diff options
Diffstat (limited to 'lzip_index.cc')
| -rw-r--r-- | lzip_index.cc | 180 |
1 files changed, 105 insertions, 75 deletions
diff --git a/lzip_index.cc b/lzip_index.cc index f70307c..66eb30d 100644 --- a/lzip_index.cc +++ b/lzip_index.cc @@ -1,18 +1,18 @@ -/* Lziprecover - Data recovery tool for the lzip format - Copyright (C) 2009-2019 Antonio Diaz Diaz. +/* Lziprecover - Data recovery tool for the lzip format + Copyright (C) 2009-2021 Antonio Diaz Diaz. - This program is free software: you can redistribute it and/or modify - it under the terms of the GNU General Public License as published by - the Free Software Foundation, either version 2 of the License, or - (at your option) any later version. + This program is free software: you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation, either version 2 of the License, or + (at your option) any later version. - This program is distributed in the hope that it will be useful, - but WITHOUT ANY WARRANTY; without even the implied warranty of - MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - GNU General Public License for more details. + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. - You should have received a copy of the GNU General Public License - along with this program. If not, see <http://www.gnu.org/licenses/>. + You should have received a copy of the GNU General Public License + along with this program. If not, see <http://www.gnu.org/licenses/>. */ #define _FILE_OFFSET_BITS 64 @@ -39,6 +39,18 @@ int seek_read( const int fd, uint8_t * const buf, const int size, } +bool Lzip_index::check_header_error( const Lzip_header & header, + const bool ignore_bad_ds ) + { + if( !header.verify_magic() ) + { error_ = bad_magic_msg; retval_ = 2; return true; } + if( !header.verify_version() ) + { error_ = bad_version( header.version() ); retval_ = 2; return true; } + if( !ignore_bad_ds && !isvalid_ds( header.dictionary_size() ) ) + { error_ = bad_dict_msg; retval_ = 2; return true; } + return false; + } + void Lzip_index::set_errno_error( const char * const msg ) { error_ = msg; error_ += std::strerror( errno ); @@ -53,22 +65,40 @@ void Lzip_index::set_num_error( const char * const msg, unsigned long long num ) retval_ = 2; } + +bool Lzip_index::read_header( const int fd, Lzip_header & header, + const long long pos ) + { + if( seek_read( fd, header.data, Lzip_header::size, pos ) != Lzip_header::size ) + { set_errno_error( "Error reading member header: " ); return false; } + return true; + } + +bool Lzip_index::read_trailer( const int fd, Lzip_trailer & trailer, + const long long pos ) + { + if( seek_read( fd, trailer.data, Lzip_trailer::size, + pos - Lzip_trailer::size ) != Lzip_trailer::size ) + { set_errno_error( "Error reading member trailer: " ); return false; } + return true; + } + + /* Skip backwards the gap or trailing data ending at pos. 'ignore_gaps' also ignores format errors and a truncated last member. If successful, push member preceding gap and set pos to member header. */ -bool Lzip_index::skip_gap( const int fd, long long & pos, +bool Lzip_index::skip_gap( const int fd, unsigned long long & pos, const bool ignore_trailing, const bool loose_trailing, const bool ignore_bad_ds, const bool ignore_gaps ) { - enum { block_size = 16384, - buffer_size = block_size + Lzip_trailer::size - 1 + Lzip_header::size }; - uint8_t buffer[buffer_size]; if( pos < min_member_size ) { - if( pos >= 0 && ignore_gaps && !member_vector.empty() ) - { pos = 0; return true; } + if( ignore_gaps && !member_vector.empty() ) { pos = 0; return true; } return false; } + enum { block_size = 16384, + buffer_size = block_size + Lzip_trailer::size - 1 + Lzip_header::size }; + uint8_t buffer[buffer_size]; int bsize = pos % block_size; // total bytes in buffer if( bsize <= buffer_size - block_size ) bsize += block_size; int search_size = bsize; // bytes to search for trailer @@ -91,47 +121,55 @@ bool Lzip_index::skip_gap( const int fd, long long & pos, if( member_size > ipos + i || !trailer.verify_consistency() ) continue; Lzip_header header; - if( seek_read( fd, header.data, Lzip_header::size, - ipos + i - member_size ) != Lzip_header::size ) - { set_errno_error( "Error reading member header: " ); return false; } - const unsigned dictionary_size = header.dictionary_size(); - if( !header.verify_magic() || !header.verify_version() || - ( !ignore_bad_ds && !isvalid_ds( dictionary_size ) ) ) continue; - if( member_vector.empty() ) // trailing data or truncated member + if( !read_header( fd, header, ipos + i - member_size ) ) return false; + if( !header.verify( ignore_bad_ds ) ) continue; + const Lzip_header & header2 = *(const Lzip_header *)( buffer + i ); + const bool full_h2 = bsize - i >= Lzip_header::size; + if( header2.verify_prefix( bsize - i ) ) // next header { - const Lzip_header & last_header = *(const Lzip_header *)( buffer + i ); - if( last_header.verify_prefix( bsize - i ) ) + if( !ignore_gaps && member_vector.empty() ) // last member { - if( !ignore_gaps ) - { error_ = "Last member in input file is truncated or corrupt."; - retval_ = 2; return false; } - const unsigned dictionary_size = - ( bsize - i >= Lzip_header::size ) ? - last_header.dictionary_size() : 0; - const unsigned long long member_size = pos - ( ipos + i ); - pos = ipos + i; - member_vector.push_back( Member( 0, 0, pos, - member_size, dictionary_size ) ); - return true; + if( !full_h2 ) error_ = "Last member in input file is truncated."; + else if( !check_header_error( header2, ignore_bad_ds ) ) + error_ = "Last member in input file is truncated or corrupt."; + retval_ = 2; return false; } + const unsigned dictionary_size = + full_h2 ? header2.dictionary_size() : 0; + const unsigned long long member_size = pos - ( ipos + i ); + pos = ipos + i; + // approximate data and member sizes for '-i -D' + member_vector.push_back( Member( 0, member_size, pos, + member_size, dictionary_size ) ); } if( !ignore_gaps && member_vector.empty() ) { - if( !loose_trailing && bsize - i >= Lzip_header::size && - (*(const Lzip_header *)( buffer + i )).verify_corrupt() ) + if( !loose_trailing && full_h2 && header2.verify_corrupt() ) { error_ = corrupt_mm_msg; retval_ = 2; return false; } if( !ignore_trailing ) { error_ = trailing_msg; retval_ = 2; return false; } } pos = ipos + i - member_size; + const unsigned dictionary_size = header.dictionary_size(); member_vector.push_back( Member( 0, trailer.data_size(), pos, member_size, dictionary_size ) ); + if( dictionary_size_ < dictionary_size ) + dictionary_size_ = dictionary_size; return true; } - if( ipos <= 0 ) - { if( ignore_gaps && !member_vector.empty() ) { pos = 0; return true; } - set_num_error( "Bad trailer at pos ", pos - Lzip_trailer::size ); - return false; } + if( ipos == 0 ) + { + if( ignore_gaps && !member_vector.empty() ) + { + const Lzip_header * header = (const Lzip_header *)buffer; + const unsigned dictionary_size = header->dictionary_size(); + // approximate data and member sizes for '-i -D' + member_vector.push_back( Member( 0, pos, 0, pos, dictionary_size ) ); + pos = 0; return true; + } + set_num_error( "Bad trailer at pos ", pos - Lzip_trailer::size ); + return false; + } bsize = buffer_size; search_size = bsize - Lzip_header::size; rd_size = block_size; @@ -144,7 +182,7 @@ bool Lzip_index::skip_gap( const int fd, long long & pos, Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, const bool loose_trailing, const bool ignore_bad_ds, const bool ignore_gaps, const long long max_pos ) - : insize( lseek( infd, 0, SEEK_END ) ), retval_( 0 ) + : insize( lseek( infd, 0, SEEK_END ) ), retval_( 0 ), dictionary_size_( 0 ) { if( insize < 0 ) { set_errno_error( "Input file is not seekable: " ); return; } @@ -155,25 +193,20 @@ Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, retval_ = 2; return; } Lzip_header header; - if( seek_read( infd, header.data, Lzip_header::size, 0 ) != Lzip_header::size ) - { set_errno_error( "Error reading member header: " ); return; } - if( !header.verify_magic() ) - { error_ = bad_magic_msg; retval_ = 2; return; } - if( !header.verify_version() ) - { error_ = bad_version( header.version() ); retval_ = 2; return; } - if( !ignore_bad_ds && !isvalid_ds( header.dictionary_size() ) ) - { error_ = bad_dict_msg; retval_ = 2; return; } + if( !read_header( infd, header, 0 ) ) return; + if( check_header_error( header, ignore_bad_ds ) ) return; // pos always points to a header or to ( EOF || max_pos ) - long long pos = ( max_pos > 0 ) ? max_pos : insize; + unsigned long long pos = ( max_pos > 0 ) ? max_pos : insize; while( pos >= min_member_size ) { Lzip_trailer trailer; - if( seek_read( infd, trailer.data, Lzip_trailer::size, - pos - Lzip_trailer::size ) != Lzip_trailer::size ) - { set_errno_error( "Error reading member trailer: " ); break; } + if( !read_trailer( infd, trailer, pos ) ) break; const unsigned long long member_size = trailer.member_size(); - if( member_size > (unsigned long long)pos || !trailer.verify_consistency() ) + // if gaps are being ignored, verify consistency of last trailer only. + if( member_size > pos || member_size < min_member_size || + ( ( !ignore_gaps || member_vector.empty() ) && + !trailer.verify_consistency() ) ) // bad trailer { if( ignore_gaps || member_vector.empty() ) { if( skip_gap( infd, pos, ignore_trailing, loose_trailing, @@ -181,12 +214,8 @@ Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, set_num_error( "Bad trailer at pos ", pos - Lzip_trailer::size ); break; } - if( seek_read( infd, header.data, Lzip_header::size, - pos - member_size ) != Lzip_header::size ) - { set_errno_error( "Error reading member header: " ); break; } - const unsigned dictionary_size = header.dictionary_size(); - if( !header.verify_magic() || !header.verify_version() || - ( !ignore_bad_ds && !isvalid_ds( dictionary_size ) ) ) + if( !read_header( infd, header, pos - member_size ) ) break; + if( !header.verify( ignore_bad_ds ) ) // bad header { if( ignore_gaps || member_vector.empty() ) { if( skip_gap( infd, pos, ignore_trailing, loose_trailing, @@ -195,10 +224,14 @@ Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, break; } pos -= member_size; + const unsigned dictionary_size = header.dictionary_size(); member_vector.push_back( Member( 0, trailer.data_size(), pos, member_size, dictionary_size ) ); + if( dictionary_size_ < dictionary_size ) + dictionary_size_ = dictionary_size; } - if( pos < 0 || pos >= min_member_size || ( pos != 0 && !ignore_gaps ) || + // block at pos == 0 must be a member unless shorter than min_member_size + if( pos >= min_member_size || ( pos != 0 && !ignore_gaps ) || member_vector.empty() ) { member_vector.clear(); @@ -217,6 +250,8 @@ Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, } if( i + 1 >= member_vector.size() ) break; member_vector[i+1].dblock.pos( end ); + if( member_vector[i].mblock.end() > member_vector[i+1].mblock.pos() ) + internal_error( "two mblocks overlap after constructing a Lzip_index." ); } } @@ -224,7 +259,7 @@ Lzip_index::Lzip_index( const int infd, const bool ignore_trailing, // All files in 'infd_vector' must be at least 'fsize' bytes long. Lzip_index::Lzip_index( const std::vector< int > & infd_vector, const long long fsize ) - : insize( fsize ), retval_( 0 ) + : insize( fsize ), retval_( 0 ), dictionary_size_( 0 ) // DS not used { if( insize < 0 ) { set_errno_error( "Input file is not seekable: " ); return; } @@ -240,8 +275,7 @@ Lzip_index::Lzip_index( const std::vector< int > & infd_vector, for( int i = 0; i < files && !done; ++i ) { const int infd = infd_vector[i]; - if( seek_read( infd, header.data, Lzip_header::size, 0 ) != Lzip_header::size ) - { set_errno_error( "Error reading member header: " ); return; } + if( !read_header( infd, header, 0 ) ) return; if( header.verify_magic() && header.verify_version() ) done = true; } if( !done ) @@ -256,17 +290,13 @@ Lzip_index::Lzip_index( const std::vector< int > & infd_vector, for( int it = 0; it < files && !done; ++it ) { const int tfd = infd_vector[it]; - if( seek_read( tfd, trailer.data, Lzip_trailer::size, - pos - Lzip_trailer::size ) != Lzip_trailer::size ) - { set_errno_error( "Error reading member trailer: " ); goto error; } + if( !read_trailer( tfd, trailer, pos ) ) goto error; member_size = trailer.member_size(); if( member_size <= (unsigned long long)pos && trailer.verify_consistency() ) for( int ih = 0; ih < files && !done; ++ih ) { const int hfd = infd_vector[ih]; - if( seek_read( hfd, header.data, Lzip_header::size, - pos - member_size ) != Lzip_header::size ) - { set_errno_error( "Error reading member header: " ); goto error; } + if( !read_header( hfd, header, pos - member_size ) ) goto error; if( header.verify_magic() && header.verify_version() ) done = true; } } @@ -323,6 +353,6 @@ long Lzip_index::blocks( const bool count_tdata ) const long n = member_vector.size() + ( count_tdata && cdata_size() < file_size() ); if( member_vector.size() && member_vector[0].mblock.pos() > 0 ) ++n; for( unsigned long i = 1; i < member_vector.size(); ++i ) - if( member_vector[i].mblock.pos() > member_vector[i-1].mblock.end() ) ++n; + if( member_vector[i-1].mblock.end() < member_vector[i].mblock.pos() ) ++n; return n; } |