Summary: | ProFTPD 1.3.0 process segfaults after uploadin files larger then +-2.8meg | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | den_RDC <rdc> |
Component: | New packages | Assignee: | Luca Longinotti (RETIRED) <chtekk> |
Status: | RESOLVED WORKSFORME | ||
Severity: | normal | CC: | amd64, humpback, net-ftp |
Priority: | High | ||
Version: | 2006.0 | ||
Hardware: | AMD64 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: | Logs + strace + emerge --info |
Description
den_RDC
2006-07-31 16:46:54 UTC
Created attachment 93161 [details]
Logs + strace + emerge --info
I cannot duplicate this on my unstable amd64 machine. Either the amd64 team has an idea or you will have to update glibc. A backtrace would be useful to locate the problem. You can obtain one by: # emerge gdb if you don't have it already setting FEATURES=nostrip and adding -ggdb to your CFLAGS re-merging proftpd restarting proftpd # gdb proftpd $(pgrep proftpd) type "cont" on the (gdb) prompt upload a file to reproduce the crash type "bt full" and attach the output to the bug. I cannot duplicate this either, works on my amd64 and x86 systems... Try without -ffast-math in the C(XX)FLAGS, that sometimes causes problems... And we need more info... Does it still happen with the latest 1.3.0-r2? What are your GCC/Glibc versions now? (GCC4.1 and Glibc 2.4 are the stable ones for desktop profiles now...) Also we really would need a backtrace to try to solve this... Best regards, CHTEKK. Forgot to mention: I too run Hardened/AMD64 systems, they all run ProFTPd for FTP serving and it works there too. Best regards, CHTEKK. |