Messages in this thread | | | Subject | Re: [PATCH v2] x86: fix output of show_stack_log_lvl() | From | Joe Perches <> | Date | Fri, 20 Feb 2015 10:03:24 -0800 |
| |
On Fri, 2015-02-20 at 09:52 -0800, Linus Torvalds wrote: > On Feb 20, 2015 9:40 AM, "Joe Perches" <joe@perches.com> wrote: > > > > There are still a few dozen uses of this pattern: > > > > pr_info("Some message line 1\nNext line: "); > > for (...) > > pr_cont(" part %d", i); > > pr_cont('\n"); > > That, btw, is a buglet anyway. > > We don't really support newlines in the middle of printouts any more. We > used to, but it for deprecated. It doesn't really work with the "printk is > a message packet" model. > > Admittedly neither does the "pr_cont()" model, but pr_cont() is > fundamentally useful, in a way that newlines in the middle are not (they > can always just be split up, while the pr_cont() cannot generally be > combined). > > So we should generally try to get rid of the newline in the middle cases.
True. Also fix the pr_debug/dev_dbg cases like drivers/dma/ppc4xx/adma.c:
static void prep_dma_pq_dbg(int id, dma_addr_t *dst, dma_addr_t *src, unsigned int src_cnt) { int i;
pr_debug("\n%s(%d):\nsrc: ", __func__, id); for (i = 0; i < src_cnt; i++) pr_debug("\t0x%016llx ", src[i]); pr_debug("dst: "); for (i = 0; i < 2; i++) pr_debug("\t0x%016llx ", dst[i]); }
| |