Message ID | 1505538646-19191-1-git-send-email-yamada.masahiro@socionext.com |
---|---|
Headers | show |
Series | Sync and consolidate Linux-derived printk, BUILD_BUG, BUG, WARN, etc. | expand |
2017-09-16 14:10 GMT+09:00 Masahiro Yamada <yamada.masahiro@socionext.com>: > > I tested this series with buildman. > > > > Masahiro Yamada (8): > stdio.h: move printf() stuff from <common.h> to <stdio.h> > printk: collect printk stuff into <linux/printk.h> with loglevel > support > treewide: replace with error() with pr_err() > common.h: remove error() > vsprintf.h: include <linux/types.h> > bug.h: sync BUILD_BUG stuff with Linux 4.13 > bug.h: move runtime BUG/WARN macros into <linux/bug.h> > dm: define dev_*() log functions in DM header I am still worried if this series is dismissed. I am being blocked from importing NAND code from Linux due to missing/incompatible Linux-derived macros.
On Wed, Oct 04, 2017 at 02:15:19PM +0900, Masahiro Yamada wrote: > 2017-09-16 14:10 GMT+09:00 Masahiro Yamada <yamada.masahiro@socionext.com>: > > > > I tested this series with buildman. > > > > > > > > Masahiro Yamada (8): > > stdio.h: move printf() stuff from <common.h> to <stdio.h> > > printk: collect printk stuff into <linux/printk.h> with loglevel > > support > > treewide: replace with error() with pr_err() > > common.h: remove error() > > vsprintf.h: include <linux/types.h> > > bug.h: sync BUILD_BUG stuff with Linux 4.13 > > bug.h: move runtime BUG/WARN macros into <linux/bug.h> > > dm: define dev_*() log functions in DM header > > I am still worried if this series is dismissed. > > I am being blocked from importing NAND code from Linux > due to missing/incompatible Linux-derived macros. I am looking at this, but a default LOGLEVEL of 5 is just too high and I'm seeing what's reasonable now. -- Tom
Hi Tom, 2017-10-05 12:06 GMT+09:00 Tom Rini <trini@konsulko.com>: > On Wed, Oct 04, 2017 at 02:15:19PM +0900, Masahiro Yamada wrote: >> 2017-09-16 14:10 GMT+09:00 Masahiro Yamada <yamada.masahiro@socionext.com>: >> > >> > I tested this series with buildman. >> > >> > >> > >> > Masahiro Yamada (8): >> > stdio.h: move printf() stuff from <common.h> to <stdio.h> >> > printk: collect printk stuff into <linux/printk.h> with loglevel >> > support >> > treewide: replace with error() with pr_err() >> > common.h: remove error() >> > vsprintf.h: include <linux/types.h> >> > bug.h: sync BUILD_BUG stuff with Linux 4.13 >> > bug.h: move runtime BUG/WARN macros into <linux/bug.h> >> > dm: define dev_*() log functions in DM header >> >> I am still worried if this series is dismissed. >> >> I am being blocked from importing NAND code from Linux >> due to missing/incompatible Linux-derived macros. > > I am looking at this, but a default LOGLEVEL of 5 is just too high and > I'm seeing what's reasonable now. > No. I set the default to 6, not 5. The reason why I chose 6 was to suppress a bunch of pr_info() from drivers/mtd/nand/, which is a copy of Linux. If you change it to a bigger number, some annoying logs would be displayed after NAND: tag. If you change it to a smaller number, pr_notice() would be suppressed. The Linux doc says: 5 (KERN_NOTICE) normal but significant condition So, I think this is worth printing by default. If you do not like it, you can change it, but please make sure to adjust git-log.
On Thu, Oct 05, 2017 at 12:20:36PM +0900, Masahiro Yamada wrote: > Hi Tom, > > > 2017-10-05 12:06 GMT+09:00 Tom Rini <trini@konsulko.com>: > > On Wed, Oct 04, 2017 at 02:15:19PM +0900, Masahiro Yamada wrote: > >> 2017-09-16 14:10 GMT+09:00 Masahiro Yamada <yamada.masahiro@socionext.com>: > >> > > >> > I tested this series with buildman. > >> > > >> > > >> > > >> > Masahiro Yamada (8): > >> > stdio.h: move printf() stuff from <common.h> to <stdio.h> > >> > printk: collect printk stuff into <linux/printk.h> with loglevel > >> > support > >> > treewide: replace with error() with pr_err() > >> > common.h: remove error() > >> > vsprintf.h: include <linux/types.h> > >> > bug.h: sync BUILD_BUG stuff with Linux 4.13 > >> > bug.h: move runtime BUG/WARN macros into <linux/bug.h> > >> > dm: define dev_*() log functions in DM header > >> > >> I am still worried if this series is dismissed. > >> > >> I am being blocked from importing NAND code from Linux > >> due to missing/incompatible Linux-derived macros. > > > > I am looking at this, but a default LOGLEVEL of 5 is just too high and > > I'm seeing what's reasonable now. > > > > No. I set the default to 6, not 5. > > The reason why I chose 6 was to suppress a bunch of pr_info() > from drivers/mtd/nand/, which is a copy of Linux. > > > If you change it to a bigger number, > some annoying logs would be displayed after NAND: tag. > > If you change it to a smaller number, > pr_notice() would be suppressed. > > > The Linux doc says: > > 5 (KERN_NOTICE) normal but significant condition > > > So, I think this is worth printing by default. > > > If you do not like it, you can change it, > but please make sure to adjust git-log. Yes, sorry. You set it to 6, I had already ruled out 5 as too much. I'm currently looking at 3 globally as otherwise most platforms are growing by 5KiB or more (often much more). And I'm going to do it as a follow-up commit so it's clear. -- Tom