Have you checked to make sure the right user space end points are
available in /sys?
Does using strace give you any hints?
-Josh
On Apr 2, 2008, at 4:28 PM, Scott Weitzenkamp (sweitzen) wrote:
I tried that, and it didn't work:
[EMAIL PROTECTED] ~]# grep eth /etc/dat.conf
OpenIB-cma u1.2 non
Ok lets forget about the single theaded thing to solve the registration
races. As Andrea pointed out this still has ssues with other subscribed
subsystems (and also try_to_unmap). We could do something like what
stop_machine_run does: First disable all running subsystems before
registering a ne
Thinking about this adventurous locking some more: I think you are
misunderstanding what a seqlock is. It is *not* a spinlock.
The critical read section with the reading of a version before and after
allows you access to a certain version of memory how it is or was some
time ago (caching effect
cheapest online drugstore. Verified by visa. Quality and Flexibility.
http://www.google.de/pagead/iclk?sa=l&ai=TTrAEI&num=35947&adurl=http://zjeT.timeminute.com
___
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/
On Wed, Apr 02, 2008 at 03:34:01PM -0700, Christoph Lameter wrote:
> Still two methods ...
Yes, the invalidate_page is called with the core VM holding a
reference on the page _after_ the tlb flush. The invalidate_end is
called after the page has been freed already and after the tlb
flush. They've
>-Original Message-
>From: [EMAIL PROTECTED]
>[mailto:[EMAIL PROTECTED] On Behalf Of Scott
>Weitzenkamp (sweitzen)
>Sent: Wednesday, April 02, 2008 4:29 PM
>To: Joshua Bernstein
>Cc: OpenFabrics EWG; [ofa_general]
>Subject: [ewg] RE: [ofa-general] how do I use uDAPL with iWARP?
>
>> >
> > I tried that, and it didn't work:
> >
> > [EMAIL PROTECTED] ~]# grep eth /etc/dat.conf
> > OpenIB-cma u1.2 nonthreadsafe default libdaplcma.so.1 dapl.1.2
> > "eth2 0"
> > ""
> > [EMAIL PROTECTED] ~]# dtest
> > 10194 Running as server - OpenIB-cma
> > 10194 Error dat_ep_create: DAT_INVALID_HAN
On Apr 2, 2008, at 4:07 PM, Scott Weitzenkamp (sweitzen) wrote:
I tried that, and it didn't work:
[EMAIL PROTECTED] ~]# grep eth /etc/dat.conf
OpenIB-cma u1.2 nonthreadsafe default libdaplcma.so.1 dapl.1.2
"eth2 0"
""
[EMAIL PROTECTED] ~]# dtest
10194 Running as server - OpenIB-cma
10194 Err
I tried that, and it didn't work:
[EMAIL PROTECTED] ~]# grep eth /etc/dat.conf
OpenIB-cma u1.2 nonthreadsafe default libdaplcma.so.1 dapl.1.2 "eth2 0"
""
[EMAIL PROTECTED] ~]# dtest
10194 Running as server - OpenIB-cma
10194 Error dat_ep_create: DAT_INVALID_HANDLE
10194 Error freeing EP: DAT_INVAL
On Thu, 3 Apr 2008, Andrea Arcangeli wrote:
> I said try_to_unmap_cluster, not get_user_pages.
>
> CPU0CPU1
> try_to_unmap_cluster:
> emm_invalidate_start in EMM (or mmu_notifier_invalidate_range_start in #v10)
> walking the list by hand in EMM (or
Scott,
On Apr 2, 2008, at 4:00 PM, Scott Weitzenkamp (sweitzen) wrote:
I have OFED 1.3 and a Chelsio S310E-SR+ iWARP 10GE NIC. I have
ib_rdma_lat working, so I know IB verbs are working.
How do I use uDAPL, though? All the default /etc/dat.conf entries
have IPoIB or bonding interfaces in
I have OFED 1.3 and a Chelsio S310E-SR+ iWARP 10GE NIC. I have
ib_rdma_lat working, so I know IB verbs are working.
How do I use uDAPL, though? All the default /etc/dat.conf entries have
IPoIB or bonding interfaces in them.
Scott Weitzenkamp
SQA and Release Manager
Data Center Access Engineer
(ankle). Striker Alan Smith was on the bench for United"It's 12 years since I
stood for elected office...and
Check out our impressive line of perfectly crafted fake designer chronometers!
Get amazingly good value for your money!http://werenbebvs.com/
John O'Shea was played. Lille started the ga
This patch updates the copyright date for files modified in 2008.
Signed-off-by: Ralph Campbell <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_common.h|2 +-
drivers/infiniband/hw/ipath/ipath_diag.c |2 +-
drivers/infiniband/hw/ipath/ipath_driver.c|2 +-
drive
From: Dave Olson <[EMAIL PROTECTED]>
This patch adds the initialization calls into the new 7220 HCA files,
changes the Makefile to compile and link the new files, and code
to handle send DMA.
Signed-off-by: Dave Olson <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/Makefile |3
From: Arthur Jones <[EMAIL PROTECTED]>
The patch adds a number of minor changes to support newer HCAs
* New send buffer control bits
* New error condition bits
* Locking and initialization changes
* More send buffers
Signed-off-by: Ralph Campbell <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ip
From: Arthur Jones <[EMAIL PROTECTED]>
A new file which allows the iba7220 send DMA engine
to be used from userland. The routines here are not
linked in yet, that will happen in a follow-on patch...
