Re: [hlds_linux] Half-Life 1: Engine update released
Gary wrote: Personally, I would make every effort to run 5, because as some others have said, it's not supported anymore, and has more fixes to the kernel (mpsafe net, more linux fixes etc) SRCDS runs fine though. I don't understand the problem here because I can get different results depending on what Linux ABI I choose to install. And (I think) the Linux ports are the same in 4.x or 5.x. Now, obviously there are others here that show having success using 5.x, but I think this is completely fruity that SRCDS works, HLDS doesn't, and now I have to got fetch a server, bring it home and do a major OS for such a problem. -- - m0gely http://quake2.telestream.com/ Q2 | Q3A | Counter-strike ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] All clients crash when changing models
Er speak to someone that actually released model change support? Ook wrote: I had a player changing models, trying different models (I have over 700 player models available on server). After doing this a few minutes, when he changed models, the other players could no longer see the models (yes, they have the models on their client) and it stuck at hgrunt. 10 seconds later, everyone's client crashed at the same time, while the server kept running. They were able to rejoine sucessfully. Is this a known exploit of some sort, where a bad model, or changing models, can cause everyone's client to crash at the same time? ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] 64Bits of Vac2
--On Sunday, August 14, 2005 7:26 PM +0100 Richard Fennell [EMAIL PROTECTED] wrote: Yes this Vac2 issue not running on 64bit is very important to a lot of us, many of us have purchased AMD64 / Opteron CPU's specifically to run HLDS-64bit in an attempt to squze as much performance as possible. There is a definate performance gain and rolling 200 servers back to 32 bit will take best part of a day and leave customers not very happy. Can you quantify the degree of the improvement? Any idea what accounts for the change? Is it the larger number of registers in native mode? Or something else? Is there something in the Steam code that actually benefits from wider words or larger pointers? (Normally fatter data will reduce performance; it reduces the effectiveness of caches because you can't fit as many words in the same amount of cache memory.) (I ask as an aspiring 64-bit developer.) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
Re: [hlds_linux] 64Bits of Vac2
The benchmarks I have run with 64-bit applications running under a 64-bit OS (needed to get the extra eight registers) have all seen a 10-15% improvement in performance. Here is a url to some published results as well, http://enterprise.amd.com/downloadables/Bioinformatics_performance.pdf ScratchMonkey wrote: --On Sunday, August 14, 2005 7:26 PM +0100 Richard Fennell [EMAIL PROTECTED] wrote: Yes this Vac2 issue not running on 64bit is very important to a lot of us, many of us have purchased AMD64 / Opteron CPU's specifically to run HLDS-64bit in an attempt to squze as much performance as possible. There is a definate performance gain and rolling 200 servers back to 32 bit will take best part of a day and leave customers not very happy. Can you quantify the degree of the improvement? Any idea what accounts for the change? Is it the larger number of registers in native mode? Or something else? Is there something in the Steam code that actually benefits from wider words or larger pointers? (Normally fatter data will reduce performance; it reduces the effectiveness of caches because you can't fit as many words in the same amount of cache memory.) (I ask as an aspiring 64-bit developer.) ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
[hlds_linux] Fwd: [HLSW-announce] HLSW 1.0.0.44 released
Small HLSW update to version 1.0.0.44 Changelog - Included BeetlesMod support - Included Battlefield 2 support - Included support for extended GameSpy protocol = BF2 Player display works on server with many players - Some small changes Download - http://www.hlsw.org Best regards Olaf Reusch (HLSW Developer) HLSW - Power is nothing without control Contact: [EMAIL PROTECTED] Homepage: http://www.hlsw.org ___ HLSW-announce mailing list [EMAIL PROTECTED] http://webmail.hlsw.org/mailman/listinfo/hlsw-announce HLSW 1.0.0.44 released :) [HLSW-announce] http://webmail.hlsw.org/mailman/listinfo/hlsw-announce -- Sincerely, Your Neighborhood List Keeper ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
RE: [hlds_linux] Half-Life 1: Engine update released
The problem appears to be a bug in the threading library exposed when two threads wait on two different condition variables at the same time. Depending on your OS and particular OS version the fix could be either upgrading GLIBC or upgrading your kernel (the 2.6.x series in Linux has integrated some threading support directly into the kernel, see http://en.wikipedia.org/wiki/Native_POSIX_Thread_Library and http://linuxdevices.com/articles/AT6753699732.html ). SRCDS doesn't have this issue because it is not running the latest VAC client (it will soon however). We are looking into other solutions (for example rewriting the code to remove the possibility of triggering the library bug) but we have no solid dates for an update. In the meantime you can upgrade your library/kernel versions to those reported to work by others on this list. - Alfred Original Message From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of m0gely Sent: Sunday, August 14, 2005 11:05 PM To: hlds_linux@list.valvesoftware.com Subject: Re: [hlds_linux] Half-Life 1: Engine update released Gary wrote: Personally, I would make every effort to run 5, because as some others have said, it's not supported anymore, and has more fixes to the kernel (mpsafe net, more linux fixes etc) SRCDS runs fine though. I don't understand the problem here because I can get different results depending on what Linux ABI I choose to install. And (I think) the Linux ports are the same in 4.x or 5.x. Now, obviously there are others here that show having success using 5.x, but I think this is completely fruity that SRCDS works, HLDS doesn't, and now I have to got fetch a server, bring it home and do a major OS for such a problem. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
