Re: mprotect, SIGBUS, and SIGSEGV

2011-05-16 Thread Petr Salinger
signal is determined by the build environment, but the next part seems to contradict that. Is this just a wording problem (the first sentence missing a not?) or something more subtle? Build environment of used eglibc (MIN_KERNEL_SUPPORTED), but not build environment of your binary. Plain FreeB

Re: mprotect, SIGBUS, and SIGSEGV

2011-05-16 Thread Benjamin Kaduk
On Sun, 15 May 2011, Petr Salinger wrote: signal is determined by the build environment, but the next part seems to contradict that. Is this just a wording problem (the first sentence missing a not?) or something more subtle? Build environment of used eglibc (MIN_KERNEL_SUPPORTED), but not bui

Re: mprotect, SIGBUS, and SIGSEGV

2011-05-15 Thread Petr Salinger
signal is determined by the build environment, but the next part seems to contradict that. Is this just a wording problem (the first sentence missing a not?) or something more subtle? Build environment of used eglibc (MIN_KERNEL_SUPPORTED), but not build environment of your binary. Plain FreeB

Re: mprotect, SIGBUS, and SIGSEGV

2011-05-15 Thread Christoph Egger
Petr Salinger writes: >>> There is no way, signal is determined by eglibc compiled for. >>> Lets assume that your binary is compiled under squeeze. >>> Under squeeze eglibc, it will receive SIGBUS, >>> under sid eglibc it will receive SIGSEGV. >> >> I don't understand this part. Your first sentenc

Re: mprotect, SIGBUS, and SIGSEGV

2011-05-11 Thread David Bremner
On Wed, 11 May 2011 07:10:41 +0200 (CEST), Petr Salinger wrote: > > Build environment of used eglibc (MIN_KERNEL_SUPPORTED), > but not build environment of your binary. > Got it, thanks. I guess there is some good reason why this doesn't change the soname of libc, since it seems to break the

Re: mprotect, SIGBUS, and SIGSEGV

2011-05-10 Thread Petr Salinger
There is no way, signal is determined by eglibc compiled for. Lets assume that your binary is compiled under squeeze. Under squeeze eglibc, it will receive SIGBUS, under sid eglibc it will receive SIGSEGV. I don't understand this part. Your first sentence seems to say the signal is determined by

Re: mprotect, SIGBUS, and SIGSEGV

2011-05-10 Thread David Bremner
On Tue, 10 May 2011 22:13:45 +0200 (CEST), Petr Salinger wrote: > What about catching both signals ? I proposed this to upstream. I'm not sure what they will think about that idea. > There is no way, signal is determined by eglibc compiled for. > Lets assume that your binary is compiled under

Re: mprotect, SIGBUS, and SIGSEGV

2011-05-10 Thread Petr Salinger
While working on getting the racket garbage collector working on kFreeBSD, I (well, really upstream) noticed mprotect was sending SIGBUS. Recently this changed so that it is sending SIGSEGV. I'm not sure what exactly changed, except that catching SIGBUS works on squeeze and catching SIGSEGV works

mprotect, SIGBUS, and SIGSEGV

2011-05-10 Thread David Bremner
While working on getting the racket garbage collector working on kFreeBSD, I (well, really upstream) noticed mprotect was sending SIGBUS. Recently this changed so that it is sending SIGSEGV. I'm not sure what exactly changed, except that catching SIGBUS works on squeeze and catching SIGSEGV works