Thread (13 messages) 13 messages, 4 authors, 2020-12-31

Re: [PATCH 1/4] block: introduce submit_bio_noacct_add_head

From: antlists <hidden>
Date: 2020-12-30 17:54:28
Also in: dm-devel, linux-block

On 30/12/2020 17:06, John Stoffel wrote:
Danny> "Provide a way for stacking block device to re-submit

Danny> the bio which should be handled first."

Danny> I will fix it.

Great, though my second question is*why*  it needs to be handled
first?  What is the difference between stacked and un-stacked devices
and how could it be done in a way that doesn't require a seperate
function like this?
Is this anything to do with what's on my mind as a database guy? I've 
heard that things like RAID and LVM have difficulty providing write 
barriers.

I want to be confident that, at EVERY level of the stack, I can put a 
barrier in that guarantees that everything written by user-space BEFORE 
the barrier is handled before anything written AFTER the barrier. That 
way, I can be confident that, after a problem, I know whether I can 
safely roll the log forward or back.

Cheers,
Wol
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help