On 7/27/2020 2:44 PM, Jan Kara wrote:
On Mon 27-07-20 10:02:11, Jane Chu wrote:
Hi,
On 7/25/2020 9:24 AM, Coly Li wrote:
It is not simple to make dax_supported() from struct dax_operations
or __generic_fsdax_supported() to return exact failure type right now.
So the simplest fix is to use pr_info() to print all the error messages
inside __generic_fsdax_supported(). Then users may find informative clue
from the kernel message at least.
I happen to notice that some servers set their printk levels at 4 by default
to minimize console messages:
# cat /proc/sys/kernel/printk
4 4 1 7
So I'm wondering if you would consider pr_error() instead of pr_info() ?
I don't think this is a good reason to raise priority of this message -
with this logic applied, all info messages should be raised to error level
because someone may find them useful :). And then people raise printk
loglevel because the kernel is too noisy... Personally I think that
pr_info() is fine because there will be error message about unsupported dax
setup from the filesystem and if sysadmin wishes, (s)he can always lookup
info messages in dmesg.
Okay, sounds like the nature of the error isn't severe enough, and it
isn't rare and random, rather, it's reproducible such that sysadmin can
easily catch when raising the printk level.
thanks,
-jane
Honza
_______________________________________________
Linux-nvdimm mailing list -- [email protected]
To unsubscribe send an email to [email protected]
