On Tue, 2006-12-19 at 08:17 +0100, Jens Axboe wrote: > On Mon, Dec 18 2006, Ming Zhang wrote: > > On Mon, 2006-12-18 at 10:54 -0500, Alan D. Brunelle wrote: > > > (Hopefully this works better than my last patch attempt - Jens: the > > > other patches got put on the back burner for some work-related stuff, > > > I'll get to those in the New Year.) > > > > > > Alan > > > plain text document attachment (bp-fix-empties) > > > Fix problem where empty input files cause premature parsing of files. > > > > > > From: Alan D. Brunelle <[EMAIL PROTECTED]> > > > > > > Had a problem where CPU 0 opened a file with 0 traces, and that caused the > > > run to end before processing other trace files. > > > --- > > > > > > blkparse.c | 4 +++- > > > 1 files changed, 3 insertions(+), 1 deletions(-) > > > > > > diff --git a/blkparse.c b/blkparse.c > > > index bded2f8..0d2ea12 100644 > > > --- a/blkparse.c > > > +++ b/blkparse.c > > > @@ -2249,8 +2249,10 @@ static int setup_file(struct per_dev_inf > > > > > > snprintf(pci->fname + len, sizeof(pci->fname)-1-len, > > > "%s.blktrace.%d", pdi->name, pci->cpu); > > > - if (stat(pci->fname, &st) < 0 || !st.st_size) > > > + if (stat(pci->fname, &st) < 0) > > > return 0; > > > + if (!st.st_size) > > > + return 1; > > > > > > not picky but what if i do blkparse from a fifo? will this work? > > It can't make things worse, as 0 is the failure condition. Alans change > makes sense, we will just read 0 events from the descriptor. >
yes, not worse, just not fixed. i know this is rarely happen. but if somebody do blkparse -i xyz and xyz is a fifo file (feed data by netcat or any fancy stuff), then setup_file will return 1. then blkparse behave incorrectly. and "blkparse -h" does not explicitly ban the use of fifo. - To unsubscribe from this list: send the line "unsubscribe linux-btrace" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
