lkml.org 
[lkml]   [2005]   [Dec]   [3]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: Linux 2.6.15-rc3 problem found - scsi order changed


On Sun, 4 Dec 2005, Helge Hafting wrote:
>
> With 2.6.15-rc1 and later, I get:
> sda,sdb,sdc,sdd: the slots in my USB card reader
> sde, sdf, sdg: harddisks connected to the sym2 pci host adapter
> sdh, sdi : harddisks connected to mainboard SATA
>
> This kernel have all drivers compiled in - no modules.
>
> So I have to ask - is this change (USB devices before
> any other scsi disks) _intentional_ ?

No. Not only isn't it intentional, it's definitely a bug.

We've had it before. The USB devices should come last.

Now, in general, we can't _guarantee_ ordering, since with modules and
lots of asynchronous events (USB is basically hotplug, we might "detect" a
disk at any time), so at best it's always just going to be very much a
"preferred ordering", but so far we've actually in _practice_ been able to
be very good at keeping the preferred ordering pretty stable.

> I guess mounting by UUID is another way of fixing this? Please tell if
> this change is intentional - it will making mounting scsi disks by device sort
> of useless for anyone with USB though :-/

Can you figure out exactly (or at least slightly more closely) where it
happened? It might be something silly like link ordering changes (like the
fact that drivers/block core files got moved to block/ - although I did
actually try to check that the link order stayed the same there, since
it was such an obvious thing), or it might be more subtle. But even just
pinpointing it to one particular nightly snapshot would be good (and using
"git bisect" to pinpoint it even more closely is obviously even better).

There's a fair number of changes wrt things like platform_device in
the -rc1 series. But I actually wonder if it's that simple Makefile
change.

Does this trivial patch fix it for you? It just makes sure that the usb/
subdirectory gets added to the list of driver subdirectories at the right
point ...

(Link order matters for the "initcall()" ordering, even if it doesn't
matter for anything else)

Linus

----
diff --git a/drivers/Makefile b/drivers/Makefile
index fac1e16..ea410b6 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -5,7 +5,7 @@
# Rewritten to use lists instead of if-statements.
#

-obj-$(CONFIG_PCI) += pci/ usb/
+obj-$(CONFIG_PCI) += pci/
obj-$(CONFIG_PARISC) += parisc/
obj-$(CONFIG_RAPIDIO) += rapidio/
obj-y += video/
@@ -49,6 +49,7 @@ obj-$(CONFIG_ATA_OVER_ETH) += block/aoe/
obj-$(CONFIG_PARIDE) += block/paride/
obj-$(CONFIG_TC) += tc/
obj-$(CONFIG_USB) += usb/
+obj-$(CONFIG_PCI) += usb/
obj-$(CONFIG_USB_GADGET) += usb/gadget/
obj-$(CONFIG_GAMEPORT) += input/gameport/
obj-$(CONFIG_INPUT) += input/
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
\
 
 \ /
  Last update: 2009-11-18 23:46    [W:0.092 / U:1.728 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site