| Summary: | Mplayer crashes on Vesa | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Francisco León <fjleon> |
| Component: | Current packages | Assignee: | Martin Schlemmer (RETIRED) <azarah> |
| Status: | RESOLVED NEEDINFO | ||
| Severity: | normal | CC: | media-video |
| Priority: | High | ||
| Version: | 1.4_rc1 | ||
| Hardware: | x86 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
What version of mplayer ? I had the same issue. It was my kernel not supporting vm86old() system call ... I have not checked yet, but with rc3 and 2.5.59 it works just fine again .... Interesting, since my strace reports the problem with vm86old() I am using rc3 and the latest "stable" gentoo-sources kernel (2.4.19-r10) I guess i will try vanilla 2.4.20 Please contact a gentoo kernel developer for more info on this. I will contact the mplayer list for more info. I believe gentoo sources 2.4.18 worked This is getting more interesting by the minute. According to a mdk9 user in the mplayer mailing list, this is a glibc bug: mdk9 + glibc-2.2.5 => segfault mdk9 + glibc-2.2.4 => it works Well, we are both using 2.3.1 right? So let's hope this is a broken kernel issue. apparently, this is a problem between mplayer and glibc >=2.2.5 the guy who tried both 2.2.4 and 2.2.5 tried kernels 2.4.18, .19 and .20 what glibc are you using? 2.3.1. I check the source .. it should use the system call '113' if I remmeber, and that should in turn call the correct vm86*() function. so, 2.3.1 with kernel 2.5 works but with 2.4.19 does not? I dont know who to blame... I'm still waiting for the developers to answer me. I also found the code you mentioned. I am tempted to take vm86.h from glibc 2.2.4 and try it out And? I keep forgetting to try at work with a 2.4.20 and glibc-2.2.5 setup. vm86.h from both the sys and the asm dir are exactly the same in 2.2.4 and 2.3.1 (i took a rpm package i found and compared those 2 files) I wouldn't know how to install 2.2.4 completely without screwing my system. Anyway, mplayer needs to port the code to the new (?) vm86() function so i think we can't do much ourselves is this still an issue ? i dont have gentoo on my machine anymore but i guess it still is an issue since it happens with recent kernels/glibc versions. if you can try it out yourself else just close this closing as it's unknown state |
Since gentoo 1.4 i havent been able to use mplayer with VESA output (to play movies without X) To do that, you do: mplayer -vo vesa <movie> All i get is "killed" If i use strace, i get: write(1, "VDec: vo config request - 352 x "..., 65VDec: vo config request - 352 x 288 (preferred csp: Planar YV12) ) = 65 write(1, "VDec: using Planar YV12 as outpu"..., 45VDec: using Planar YV12 as output csp (no 0) ) = 45 write(1, "Movie-Aspect is undefined - no p"..., 51Movie-Aspect is undefined - no prescaling applied.) = 51 write(1, "VO: [vesa] 352x288 => 352x288 Pl"..., 48VO: [vesa] 352x288 => 352x288 Planar YV12 [fs] ) = 48 open("/dev/zero", O_RDONLY|O_LARGEFILE) = 10 mmap2(0x10000, 65536, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE|MAP_FIXED, 10, 0) = 0x10000 open("/dev/mem", O_RDWR|O_LARGEFILE) = 11 mmap2(NULL, 1282, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE|MAP_FIXED, 11, 0) = 0 mmap2(0xa0000, 393216, PROT_READ|PROT_WRITE, MAP_SHARED|MAP_FIXED, 11, 0xa0) = 0xa0000 ioperm(0, 0x400, 0x1) = 0 iopl(0x3) = 0 vm86old(0x83871ac) = -1 ENOSYS (Function not implemented) --- SIGSEGV (Segmentation fault) --- --- SIGSEGV (Segmentation fault) --- --- SIGSEGV (Segmentation fault) --- --- SIGSEGV (Segmentation fault) --- --- SIGSEGV (Segmentation fault) --- --- SIGSEGV (Segmentation fault) --- getpid() = 3143 kill(3143, SIGKILL) = 0 +++ killed by SIGKILL +++ I have been in contact with the mplayer developers on their mailing list, with no luck. I have an nvidia gf4 ti 4200, using vga=788 on lilo.conf Perhaps devfs is making trouble or i have other permission issues? (i am trying as root!) Reproducible: Always Steps to Reproduce: 1.mplayer -vo vesa movie 2. 3. Actual Results: crash Expected Results: no crash i dont have proof that this is a gentoo bug other that it used to work on 1.2