When I added bfq to Kconfig.iosched, I forgot to add the line that
sets to yes the default value of IOSCHED_BFQ, as it is the case for
the other two I/O schedulers. Such a piece of information is used by
popular build systems, when making kernel configs automatically. So,
the lack of this line for bfq makes it hard and confusing to have bfq
in the kernel generated by these systems.
This commit adds this missing line.
Reported-by: Anders Roxell <[email protected]>
Signed-off-by: Paolo Valente <[email protected]>
---
block/Kconfig.iosched | 1 +
1 file changed, 1 insertion(+)
diff --git a/block/Kconfig.iosched b/block/Kconfig.iosched
index 4626b88b2d5a..01d05ab68322 100644
--- a/block/Kconfig.iosched
+++ b/block/Kconfig.iosched
@@ -20,6 +20,7 @@ config MQ_IOSCHED_KYBER
config IOSCHED_BFQ
tristate "BFQ I/O scheduler"
+ default y
---help---
BFQ I/O scheduler for BLK-MQ. BFQ distributes the bandwidth of
of the device among all processes according to their weights,
--
2.20.1
On 4/13/19 3:16 AM, Paolo Valente wrote:
> When I added bfq to Kconfig.iosched, I forgot to add the line that
> sets to yes the default value of IOSCHED_BFQ, as it is the case for
> the other two I/O schedulers. Such a piece of information is used by
> popular build systems, when making kernel configs automatically. So,
> the lack of this line for bfq makes it hard and confusing to have bfq
> in the kernel generated by these systems.
>
> This commit adds this missing line.
This only makes sense to do for newly introduced symbols. Apart from
that, we should not default to 'y' (or 'm' for that matter) what isn't
strictly needed. Looking at the io schedulers, kyber shouldn't have had
that added.
--
Jens Axboe