On 06/16/2014 09:16 PM, li...@fixnum.org wrote: > From: Wim Vander Schelden <w...@fixnum.org> > > ---
Missing a Signed-off-by declaration; without that legal hurdle, we are unwilling to take the patch. For more details: http://wiki.qemu.org/Contribute/SubmitAPatch > linux-user/syscall.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 49 insertions(+) Run your patch through scripts/checkpatch.pl to flag many of the complaints below. > > diff --git a/linux-user/syscall.c b/linux-user/syscall.c > index c134c32..f9fed3e 100644 > --- a/linux-user/syscall.c > +++ b/linux-user/syscall.c > @@ -4947,6 +4947,54 @@ int host_to_target_waitstatus(int status) > return status; > } > > +static int open_self_cmdline(void *cpu_env, int fd) > +{ > + int fd_orig = -1; > + bool word_skipped = false; > + > + fd_orig = open("/proc/self/cmdline", O_RDONLY); Why a double space? > + if(fd_orig < 0) { Wrong style - we use space after 'if' > + return fd_orig; > + } > + > + while(true) { and space after 'while' > + if(! word_skipped) { > + // Skip the first string, which is the path to qemu-*-static > instead of the actual command. /* */ comments instead of //; fit 80 columns. > + size_t command_length = strnlen(buf, sizeof(buf)); > + if(command_length != sizeof(buf)) { > + // Null byte found, skip one string > + nb_read -= command_length + 1; > + cp_buf += command_length + 1; > + word_skipped = true; This feels a bit complicated. Why not just use memchr() to locate the first NUL byte, rather than messing with strnlen()? > @@ -5148,6 +5196,7 @@ static int do_open(void *cpu_env, const char *pathname, > int flags, mode_t mode) > { "maps", open_self_maps, is_proc_myself }, > { "stat", open_self_stat, is_proc_myself }, > { "auxv", open_self_auxv, is_proc_myself }, > + { "cmdline", open_self_cmdline, is_proc_myself}, Match the style of the lines nearby (space before '}') -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org
signature.asc
Description: OpenPGP digital signature
