Hi,
Linaro kernels include a patch by Tony Lindgren that adds a printascii() to every printk() call. Currently in linux-linaro-tracking, that's 892a9def (ARM: Make low-level printk work). With this patch, every line is printed twice once you've registered a console driver which is a bit of a pain. Speaking to him, he acknowledges that this hack is no longer needed since we now have earlyprintk, so can this patch be removed from linaro kernels?
Cheers, Javi
PS: Please CC me, I'm not subscribed to the list
In my tree I have this patch on top of llct to get rid of that nasty echo
diff --git a/kernel/printk.c b/kernel/printk.c index 1e75d13..97f950d 100644 --- a/kernel/printk.c +++ b/kernel/printk.c @@ -1547,10 +1547,6 @@ asmlinkage int vprintk_emit(int facility, int level, */ text_len = vscnprintf(text, sizeof(textbuf), fmt, args);
-#ifdef CONFIG_DEBUG_LL - printascii(text); -#endif - /* mark and strip a trailing newline */ if (text_len && text[text_len-1] == '\n') { text_len--;
On 13 June 2013 17:29, Javi Merino javi.merino@arm.com wrote:
On Thu, 2013-06-13 at 10:29 +0100, Javi Merino wrote:
By the looks of things, it's already been removed, and was done so a couple of weeks ago before the Linaro 13.05 release. The current commit which reverts it is... https://git.linaro.org/gitweb?p=kernel/linux-linaro-tracking.git%3Ba=commit%...
On 06/13/2013 03:27 PM, Jon Medhurst (Tixy) wrote:
Right. It has been re-added in the middle of 13.05 cycle when half of this patch (the 2nd of its two hunks) has got into one of the topics.
I haven't noticed that the Android topic reverted this change. Now I've removed the 892a9def from llct-v3.10-misc-fixes topic. So the "Make low-level printk work" commit is now removed from linux-linaro-core-tracking, and will get removed from linux-linaro on its next update (should be tomorrow).
Thanks, Andrey
On 06/13/2013 02:29 AM, Javi Merino wrote:
Yea. As Andy mentioned, we carry a patch to revert that change in the Linaro tree (though occasionally I get forgetful and it gets dropped when migrating to new AOSP releases).
I tried to push the revert w/ Andy's rational to AOSP but the Android devs apparently have a fondness for the patch, and still use it: https://android-review.googlesource.com/#/c/59074/
(My guess is they consider it easier to just keep the patch then to have to remember the boot arguments required when debugging difficult problems.)
thanks -john
linaro-kernel@lists.linaro.org
-
Andrey Konovalov -
Andy Green -
Javi Merino -
John Stultz -
Jon Medhurst (Tixy)