November 2012 - DVDnav-discuss - lists.mplayerhq.hu

[PATCH] Sanitize PTT start offsets
by John Stebbins 05 Oct '13

05 Oct '13

If the start offset points past the last_byte (outside the region read) stop reading PTTs and adjust nr_of_srpts. Fixes Transformers 3 -- John GnuPG fingerprint: D0EC B3DB C372 D1F1 0B01 83F0 49F1 D7B2 60D4 D0F7

3 5

[PATCH] "dvdnav_jump_to_sector" as an alternative to "dvdnav_time_search" (REV 5: combined args to structs)
by gnosygnu 29 Jul '13

29 Jul '13

This revision includes the following changes: 1) combined multiple arguments into structures (See dvdnav_pos_data_t, dvdnav_cell_data_t, dvdnav_jump_args_t) 2) changed multi-line statements to remove leading commas 3) removed tabs/trailing whitespace 4) replaced "discontinuity tmap indexes" approach with simpler "use entry sector of current cell" to handle DVDs with initial dummy cells. Thanks to all who have read through the patch and/or offered comments. As far as I know, there are no other known pending changes/requests. I am hoping that this patch is at a near-final state. Please let me know what else should be done to get this patch accepted. Thanks.

3 19

[PATCH] "dvdnav_jump_to_sector" as an alternative to "dvdnav_time_search" (REV 6: add mode argument)
by gnosygnu 16 Dec '12

16 Dec '12

This revision features one minor change. It adds a new "mode" parameter to "dvdnav_jump_to_sector". As per the included comment: /* NOTE: Mode is currently unimplemented. Only 0 should be passed. */ /* 1 and -1 are for future implementation */ /* 0: Default. Jump to a time which may be either <> time_in_pts_ticks */ /* 1: After. Always jump to a time that is > time_in_pts_ticks */ /* -1: Before. Always jump to a time that is < time_in_pts_ticks */ dvdnav_status_t dvdnav_jump_to_sector_by_time(dvdnav_t *this, uint64_t time_in_pts_ticks, int32_t mode) ... The corresponding .h interface was also updated. My apologies for not implementing the new parameter, but as I detailed in another email, the change looks quite significant. As far as I know, there are no other known pending changes/requests. I am hoping that this patch is at a final state. Please let me know what else should be done to get this patch accepted. Thanks.

4 6

[PATCH] Fix out of array pointer access
by Bryce Harrington 10 Dec '12

10 Dec '12

This patch fixes a segmentation fault hit when reading the DVD 'The Express'. It prevents a read/write beyond end of an array due to using a length value taken from the DVD, which can exceed the allocated size. https://bugs.launchpad.net/ubuntu/+source/libdvdread/+bug/894170 The patch was originally written by rickyrockrat (sorry, I don't have his email address) for 4.1.3. I got the DVD and reproduced the segfault using 4.2.0 and verified the patch stops the segfault from happening. We're not confident this is the best fix though, so are posting it here for review. Signed-off-by: Bryce Harrington <bryce(a)canonical.com> diff -Nrup libdvdread-4.2.0//src/ifo_read.c libdvdread-4.2.0-working//src/ifo_read.c --- libdvdread-4.2.0//src/ifo_read.c 2011-12-13 20:24:16.000000000 -0800 +++ libdvdread-4.2.0-working//src/ifo_read.c 2011-12-13 21:21:53.915978130 -0800 @@ -1082,6 +1082,12 @@ int ifoRead_TT_SRPT(ifo_handle_t *ifofil return 0; } + if(tt_srpt->nr_of_srpts>info_length/sizeof(title_info_t)){ + fprintf(stderr,"libdvdread: data mismatch: info_length (%ld)!= nr_of_srpts (%d). Truncating.\n", + info_length/sizeof(title_info_t),tt_srpt->nr_of_srpts); + tt_srpt->nr_of_srpts=info_length/sizeof(title_info_t); + } + for(i = 0; i < tt_srpt->nr_of_srpts; i++) { B2N_16(tt_srpt->title[i].nr_of_ptts); B2N_16(tt_srpt->title[i].parental_id);

3 3

[PATCH] Fix strict-aliasing warnings
by Reimar Döffinger 09 Dec '12

09 Dec '12

Hello, gcc complains about this both in SVN and Erik's git repo. Using a union is supposedly safe for gcc at least, but I don't know about other compilers. I think the bitfield is simply a bad idea in general... An alternative would be to use memcpy. Index: libdvdnav/dvdnav.c =================================================================== --- libdvdnav/dvdnav.c (revision 1243) +++ libdvdnav/dvdnav.c (working copy) @@ -1162,7 +1161,7 @@ union { user_ops_t ops_struct; uint32_t ops_int; - } ops; + } ops, tmp; ops.ops_int = 0; @@ -1172,10 +1171,12 @@ } pthread_mutex_lock(&this->vm_lock); - ops.ops_int |= *(uint32_t*)&this->pci.pci_gi.vobu_uop_ctl; + ops.ops_struct = this->pci.pci_gi.vobu_uop_ctl; - if(this->vm && this->vm->state.pgc) - ops.ops_int |= *(uint32_t*)&this->vm->state.pgc->prohibited_ops; + if(this->vm && this->vm->state.pgc) { + tmp.ops_struct = this->vm->state.pgc->prohibited_ops; + ops.ops_int |= tmp.ops_int; + } pthread_mutex_unlock(&this->vm_lock); return ops.ops_struct;

