summaryrefslogtreecommitdiff
path: root/lib/libc
AgeCommit message (Collapse)Author
2002-01-16Use the volatile specifier to fix warnings about variables beingTodd C. Miller
clobbered by longjmp / vfork instead of the gcc "(void)&foo;" hack.
2002-01-15MNAMELEN is 90, not 32; peterw AT ifost.org.auTodd C. Miller
2002-01-14If waitpid() return ECHILD that should not be a fatal error sinceTodd C. Miller
someone else could have waited for the process or zombies could be disabled via SIG_IGN or SA_NOCLDWAIT.
2002-01-12If the user passes in "" as the string to resolve the lstat() willTodd C. Miller
fail anyway so check for that. Also convert "." to "" since that way we avoid the lstat() (which we don't need) and the subsequent chdir() and some dir checks.
2002-01-12document net.inet.icmp.redir{accept,timeout}Eric Jackson
2002-01-11typoJun-ichiro itojun Hagino
From: Thomas Klausner <wiz@danbala.ifoer.tuwien.ac.at>
2002-01-11document new sysctl for vfs.nfs.norsvportNathan Binkert
2002-01-10sigprocmask should return 0 on success, fixes powerpc ntpdate problem.Dale Rahn
2002-01-07If the handler for SIGCHLD is set to SIG_IGN, act as if theTodd C. Miller
SA_NOCLDWAIT (don't create zombies) flag has been specified. This is consistent with most other operating systems and is what XPG4.2 specifies.
2002-01-04system calls are made thru _thread_sys_* with weak symbols of the real nameDale Rahn
pointing to the _thread_sys_ version. ok pval, fgs
2002-01-04longjmp(..., 0) is not really legal, but all other archs I've lookedArtur Grabowski
at make sure that setjmp doesn't return 0 if longjmp is passed a 0. Avoid surprises.
2002-01-03Always have the real entry point to the syscall called _thread_sys_Artur Grabowski
and provide the real name as a weak alias. This is the last piece needed for libc_r on sparc64 (plus debugging).
2002-01-02label abort() from inside library as ILLEGALTheo de Raadt
2002-01-02KNFTheo de Raadt
2002-01-02more pid_t useTheo de Raadt
2002-01-02Grammar and format fixes.Aaron Campbell
2002-01-02Remove extra reference to poll.2. Spotted by form@.Mike Pechkin
millert@ ok;
2002-01-02Move the const qualifier to the right place. deraadt@ okThomas Nordin
2002-01-02Add missing .Os. deraadt@ okThomas Nordin
2001-12-29Two functions can return errors here, not just one.Miod Vallat
2001-12-26no way for this to be 'shared only', duhPeter Valchev
2001-12-15Document SIGTTIN and SIGTTOU now that we deal with them too.Todd C. Miller
2001-12-15Catch SIGTTIN and SIGTTOU too and treat them like SIGTSTP.Todd C. Miller
2001-12-08when strerror() has an Unknown error, also set EINVALTheo de Raadt
2001-12-07Add ERRORS and STANDARDS sectionsTodd C. Miller
2001-12-07o Turn off ECHONL in addition to ECHOTodd C. Miller
o Return NULL of read(2) returns -1 o Add ERRORS and STANDARDS sections
2001-12-07From XPG4.2: read(2) can return EIO if the process is a member ofTodd C. Miller
a background process attempting to read from its controlling terminal, the process is ignoring or blocking the SIGTTIN signal or the process group is orphaned. Verified against reality...
2001-12-07missing commaTodd C. Miller
2001-12-07Instead of restoring the tty mode, restoring the signal handler,Todd C. Miller
unblocking the signal and redelivering it just make all our signal handler interupt system calls and set a flag. We can just deliver the signal at the end right before we would normally return. This solves the SIG_IGN problem nicely and causes readpassphrase() to return when someone hits ^C even if the handler is SIG_IGN.
2001-12-07Document signal handling in readpassphrase(3)Todd C. Miller
2001-12-07Add CAVEATS section pointing out the dangers of closing fds <= 2Todd C. Miller
as found by Markus.
2001-12-06Use TCSAFLUSH not TCSANOW. I was only using TCSANOW during debugging...Todd C. Miller
2001-12-06Fix a long standing annoyance with getpass/readpassphrase. Instead ofTodd C. Miller
blocking SIGINT and SIGTSTP, catch them (along with SIGHUP, SIGQUIT, SIGTERM). We restore the tty mode as needed and then restore the original signal handler and resend the signal. For SIGTSTP, upon return from suspend the user is re-prompted for the passphrase.
2001-12-06alloca cannot check if the allocation is valid. mention the consequences; ↵Theo de Raadt
millert ok
2001-12-05correct an alignment mis-conception for malloc(0) returned regions.Thierry Deval
OK deraadt@
2001-12-05KNF, i need this for what comes next..Theo de Raadt
2001-12-05document how malloc(0) blobbies are stored, and their characterTheo de Raadt
2001-12-03for MAP_ANON, dont mention giving an fd of anything but -1, this never worked.Eric Jackson
2001-11-29chflags(2) can return EPERM for other reasons.Todd C. Miller
From NetBSD via Brian J. Kifiak
2001-11-29getfh(2) can fail with EPERM; Brian J. KifiakTodd C. Miller
2001-11-212nd param is constTheo de Raadt
2001-11-20a few thoughts about longjmp in signal handlersTheo de Raadt
2001-11-20shared onlyPeter Valchev
2001-11-20Put dl* function definitions in libc on ELF, removing the need for libdl.Peter Valchev
These are just stubs, the real functions are resolved by ld.so when linking is performed. ok art
2001-11-15sync with reality; scope identifier comes even without NI_WITHSCOPEID.Jun-ichiro itojun Hagino
2001-11-15make NI_WITHSCOPEID a default (always on), to synchronize with recent 2553bis.Jun-ichiro itojun Hagino
sync with kame.
2001-11-14sync with kame better. open some renaming #defines (ENI_xx).Jun-ichiro itojun Hagino
2001-11-09Typos, siz -> size.Aaron Campbell
2001-11-08My part of help, while i'm here:Mike Pechkin
o) better -mdoc design; o) add section number to .Xr; o) close .Bd; o) start new sentence on a new line; jjbg@, millert@ ok
2001-11-07list syslog_r as safe in signal handlersTheo de Raadt
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-28 15:48:32 +0000