|
Bugzilla – Full Text Bug Listing |
| Summary: | flashplayer problem opensuse 13.1 | ||
|---|---|---|---|
| Product: | [openSUSE] openSUSE 13.1 | Reporter: | Forgotten User _3TOh92WgY <forgotten__3TOh92WgY> |
| Component: | Other | Assignee: | Stanislav Brabec <sbrabec> |
| Status: | RESOLVED FIXED | QA Contact: | E-mail List <qa-bugs> |
| Severity: | Normal | ||
| Priority: | P3 - Medium | CC: | coolo, forgotten_B-7XM5iatA, wbauer |
| Version: | Final | ||
| Target Milestone: | --- | ||
| Hardware: | 32bit | ||
| OS: | openSUSE 13.1 | ||
| Whiteboard: | maint:released:sle11-sp3:60225 | ||
| Found By: | --- | Services Priority: | |
| Business Priority: | Blocker: | --- | |
| Marketing QA Status: | --- | IT Deployment: | --- |
|
Description
Forgotten User _3TOh92WgY
2013-12-19 18:14:54 UTC
flashplayer is included in openSUSE 13.1 in the official non-oss repository and works on www.spiegel.de (I tested with version 11.2.202.327-14.1). You can do a 1-click install here: http://software.opensuse.org/package/flash-player or just install through your package manager (YaST or zypper) assuming you have the non-oss repository active. Before you install, remove the downloaded flashplayer you received from Adobe. I cannot reproduce as well. Spiegel videos work correctly for me with openSUSE 13.1 Online Update flash-player. The Flash Player plugin in openSUSE is exactly equal to the one from Adobe. Just the package format is converted to RPM and signed by SUSE. Maybe you have something bad in your installation. Could you reproduce it with a fresh account with no user data? Unfortunately I've also problems with flash player and this independent of Adobe or openSuSE package. Flash player do not start. I tried to start via terminal and got the message WARNING: environment variable G_FILENAME_ENCODING set and it is not UTF-8 Ungültiger Maschinenbefehl I checked this variable and got the output echo $G_FILENAME_ENCODING @locale,UTF-8,ISO-8859-15,CP1252 How did you reproduce this problem? Which browser and which web page? This warning is well known and it is an Adobe fault. G_FILENAME_ENCODING specifies possible encodings of files on the file system, and it is correctly formated: It is a list: @locale: try current locale encoding first (most probably UTF-8) UTF-8: try UTF-8 (it has effect only of your locale is not UTF-8) ISO-8859-15: Most probable file name encoding for old UNIX and Linux systems in Germany. CP1252 Most probable 8-bit file name encoding for old Windows file systems in Germany. This setting makes possible to mount USB flash from nearly any source and don't care about file name encoding in GNOME. Can you reproduce this problem without this setting? e. g. from command line G_FILENAME_ENCODING="" firefox I'm using seamonkey as web browser. If I try to play a video from a website (e. g. youtube) by using flashplayer, there is only an empty window shown. Also the standalone player don't start. I tried to start the standalone player from command line and got the message: WARNING: environment variable G_FILENAME_ENCODING set and it is not UTF-8 Ungültiger Maschinenbefehl Now I tried to start standalone player from command line by typing: G_FILENAME_ENCODING="UTF-8" flashplayer and got only the message " Ungültiger Maschinenbefehl " without warning of environment variable. After that I tried to start seamonkey from command line and got the error message: (process:7515): GLib-CRITICAL **: g_slice_set_config: assertion 'sys_page_size == 0' failed If seamonkey was started from command line and I tried to play a flash video by using flash plugin, I got the following error message at the terminal: ###!!! [Parent][MessageChannel::Call] Error: Channel error: cannot send/recv Last week I installed openSuSE 13.1 on an other old Computer. Here the same problem with the flash player. Only a white empty window where flash player should play the video. Start of seamonkey via terminal show the known error messages: ###!!! [Parent][MessageChannel::InterruptCall] Error: Channel error: cannot send/recv ###!!! [Parent][MessageChannel::Call] Error: Channel error: cannot send/recv ###!!! [Parent][MessageChannel::Call] Error: Channel error: cannot send/recv ###!!! [Parent][MessageChannel::InterruptCall] Error: Channel error: cannot send/recv ###!!! [Parent][MessageChannel::InterruptCall] Error: Channel error: cannot send/recv ###!!! [Parent][MessageChannel::InterruptCall] Error: Channel error: cannot send/recv ###!!! [Parent][MessageChannel::InterruptCall] Error: Channel error: cannot send/recv ###!!! [Parent][MessageChannel::InterruptCall] Error: Channel error: cannot send/recv Btw. recently I installed openSuSE 13.2 on my old computer, the same behave there. I have no idea what the reason could be. On all other machines I installed openSuSE, flash player is working without problems. Someone an idea how to find out why flash player don't work on this machines ? OS: openSuSE 13.1 and 13.2 ( i586 ) Browser: Seamonkey 2.31 Flashplayer: 11.2.202.425 (In reply to Norman Gorlt from comment #6) > Someone an idea how to find out why flash player don't work on this machines > ? Flash-Player _requires_ SSE2 support since version 11. But certain older 32bit CPUs (in particular AMD Athlons and Semprons) do not have that. "Ungültiger Maschinenbefehl" would indicate that this is indeed your problem. "cat /proc/cpuinfo" will show you whether your CPU has SSE2 support or not (the "flags" line should include "sse2" if it is supported). If your CPU does not support it, there's no way to get a current Flash-Player to work, your only option would be to install the ancient version 10 (e.g. download and install the package included in openSUSE 11.4). And since Flash-Player is closed-source, there's nothing openSUSE can do about that. Thanks for the feedback. Reporter's User-Agent is x86_64, but the bug itself is marked as bug for 32bit, so it could be an explanation. There is no fix for this bug. The only thing I can do about it is a a change: ExclusiveArch: %ix86 x86_64 to ExclusiveArch: i686 x86_64 But it cannot be done in maintained products, as it forces users to stick at old vulnerable version. So we can do: - Do it only for Factory - Do it for all products, and release a dummy flash-player package for i586 that will completely remove it. (Well, Adobe support of NPAPI (i. e. Firefox on Linux) is going to expire soon, so we have to do it in future for all versions.) We cannot provide 11.4 version for them. It is vulnerable and insecure. (In reply to Stanislav Brabec from comment #8) > Reporter's User-Agent is x86_64, but the bug itself is marked as bug for > 32bit, so it could be an explanation. Well, the reporter apparently did get it to work with "the latest flash player.rpm from adobe", so probably did not have the same problem. My reply was mainly directed at Norman Gorlt. > There is no fix for this bug. None that I can imagine. > The only thing I can do about it is a a change: > ExclusiveArch: %ix86 x86_64 > to > ExclusiveArch: i686 x86_64 Wouldn't help at all I think. i686 does not imply SSE2 support, those affected CPUs are in fact i686 compatible AFAIK. And, according to http://en.wikipedia.org/wiki/SSE2, SSE2 was introduced by intel with the Pentium4 first, whereas i686 means PentiumPro IIANM. Well, this is bad. => We cannot do it in RPM level. Stephan: Is there any chance for update that behaves differently on SSE2 equipped machines and SSE2 unequipped machines? Does ZYPP have a hardware tag that it could make it possible? This update would need an additional RPM or ZYPP magic: With SSE2: update from flash-player to new version of flash-player (possibly i686, preferably without asking for architecture change). Without SSE2: update from flash-player to flash-player obsoleting package (e. g. flash-player-eol) or completely remove flash-player. I would like to do it together with the next security update. The easy hack solution, check in %pre which fails for SSE2-less machines, would violate contract, as it leave vulnerable packages on SSE2-less machines (and broken, as the current version already does not support SSE2). Wolfgang Bauer: Maybe the original reporter downloaded SSE2 requiring version from Adobe. no zypp/rpm magic - if you prefer no flash over broken flash (you will get bug reports both ways IMO), do that check in %post and rm /usr/lib/browser-plugins/libflashplayer.so (In reply to Wolfgang Bauer from comment #7) > Flash-Player _requires_ SSE2 support since version 11. > But certain older 32bit CPUs (in particular AMD Athlons and Semprons) do not > have that. Thanks for quick answer. One machine is working with an AMD Athlon XP 2500 ( Barton ) , the other one with an AMD Athlon XP 2400 ( Thorton ). Both CPU's do