[hlds_linux] Whats up at valve
seems like any new servers tonight are not updateing are wont run what happend? To subscribe or unsubscribe via the World Wide Web, visit http://list.valvesoftware.com/mailman/listinfo/hlds_linux or, via email, send a message with subject or body 'help' to [EMAIL PROTECTED] You can reach the person managing the list at [EMAIL PROTECTED] When replying, please edit your Subject line so it is more specific than Re: Contents of hlds_linux digest... Today's Topics: 1. New/Old Cheat/Hack (Husayn ibn al-Samarqandi) 2. Re: Half-Life 1: Engine update released (James D) 3. Steam Ticket Expired (Kennycom) 4. New/old cheat/hack ([EMAIL PROTECTED]) 5. Re: Half-Life 1: Engine update released (m0gely) 6. Re: All clients crash when changing models (James Tucker) 7. Re: 64Bits of Vac2 (ScratchMonkey) --__--__-- Message: 1 Date: Sun, 14 Aug 2005 18:08:11 -0500 (EST) From: Husayn ibn al-Samarqandi [EMAIL PROTECTED] To: hlds_linux@list.valvesoftware.com Subject: [hlds_linux] New/Old Cheat/Hack Reply-To: hlds_linux@list.valvesoftware.com Folks: I run a DoD server. Just got the following from one of my admins. Comments? Insane Husayn -- Forwarded message -- The old spec hax seems to be back in some form or another Yesterday Christman and I saw one. He joined the game... picked a team as normal then quickly killed himslef and joined a new team that was not in the game. Take a look: L 08/13/2005 - 21:23:26: roflcakes704STEAM_0:0:2112501Allies joined team 19IVoiceGameMgrHelper He then could drop out of spec mode anywhere on the map and kill. He looked like an Axis and had the Allied Bar... When Christman saw him he thought he was a Axis so he let him go and then was killed by him. --__--__-- Message: 2 Date: Sun, 14 Aug 2005 23:51:42 -0500 From: James D [EMAIL PROTECTED] To: hlds_linux@list.valvesoftware.com Subject: Re: [hlds_linux] Half-Life 1: Engine update released Reply-To: hlds_linux@list.valvesoftware.com I can't even get the steam update script to work. I'm a Linux newbie so updating glibc scares the hell out of me. Not to mention the fact that I wouldn't know how to do it anyway. lol -James Joachim Sehlstedt wrote: Hi, What I did find out is that the server will only crash when running glibc 2.3.2 on Linux kernel 2.6.12.1, running kernel 2.6.8 works without no problem. I have not tested the kernels between those two. Upgrading to glibc 2.3.5 on kernel 2.6.12.1 made the problem go away. I've also talked a bit with Alfred off-list and he confirmed the problem on my server and ran some debugging on it. I've also gave him access to another server we have here so he could debug it better, but I guess he left work before reading my last reply. He also told me that he might have a solution for it. If you are running Linux and have this problem just upgrade the glibc to 2.3.5 and you should be all right. greetings Joachim Sehlstedt Gary wrote: Here's what I've found for BSD users (linux users might want to read). I believe glibc 2.3.2 and 2.3.3 are buggy (and old, 2.3.2 is from 2001 era).. VERY old glibc's, linux_base is redhat-6 :) [EMAIL PROTECTED] egrep -r glibc-2.2 linux_* linux_base/Makefile: glibc-2.2.5-44.legacy.3.i386.rpm:legacy \ linux_base-src-rh-7.3/Makefile: glibc-2.2.5-44.legacy.3.src.rpm:legacy \ Somewhat newer glibc's: [EMAIL PROTECTED] egrep -r glibc-2.3 linux_* linux_base-8/Makefile: glibc-2.3.2-4.80.8.${ARCH}.rpm:new \ linux_base-rh-9/Makefile:DISTFILES= glibc-2.3.2-27.9.7.i386.rpm:new \ linux_base-src-rh-8/Makefile: glibc-2.3.2-4.80.8.src.rpm:new \ linux_base-suse-9.1/Makefile: glibc-2.3.3-98.i586.rpm \ linux_base-suse-9.2/Makefile: glibc-2.3.3-118.i586.rpm \ linux_base-suse-9.3/files/rpmlist.conf:glibc|glibc-2.3.4-23.2.i586.rpm||on I removed linux_base-8 (glibc 2.3.2) and used the suse 9.3 base (glibc 2.3.4) and I don't have any problems. If you chose to upgrade glibc (older linux versions like redhat 8 do not have updated glibc packages so you might be out of luck), so in effect, glibc versions 2.3.4 aren't recommended for newer hl1.6 binaries. -G At 01:31 AM 8/13/2005, Joachim Sehlstedt wrote: Hi! After this release all my servers just crashes, it happens right after all cvars are printed to the console for logging. Tried removing the metamod and just running a plain server, same problem. Server is fully updated and verified with -verify_all, operating system is Linux 2.6.x.y and glibc 2.3.2. Servers are started with this argument: ./hlds_run -autoupdate -debug -game cstrike -maxplayers 17 +map de_dust2 +ip XXX.XXX.XXX.XXX. This is what I can see in the console: L 08/13/2005 - 07:26:32: Server cvars end L 08/13/2005 - 07:26:33: Server cvar sv_maxspeed = 900 L 08/13/2005 - 07:26:33: Server cvar pausable = 0 ./hlds_run: line 424: 11837 Segmentation fault (core dumped) $HL_CMD /game_cs_4/core.11837 is not a core dump: File