Thread (72 messages) 72 messages, 7 authors, 2021-12-01

Re: [PATCH v2 01/20] block: Add a flag for internal commands

From: Bart Van Assche <bvanassche@acm.org>
Date: 2021-11-22 17:38:45

On 11/22/21 12:46 AM, John Garry wrote:
On 19/11/2021 19:57, Bart Van Assche wrote:
quoted
From: Hannes Reinecke <hare@suse.de> >
Some drivers use a single tag space for requests submitted by the block
layer and driver-internal requests. Driver-internal requests will never
pass through the block layer but require a valid tag. This patch adds a
new request flag REQ_INTERNAL.
I'm not sure on the name. Don't we already use term "internal" for 
elevator request tag?
I don't see how any confusion could arise between "internal_tag" and 
"REQ_INTERNAL" since in both cases the context is made clear - either 
the request tag or the request in its entirety.
quoted
to mark such requests and a terminates any
such commands in blk_execute_rq_nowait() with a WARN_ON_ONCE() to signal
such an invalid usage.
 >
FYI, I have been working on a different stream, that allows us to send 
the reserved request through the block layer, as we need it for poll 
mode support. The reason is that we need to send reserved requests on 
specific HW queues, which may be polling. However poll mode support only 
allows us to poll requests with bios, so that's a problem ATM.
Please use REQ_OP_DRV_* without REQ_INTERNAL for requests that need to 
be sent through the block layer.

Thanks,

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