not support SSE2 Now I know the reason. It would never have occurred to me. Tried for a test flash player 10.2 on the XP 2400 machine and yes this version is working there. Of course not useable because of safety reasons. Thank you once again Wolfgang for this important hint. This is an autogenerated message for OBS integration: This bug (856386) was mentioned in https://build.opensuse.org/request/show/281323 Factory:NonFree / flash-player I created an update that deletes both binaries in case of SSE2-less CPU and it writes a message to stderr. I cannot make it fail because transaction could be undone and system could stick with the old version. Package removal is not possible as well because the RPM database is locked by ZYPP trasaction. It is an ugly hack that causes warnings in rpm -V, but it prevents invalid instruction exceptions in the browser. I am afraid that we cannot do anything better. Maybe obsoleting empty package for i586, but it will not help for SSE2-less i686 CPUs anyway. SUSE-SU-2015:0052-1: An update that fixes 9 vulnerabilities is now available. Category: security (important) Bug References: 856386,913057 CVE References: CVE-2015-0301,CVE-2015-0302,CVE-2015-0303,CVE-2015-0304,CVE-2015-0305,CVE-2015-0306,CVE-2015-0307,CVE-2015-0308,CVE-2015-0309 Sources used: openSUSE-SU-2015:0059-1: An update that fixes 9 vulnerabilities is now available. Category: security (important) Bug References: 856386,913057 CVE References: CVE-2015-0301,CVE-2015-0302,CVE-2015-0303,CVE-2015-0304,CVE-2015-0305,CVE-2015-0306,CVE-2015-0307,CVE-2015-0308,CVE-2015-0309 Sources used: SUSE-SU-2015:0062-1: An update that fixes 9 vulnerabilities is now available. Category: security (important) Bug References: 856386,913057 CVE References: CVE-2015-0301,CVE-2015-0302,CVE-2015-0303,CVE-2015-0304,CVE-2015-0305,CVE-2015-0306,CVE-2015-0307,CVE-2015-0308,CVE-2015-0309 Sources used: SUSE Linux Enterprise Desktop 11 SP3 (src): flash-player-11.2.202.429-0.4.1 This is an autogenerated message for OBS integration: This bug (856386) was mentioned in https://build.opensuse.org/request/show/283284 12.3:NonFree / flash-player openSUSE-SU-2015:0174-1: An update that fixes 11 vulnerabilities is now available. Category: security (critical) Bug References: 856386,913057,914333,914463 CVE References: CVE-2015-0301,CVE-2015-0302,CVE-2015-0303,CVE-2015-0304,CVE-2015-0305,CVE-2015-0306,CVE-2015-0307,CVE-2015-0308,CVE-2015-0309,CVE-2015-0310,CVE-2015-0311 Sources used: Note: However the bug is marked as RESOLVED FIXED, it is not really fixed, and it cannot be fixed by us. The fix just completely removes Adobe Flash Player from machines where it cannot work. If somebody still has SSE2 incapable machine, it would be nice to get confirmation that it really “works”, i. e. flash-player package is installed but Flash Player plugin is not available. Installed for test flash-player 11.2.202.440-2-29.1 on the old Athlon XP 2400 machine. Installation was possible without any error message. libflashplayer.so has not been installed in /usr/lib/browser-plugins All other files has been installed. So Flash Player plugin is not available. Unfortunately user don't get to know why flashplayer do not work. A message during installation would be good. Prevention of installation of flash player at such machines would be better. openSUSE-SU-2015:0725-1: An update that fixes 45 vulnerabilities is now available. Category: security (important) Bug References: 856386,901334,905032,907257,909219,913057,914333,914463,922033,927089 CVE References: CVE-2014-0558,CVE-2014-0564,CVE-2014-0569,CVE-2014-0573,CVE-2014-0574,CVE-2014-0576,CVE-2014-0577,CVE-2014-0581,CVE-2014-0582,CVE-2014-0583,CVE-2014-0584,CVE-2014-0585,CVE-2014-0586,CVE-2014-0588,CVE-2014-0589,CVE-2014-0590,CVE-2014-8437,CVE-2014-8438,CVE-2014-8440,CVE-2014-8441,CVE-2014-8442,CVE-2015-0331,CVE-2015-0332,CVE-2015-0346,CVE-2015-0347,CVE-2015-0348,CVE-2015-0349,CVE-2015-0350,CVE-2015-0351,CVE-2015-0352,CVE-2015-0353,CVE-2015-0354,CVE-2015-0355,CVE-2015-0356,CVE-2015-0357,CVE-2015-0358,CVE-2015-0359,CVE-2015-0360,CVE-2015-3038,CVE-2015-3039,CVE-2015-3040,CVE-2015-3041,CVE-2015-3042,CVE-2015-3043,CVE-2015-3044 Sources used: Removing the NEEDINFOs. This doesn't matter anymore I suppose. In fact, it is not fixed, and it cannot be fixed by us in any way. It was just worked around by disabling Flash on pure i586. |