|
From: Subhashish P. <sp...@gm...> - 2014-01-09 14:11:05
|
Hello! I have some theoretical queries. What does "teaching valgrind some syscalls" mean? What are the read and write primitives of valgrind and where are they handled? Forgive me for being such a newbie, in advance. Regards, Subhashish P.S. - Which documentation should provide the essential understanding of how valgrind works, so that those stuff could be worked upon when porting valgrind? |
|
From: Philippe W. <phi...@sk...> - 2014-01-09 21:26:27
|
On Thu, 2014-01-09 at 19:40 +0530, Subhashish Pradhan wrote: > Hello! > > > I have some theoretical queries. What does "teaching valgrind some > syscalls" mean? > > > What are the read and write primitives of valgrind and where are they > handled? Basically, 'teaching a (new) syscall to valgrind' means to write a wrapper for this syscall. The wrapper describes the effect of the syscall (typically, for memcheck, describes what memory is read by the syscall, and what memory is written by the syscall). The easiest to undersatnd all that is to look at some existing wrappers e.g. in coregrind/m_syswrap/syswrap-generic.c Philippe NB: valgrind-developers might have been a better mailing list for this question, as writing a syscall wrapper is more a valgrind dev activity. |
|
From: Subhashish P. <sp...@gm...> - 2014-01-10 11:36:52
|
Thanks so much! I googled a lot but couldn't even find a hint about it. I would take a look into coregrind/m_syswrap/syswrap-generic.c and ask my doubts on #valgrind-dev. Regards, Subhashish On Fri, Jan 10, 2014 at 2:56 AM, Philippe Waroquiers < phi...@sk...> wrote: > On Thu, 2014-01-09 at 19:40 +0530, Subhashish Pradhan wrote: > > Hello! > > > > > > I have some theoretical queries. What does "teaching valgrind some > > syscalls" mean? > > > > > > What are the read and write primitives of valgrind and where are they > > handled? > Basically, 'teaching a (new) syscall to valgrind' means to write a > wrapper for this syscall. The wrapper describes the effect of the > syscall (typically, for memcheck, describes what memory is read > by the syscall, and what memory is written by the syscall). > > The easiest to undersatnd all that is to look at some existing > wrappers e.g. in coregrind/m_syswrap/syswrap-generic.c > > Philippe > > NB: valgrind-developers might have been a better mailing list for this > question, as writing a syscall wrapper is more a valgrind dev activity. > > > |
|
From: Subhashish P. <sp...@gm...> - 2014-01-10 21:47:14
|
Hello again, One last query: Are the kernel traps also implemented there in coregrind or somewhere else? May I have the location like that for syscalls? Regards, Subhashish On Fri, Jan 10, 2014 at 5:06 PM, Subhashish Pradhan <sp...@gm...>wrote: > Thanks so much! I googled a lot but couldn't even find a hint about it. > > I would take a look into coregrind/m_syswrap/syswrap-generic.c and ask my > doubts on #valgrind-dev. > > Regards, > Subhashish > > > On Fri, Jan 10, 2014 at 2:56 AM, Philippe Waroquiers < > phi...@sk...> wrote: > >> On Thu, 2014-01-09 at 19:40 +0530, Subhashish Pradhan wrote: >> > Hello! >> > >> > >> > I have some theoretical queries. What does "teaching valgrind some >> > syscalls" mean? >> > >> > >> > What are the read and write primitives of valgrind and where are they >> > handled? >> Basically, 'teaching a (new) syscall to valgrind' means to write a >> wrapper for this syscall. The wrapper describes the effect of the >> syscall (typically, for memcheck, describes what memory is read >> by the syscall, and what memory is written by the syscall). >> >> The easiest to undersatnd all that is to look at some existing >> wrappers e.g. in coregrind/m_syswrap/syswrap-generic.c >> >> Philippe >> >> NB: valgrind-developers might have been a better mailing list for this >> question, as writing a syscall wrapper is more a valgrind dev activity. >> >> >> > |
|
From: Tom H. <to...@co...> - 2014-01-10 22:30:32
|
On 10/01/14 21:47, Subhashish Pradhan wrote: > One last query: Are the kernel traps also implemented there in coregrind > or somewhere else? What do you mean by "kernel traps" exactly? Tom -- Tom Hughes (to...@co...) http://compton.nu/ |
|
From: Subhashish P. <sp...@gm...> - 2014-01-11 05:34:35
|
Hello again! I asked around and found that it was another name for syscalls in gnumach. Thanks again! Regards, Subhashish On Sat, Jan 11, 2014 at 4:00 AM, Tom Hughes <to...@co...> wrote: > On 10/01/14 21:47, Subhashish Pradhan wrote: > > One last query: Are the kernel traps also implemented there in coregrind >> or somewhere else? >> > > What do you mean by "kernel traps" exactly? > > Tom > > -- > Tom Hughes (to...@co...) > http://compton.nu/ > |