Re: complete boot failure in 4.5-rc1 caused by nvme: make SG_IO support optional

From: Jens Axboe
Date: Tue Feb 09 2016 - 12:14:12 EST


On 02/09/2016 10:12 AM, Christoph Hellwig wrote:
Does this looks reasonable?

---
From 7843fae979df3fc14007735f54cc6bb2f6f66dc5 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@xxxxxx>
Date: Tue, 9 Feb 2016 18:11:32 +0100
Subject: nvme: fix Kconfig description for BLK_DEV_NVME_SCSI

Signed-off-by: Christoph Hellwig <hch@xxxxxx>
---
drivers/nvme/host/Kconfig | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/nvme/host/Kconfig b/drivers/nvme/host/Kconfig
index 59307f8..2e24156 100644
--- a/drivers/nvme/host/Kconfig
+++ b/drivers/nvme/host/Kconfig
@@ -17,8 +17,8 @@ config BLK_DEV_NVME_SCSI
and block devices nodes, as well a a translation for a small
number of selected SCSI commands to NVMe commands to the NVMe
driver. If you don't know what this means you probably want
- to say N here, and if you know what it means you probably
- want to say N as well.
+ to say N here, unless you run a distro that abuses this for
+ stable device names like some OpenSuSE and SLES versions.

Yep, that looks a lot more reasonable to me. Might be wort including that it impacts the mount-by-id on those distros.

--
Jens Axboe