2 1

[PATCH] fix C++ compatibility broken in r1208
by Anssi Hannula 09 Dec '12

09 Dec '12

Revision 1208 ("Add dvdnav_program_play & dvdnav_current_title_program") added dvdnav_program_play function to dvdnav.h. However, unlike the other API functions, its first parameter is named 'this' instead of 'self'. 'this' is a C++ keyword and prevents compilation of C++ projects using libdvdnav. Restore C++ compatibility by renaming the parameter to 'self' like in the other functions. -- Anssi Hannula

3 2

[PATCH] Make sure we get nav packets for all cells in multi angle
by Joakim Plate 09 Dec '12

09 Dec '12

Hi, This is a resend of an old patch that we still keep in xbmc's version of libdvdnav. The problem is that libdvdnav jumps over NAV packets when playing multi angle features. I don't remember all the exact details on this patch anymore since it was many years since i tracked it down. But it is very obvious in xbmc since we use nav packet end/start times to correct timestamps. Without the patch we git gaps in nav packets that doesn's exist in the streams. I think i still have a dvd sample that shows the issue. /Joakim

3 5

libdvdnav 4.1.4
by Jean-Jacques Sarton 09 Dec '12

09 Dec '12

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello, On Fedora 14 totem and vlc end both with folowing message: libdvdread: Found 1 VTS's libdvdread: Elapsed time 0 libdvdnav: Language 'en' not found, using '' instead libdvdnav: Menu Languages available: libdvdnav: Language 'en' not found, using '' instead libdvdnav: Menu Languages available: libdvdnav: Language 'en' not found, using '' instead libdvdnav: Menu Languages available: libdvdnav: *** pgci_ut handle is NULL *** totem: /builddir/build/BUILD/libdvdnav-4.1.4/src/vm/vm.c:1772: get_ID: Assertion `pgcit != ((void *)0)' failed. Of course tje line totem:... is to be replaced withz vlc:... according to the used player The dvd contain no menu and contain a privat video. If I rename libdvdnav totem/vlc works. Regards, Jean-Jacques Sarton -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) iEYEARECAAYFAk0h3bAACgkQM9JbiR3CwQsEEgCggN7qJrKQXspNYBJbCGRh237R +joAn1pQaRWHFN+nASDqRbL6Jgj3CM5S =iNgb -----END PGP SIGNATURE-----

4 6

[PATCH] consuming all memory when scanning dvd "Up"
by John Stebbins 09 Dec '12

09 Dec '12

First a disclaimer. This is a big invasive patch. If you don't like it, my feelings won't be hurt. I will be using it in HandBrake. So if you see problems with it, please let me know. Many dvd's are showing up that have titles that have numerous repeated language unit tables and pgc's. The new release "Up" is an example. I believe "Dark Knight" also exhibits this. When scanning such discs with lsdvd (and other programs that scan all titles), libdvdread will consume many GB of data for these repeated elements and gets very slow (e.g hours to read all titles) on OS X due to reading the duplicate data from uncached raw devices. This patch detects duplicates and reference counts pgc's and pgcit's. When a duplicate is detected, a reference count is incremented instead of allocating new memory and re-reading the data. I opted for reference counting instead of simply returning a failure on duplicate detection because it is perfectly valid to have a title with such duplicates. I could easily imagine titles adding duplicates, then navigating around them.

3 5

[PATCH] Only -1 as open() return value indicates error
by Reimar Döffinger 06 Nov '12

06 Nov '12

Hello, mostly purely theoretical, but we might get 0 as fd from open. I used the exact -1 check, because e.g. WinCE will also use other negative value as file descriptors (it basically just uses memory pointers cast to int as far as I know) and I don't see a good reason to check for >= 0 instead. Index: vm/vm.c =================================================================== --- vm/vm.c (revision 1243) +++ vm/vm.c (working copy) @@ -172,7 +172,7 @@ /* Read DVD name */ fd = open(device, O_RDONLY); - if (fd > 0) { + if (fd != -1) { off = lseek( fd, 32 * (off_t) DVD_VIDEO_LB_LEN, SEEK_SET ); if( off == ( 32 * (off_t) DVD_VIDEO_LB_LEN ) ) { off = read( fd, data, DVD_VIDEO_LB_LEN );

2 2