Summary: | media-gfx/inkscape - i686-pc-linux-gnu-g++: Internal error: Killed (program cc1plus) | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Zhivko Asenov <hobot4e> |
Component: | Current packages | Assignee: | The Gentoo Linux Hardened Team <hardened> |
Status: | RESOLVED OBSOLETE | ||
Severity: | normal | CC: | graphics+disabled, mjo |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: | build log |
Description
Zhivko Asenov
2008-08-06 19:30:18 UTC
How much RAM does your system have? top gives: Mem: 506444k total, 499840k used, 6604k free, 64096k buffers Swap: 497972k total, 42284k used, 455688k free, 342676k cached Could you try with MAKEOPTS=-j1 please? I have plenty of experience building inkscape with just 512MB (which I assume your data represent) while running a desktop on the same system and most of the time I would see memory allocation problems when MAKEOPTS was set to >-j1... i tried: MAKEOPTS="-j1" emerge -1va =media-gfx/inkscape-0.46-r2 and i've got the same error. What exactly sais this error message: make[2]: {standard input}:701: Error: suffix or operands invalid for `sub' (In reply to comment #4) > i tried: > MAKEOPTS="-j1" > emerge -1va =media-gfx/inkscape-0.46-r2 > and i've got the same error. > What exactly sais this error message: > make[2]: {standard input}:701: Error: suffix or operands invalid for `sub' What is the output this time? Is it color-profile.cpp that fails again or a different file? (In reply to comment #5) > (In reply to comment #4) > > i tried: > > MAKEOPTS="-j1" > > emerge -1va =media-gfx/inkscape-0.46-r2 > > and i've got the same error. > > What exactly sais this error message: > > make[2]: {standard input}:701: Error: suffix or operands invalid for `sub' > > What is the output this time? Is it color-profile.cpp that fails again or a > different file? > yes it was the same failure. It was managed to be compiled if it is chosen vanilla compile from gcc-config. However I want to know how different ("unsafe") is when I use vanilla instead hardened gcc-configuration? Does dmesg give you any relevant output? (In reply to comment #7) > Does dmesg give you any relevant output? > actually yes. Here a piece of the output: ----------------------------------------------------------------------------- Out of memory: kill process 29241 (cc1plus) score 308200 or a child Killed process 29241 (cc1plus) kdm_greet invoked oom-killer: gfp_mask=0x1201d2, order=0, oomkilladj=0 Pid: 4265, comm: kdm_greet Not tainted 2.6.24-hardened-r3 #3 [<c084d2ce>] oom_kill_process+0x10e/0x120 [<c084d47b>] out_of_memory+0xbb/0x120 [<c084ee2b>] __alloc_pages+0x2bb/0x360 [<c0851b08>] __do_page_cache_readahead+0x108/0x1a0 [<c0851c66>] do_page_cache_readahead+0x46/0x70 [<c084aa8f>] filemap_fault+0x24f/0x380 [<c085a274>] __do_fault+0x64/0x370 [<c085a9d1>] handle_mm_fault+0x211/0x2b0 [<c08166b9>] do_page_fault+0x149/0x640 [<c087a038>] sys_select+0x148/0x1a0 [<c0816570>] do_page_fault+0x0/0x640 [<c0ad64e5>] error_code+0x75/0x80 [<c0ad0000>] xdr_process_buf+0x60/0x1f0 ======================= Mem-info: DMA per-cpu: CPU 0: Hot: hi: 0, btch: 1 usd: 0 Cold: hi: 0, btch: 1 usd: 0 Normal per-cpu: CPU 0: Hot: hi: 186, btch: 31 usd: 178 Cold: hi: 62, btch: 15 usd: 48 Active:60682 inactive:50370 dirty:0 writeback:60 unstable:0 free:1191 slab:2105 mapped:1 pagetables:449 bounce:0 DMA free:2020kB min:88kB low:108kB high:132kB active:4924kB inactive:4576kB present:16256kB pages_scanned:20579 all_unreclaimable? yes lowmem_reserve[]: 0 483 483 483 Normal free:2744kB min:2768kB low:3460kB high:4152kB active:237804kB inactive:196904kB present:495300kB pages_scanned:1008742 all_unreclaimable? yes lowmem_reserve[]: 0 0 0 0 DMA: 1*4kB 2*8kB 1*16kB 20*32kB 17*64kB 2*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 2020kB Normal: 28*4kB 1*8kB 0*16kB 0*32kB 1*64kB 2*128kB 1*256kB 0*512kB 0*1024kB 1*2048kB 0*4096kB = 2744kB Swap cache: add 124721, delete 124657, find 162/184, race 0+0 Free swap = 15128kB Total swap = 497972kB Free swap: 15128kB 128896 pages of RAM 0 pages of HIGHMEM 2351 reserved pages 11689 pages shared 64 pages swap cached 0 pages dirty 30 pages writeback 1 pages mapped 2105 pages slab 449 pages pagetables Out of memory: kill process 29249 (cc1plus) score 142192 or a child Killed process 29249 (cc1plus) cc1plus invoked oom-killer: gfp_mask=0x1280d2, order=0, oomkilladj=0 Pid: 5014, comm: cc1plus Not tainted 2.6.24-hardened-r3 #3 [<c084d2ce>] oom_kill_process+0x10e/0x120 [<c084d47b>] out_of_memory+0xbb/0x120 [<c084ee2b>] __alloc_pages+0x2bb/0x360 [<c085a129>] do_anonymous_page+0x49/0x130 [<c085a8d1>] handle_mm_fault+0x111/0x2b0 [<c085d15f>] expand_stack+0x6f/0xb0 [<c08166b9>] do_page_fault+0x149/0x640 [<c0825be5>] __do_softirq+0x75/0xf0 [<c0816570>] do_page_fault+0x0/0x640 [<c0ad64e5>] error_code+0x75/0x80 [<c0ad0000>] xdr_process_buf+0x60/0x1f0 ======================= I suspect hardware, but I do notice that you have it compile fine with a vanilla toolchain. It's not hardware, its another gcc 3.4.6 SSP on C++ interaction. Similar to bugs 230119 and 230415. Workaround is to use -hardenednossp compiler specs or add "-fno-stack-protector -fno-stack-protector-all" to CXXFLAGS. (In reply to comment #10) > It's not hardware, its another gcc 3.4.6 SSP on C++ interaction. Similar to > bugs 230119 and 230415. Workaround is to use -hardenednossp compiler specs or > add "-fno-stack-protector -fno-stack-protector-all" to CXXFLAGS. > yes this is how we compile some packages from time to time. The question is where can i find more information about the difference in these compile-options. What risks are hidden there if any. The bug is not fixed, leave it to the assignee to close. Seconded. Created attachment 218969 [details]
build log
Notice that 0.46 is not in portage anymore... (In reply to comment #15) > Notice that 0.46 is not in portage anymore... And hardened gcc-3.x is long gone. This is certainly obsolete. |