This is the mail archive of the glibc-bugs@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug libc/16060] New: dprintf fails to be async-signal safe


https://sourceware.org/bugzilla/show_bug.cgi?id=16060

            Bug ID: 16060
           Summary: dprintf fails to be async-signal safe
           Product: glibc
           Version: unspecified
            Status: NEW
          Severity: enhancement
          Priority: P2
         Component: libc
          Assignee: unassigned at sourceware dot org
          Reporter: bugdal at aerifal dot cx
                CC: drepper.fsp at gmail dot com

This is NOT a conformance bug, as POSIX does not require dprintf to be AS-safe
or even mention it as an option. However, Roland McGrath has stated on
libc-alpha that, as the original inventor of this interface, he intended it to
be AS-safe and in general not to have unnecessary failure cases. At present,
dprintf fails to be AS-safe for at least the following reasons:

1. It adds and removes the temporary FILE structure it creates to/from the
global open streams list, requiring a lock. This operation is entirely
unnecessary and nonsensical (and also documented in bug #12847).

2. The printf core uses malloc in various places (wide string conversion, i18n
%n$ arguments, ...). Fixing these uses is less trivial, and may not matter to
the most common usage cases for dprintf which are unlikely to hit the code
paths that use malloc, but they are bugs in themselves anyway (unnecessary
failure cases for the entire printf family) which should be fixed.

There may also be other AS-safety issues I am unaware of.

-- 
You are receiving this mail because:
You are on the CC list for the bug.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]