Re: [openib-general] [PATCH 6/6] iSER Kconfig and Makefile

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 5/11/06, Roland Dreier <[email protected]> wrote:
I fixed up this patch so that it actually hooks into the build (as
below).  (BTW, when sending patches in the future, please make them
apply with "-p1" -- your patches had paths like

    /usr/src/linux-2.6.17-rc3/drivers/infiniband/ulp/iser-x/Kconfig

so I had to manually strip off the /usr/src/)

OK, will do that.


Anyway, with a config like

    CONFIG_SCSI_ISCSI_ATTRS=y
    # CONFIG_ISCSI_TCP is not set
    CONFIG_INFINIBAND_ISER=y

My build fails with a bunch of errors like:

    drivers/built-in.o: In function `iser_comp_error_worker':iser_verbs.c:(.text+0x7d7cd): undefined reference to `iscsi_conn_failure'

and so on.  Is the correct fix for this to add

obj-$(CONFIG_INFINIBAND_ISER)   += libiscsi.o

to drivers/scsi/Makefile?

Indeed since libiscsi does not have a CONFIG_ of its own, you need to
set CONFIG_ISCSI_TCP to have libiscsi being built and then iser links
with it.

So there are two options here: either to set a CONFIG_LIBISCSI and
select it by both CONFIG_ISCSI_TCP and CONFIG_INFINIBAND_ISER or the
approach you were suggesting. I am cc-ing Mike Christie and James
Bottomley on this email, if you guys have a preference, let me know
and i can produce a patch to drivers/scsi/Kconfig and Makefile.

Also, I get the following sparse warning:

    drivers/infiniband/ulp/iser/iser_initiator.c:610:25: error: incompatible types for operation (&)

and the code there does look fishy:

        itt = hdr->itt & ISCSI_ITT_MASK; /* mask out cid and age bits */

hdr->itt is __be32 but ISCSI_ITT_MASK is just (0xfff), so it seems
that there is something wrong, either with the iSCSI endianness
annotation or with the code itself.

What's little wrong here is that hdr->itt is __be32 but it never gets
htonl-ed before placing on the wire so i can't ntohl it back before
AND-ing it with the ITT_MASK. Its only little wrong b/c in a way the
target treats the ITT (Initiator Task Tag) as an opaque tag and hence
it just returns on the command/control response the itt of the
command/control.

At the bottom line, its need to be fixed by converting the itt to
network order in libiscsi and then converting it to host order in iser
before doing the arthimetic AND, so the sparse endianess related
warning will be gone. I  will be able to fix that on Sunday and send
the fixes to linux-scsi and openib.

Thanks,
  Roland

Thanks to you for pointing these issues, also, in drivers/infiniband/Makefile

+obj-$(CONFIG_INFINIBAND_SRP)           += ulp/iser/

should be

+obj-$(CONFIG_INFINIBAND_ISER)           += ulp/iser/

Or.

diff-tree 9120bc6c8b5bdd1f4c85df7a04779ae8faa0c1a5 (from 4161cba09429dae06d249584ee1c7d63c672422c)
Author: Or Gerlitz <[email protected]>
Date:   Thu May 11 10:03:30 2006 +0300

    IB/iser: iSER Kconfig and Makefile

    Kconfig and Makefile for iSER.

    Signed-off-by: Or Gerlitz <[email protected]>
    Signed-off-by: Roland Dreier <[email protected]>

diff --git a/drivers/infiniband/Kconfig b/drivers/infiniband/Kconfig
index ba2d650..69a53d4 100644
--- a/drivers/infiniband/Kconfig
+++ b/drivers/infiniband/Kconfig
@@ -41,4 +41,6 @@ source "drivers/infiniband/ulp/ipoib/Kco

 source "drivers/infiniband/ulp/srp/Kconfig"

+source "drivers/infiniband/ulp/iser/Kconfig"
+
 endmenu
diff --git a/drivers/infiniband/Makefile b/drivers/infiniband/Makefile
index eea2732..abeaf79 100644
--- a/drivers/infiniband/Makefile
+++ b/drivers/infiniband/Makefile
@@ -3,3 +3,4 @@ obj-$(CONFIG_INFINIBAND_MTHCA)          += hw/mt
 obj-$(CONFIG_IPATH_CORE)               += hw/ipath/
 obj-$(CONFIG_INFINIBAND_IPOIB)         += ulp/ipoib/
 obj-$(CONFIG_INFINIBAND_SRP)           += ulp/srp/
+obj-$(CONFIG_INFINIBAND_SRP)           += ulp/iser/
diff --git a/drivers/infiniband/ulp/iser/Kconfig b/drivers/infiniband/ulp/iser/Kconfig
new file mode 100644
index 0000000..fead87d
--- /dev/null
+++ b/drivers/infiniband/ulp/iser/Kconfig
@@ -0,0 +1,11 @@
+config INFINIBAND_ISER
+       tristate "ISCSI RDMA Protocol"
+       depends on INFINIBAND && SCSI
+       select SCSI_ISCSI_ATTRS
+       ---help---
+         Support for the ISCSI RDMA Protocol over InfiniBand.  This
+         allows you to access storage devices that speak ISER/ISCSI
+         over InfiniBand.
+
+         The ISER protocol is defined by IETF.
+         See <http://www.ietf.org/>.
diff --git a/drivers/infiniband/ulp/iser/Makefile b/drivers/infiniband/ulp/iser/Makefile
new file mode 100644
index 0000000..fe6cd15
--- /dev/null
+++ b/drivers/infiniband/ulp/iser/Makefile
@@ -0,0 +1,4 @@
+obj-$(CONFIG_INFINIBAND_ISER)  += ib_iser.o
+
+ib_iser-y                      := iser_verbs.o iser_initiator.o iser_memory.o \
+                                  iscsi_iser.o
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[Index of Archives]     [Kernel Newbies]     [Netfilter]     [Bugtraq]     [Photo]     [Stuff]     [Gimp]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Video 4 Linux]     [Linux for the blind]     [Linux Resources]
  Powered by Linux