On Sun, Mar 17, 2013 at 12:35:20AM +0200, Mikolaj Golub wrote: > Ah, this is a thing I wanted to discuss but forgot! As I understand > the idea of the 'ABI hack' is: if the output buffer is less than the > size of data we have, truncate our data to the last successfully > written kinfo_file structure and return without error. > > In my code I do reset ENOMEM to 0 (see sysctl_kern_proc_filedesc), but > I don't see a way how using sbuf interface I can truncate req->oldidx > to the last successfully written file: sbuf_bcat() (to internal > buffer) may be successful and the error might appear only later, when > draining. I suspect it will break things if I return with a partial > kinfo_file, but haven't come with a solution yet...
A solution I am going to try is to provide maxlen argument to kern_proc_filedesc_out(), and if it is not 0, output files that do not exceed the limit, so sysctl_kern_proc_filedesc would call: kern_proc_filedesc_out(p, &sb, req->oldlen); -- Mikolaj Golub _______________________________________________ freebsd-hackers@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to "freebsd-hackers-unsubscr...@freebsd.org"
