On Thu, 2006-12-21 at 10:49 -0500, Ming Zhang wrote:
<snip>
> >
> > That's fine with me, I'm not suggesting an elaborate scheme. I just want
> > to prevent people accidentally doin foo.blktrace.0 from missing events
> > from CPUs 1...N. So:
> >
> > * - for stdin
> >
> > * 'x' is a fifo, open that.
> >
> > * foo and foo.blktrace.[0...N] exists, open foo.blktrace.[0...N]
> >
> > Up until now, that is no change from what blkparse currently does, I'm
> > just describing it. So the new rule I'm proposing is:
> >
> > * foo.blktrace.0 given, and foo.blktrace.[1...N] exists, print a warning
> > and add those files.
> >
> > If someone really wants to use only that file (obscure case, I cannot
> > imagine a real world scenario where that is the case. The weird HT
> > experimentation case that Seelam gave is easy - just delete the damn
> > file, it wont even be valid anymore since the first file is
> > overwritten), then they can use '-' as input and cat the file.
> >
>
> no warning is given. add silently for 4th.
any comment about this patch?
>
> diff --git a/blkparse.c b/blkparse.c
> index 6bb5308..11e81c6 100644
> --- a/blkparse.c
> +++ b/blkparse.c
> @@ -2307,9 +2307,22 @@ static int handle(struct ms_stream *msp)
> return 1;
> }
>
> +static int name_check(char *name)
> +{
> + char *b;
> +
> + if (!name)
> + return -EINVAL;
> +
> + b = strstr(name, ".blktrace.");
> + if (b)
> + *b = '\0';
> + return 0;
> +}
> +
> static int do_file(void)
> {
> - int i, cpu;
> + int i, cpu, res;
> struct per_dev_info *pdi;
>
> /*
> @@ -2317,6 +2330,9 @@ static int do_file(void)
> */
> for (i = 0; i < ndevices; i++) {
> pdi = &devices[i];
> + res = name_check(pdi->name);
> + if (res)
> + return res;
> for (cpu = 0; setup_file(pdi, cpu); cpu++)
> ;
> }
>
>
-
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
