is_invalid_fragment_header((fragment_header_t *) header) == 0);
+    assert(memcmp(header, orig_header, sizeof(orig_header)) == 0);
+
+    // Switch it to zlib's implementation
+    orig_header[67] = '\xe3';
+    orig_header[68] = '\x73';
+    orig_header[69] = '\x88';
+    orig_header[70] = '\xa0';
+    memcpy(header, orig_header, sizeof(orig_header));
+
+    assert(liberasurecode_get_fragment_metadata(header, &res) == 0);
+    assert(memcmp(header, orig_header, sizeof(orig_header)) == 0);
+    assert(res.backend_version == _VERSION(2, 14, 1));
+    assert(is_invalid_fragment_header((fragment_header_t *) header) == 0);
+    assert(memcmp(header, orig_header, sizeof(orig_header)) == 0);
 
     // Write down the wrong thing
     header[70] = '\xff';
@@ -1832,7 +1880,8 @@ struct testcase testcases[] = {
     TEST(test_fragments_needed_invalid_args, EC_BACKENDS_MAX, CHKSUM_TYPES_MAX),
     TEST(test_get_fragment_partition, EC_BACKENDS_MAX, CHKSUM_TYPES_MAX),
     TEST(test_liberasurecode_get_version, EC_BACKENDS_MAX, CHKSUM_TYPES_MAX),
-    TEST(test_metadata_crcs, EC_BACKENDS_MAX, 0),
+    TEST(test_metadata_crcs_le, EC_BACKENDS_MAX, 0),
+    TEST(test_metadata_crcs_be, EC_BACKENDS_MAX, 0),
     // NULL backend test
     TEST(test_create_and_destroy_backend, EC_BACKEND_NULL, CHKSUM_NONE),
     TEST(test_simple_encode_decode, EC_BACKEND_NULL, CHKSUM_NONE),