Signed-off-by: Arthur Jones <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_user_sdma
From: Arthur Jones <[EMAIL PROTECTED]>
A new header file which allows the iba7220 send DMA engine
to be used from userland. The definitions here are not
used yet, that will happen in a follow-on patch...
Signed-off-by: Arthur Jones <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_user
From: John Gregor <[EMAIL PROTECTED]>
The IBA7220 HCA has a new feature to DMA data to the on chip send buffers
instead of or in addition to the host CPU doing the data transfer.
This patch adds code to support the send DMA queue.
Signed-off-by: John Gregor <[EMAIL PROTECTED]>
---
drivers/infin
This patch adds binary data to initialize the IB SERDES.
Signed-off-by: Michael Albaugh <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_sd7220_img.c | 1082
1 files changed, 1082 insertions(+), 0 deletions(-)
diff --git a/drivers/infiniband/hw/ipath/ipath_sd72
From: Michael Albaugh <[EMAIL PROTECTED]>
The control and initialization of the SerDes blocks of the
IBA7220 is sufficiently complex to merit a separate file.
This is that file.
Signed-off-by: Michael Albaugh <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_sd7220.c | 1462
From: Michael Albaugh <[EMAIL PROTECTED]>
This patch adds a new ASIC-specific header file for the HCAs
using the IBA7220.
Signed-off-by: Michael Albaugh <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_7220.h | 57 ++
1 files changed, 57 insertions(+), 0 d
This is part of a patch series to add support for a new HCA.
This patch adds new fields to the header files.
Signed-off-by: Ralph Campbell <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_common.h|3
drivers/infiniband/hw/ipath/ipath_kernel.h| 165 -
This patch makes chip reset more robust and reduces lock contention
between user and kernel TID register updates.
Signed-off-by: Ralph Campbell <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_iba6120.c | 79 -
drivers/infiniband/hw/ipath/ipath_init_chip.c |
From: Dave Olson <[EMAIL PROTECTED]>
Newer HCAs support MSI interrupts and also INTx interrupts.
Fix the code so that INTx can be reliably enabled if MSI interrupts
are not working.
Signed-off-by: Dave Olson <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_driver.c| 23 +++---
From: Dave Olson <[EMAIL PROTECTED]>
Newer HCAs have a threshold counter to reduce the number of DMAs
the chip makes to update the PIO buffer availability status bits.
This patch enables the feature.
Signed-off-by: Dave Olson <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_file_ops.c
From: Dave Olson <[EMAIL PROTECTED]>
Whenever the LID is set, notify the HCA specific code so that
the appropriate HW registers can be updated.
Also log the info on the console at low priority.
Signed-off-by: Dave Olson <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_driver.c |9 +
From: Dave Olson <[EMAIL PROTECTED]>
This patch adds code to enable/disable the IBTA 1.2 heartbeat for testing
if the HCA supports it.
Signed-off-by: Dave Olson <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_common.h |2 ++
drivers/infiniband/hw/ipath/ipath_driver.c | 31 ++
From: Dave Olson <[EMAIL PROTECTED]>
The hardware-based recovery doesn't need any intervention, and in a few
cases we can get a bit confused about state and skip steps such as
turning off the link state LED when we consider recovery to be "down".
So ignore this transition, and either we recover in
Newer HCAs have a HW option to write a sequence number to each receive
queue entry and avoid a separate DMA of the tail register to memory.
This patch adds support for these changes.
Signed-off-by: Ralph Campbell <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_common.h| 31
This patch makes some white space changes and minor non-functional changes
to more closely match the code in OFED-1.3.
Signed-off-by: Ralph Campbell <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_driver.c| 29 ++---
drivers/infiniband/hw/ipath/ipath_init_chip
From: Michael Albaugh <[EMAIL PROTECTED]>
This patch checks for old and new format writes to send a
packet via the diagnostic interface.
Signed-off-by: Michael Albaugh <[EMAIL PROTECTED]>
---
drivers/infiniband/hw/ipath/ipath_diag.c |9 +++--
1 files changed, 7 insertions(+), 2 deletion
The following patches add the remaining changes needed to fully
support the QLogic 7220 DDR HCAs. These also will make 2.6.26
match what is in OFED-1.3 plus some recent minor fixes and code style
clean up.
These can also be pulled into Roland's infiniband.git for-2.6.26 repo using:
git pull git://
On Thu, 3 Apr 2008, Andrea Arcangeli wrote:
> Why can't it fire on the mm_struct where GRU just registered? That
> mm_struct existed way before GRU registered, and VM is free to unmap
> it w/o mmap_sem if there was any memory pressure.
Right. Hmmm... Bad situation. We would have invalidate_start
On Wed, 2 Apr 2008, Andrea Arcangeli wrote:
> + void (*invalidate_page)(struct mmu_notifier *mn,
> + struct mm_struct *mm,
> + unsigned long address);
> +
> + void (*invalidate_range_start)(struct mmu_notifier *mn,
> +
On Wed, Apr 02, 2008 at 03:06:19PM -0700, Christoph Lameter wrote:
> On Thu, 3 Apr 2008, Andrea Arcangeli wrote:
>
> > That would work for #v10 if I remove the invalidate_range_start from
> > try_to_unmap_cluster, it can't work for EMM because you've
> > emm_invalidate_start firing anywhere outsid
On Wed, Apr 02, 2008 at 02:56:25PM -0700, Christoph Lameter wrote:
> I am a bit surprised that brk performance is that important. There may be
I think it's not brk but fork that is being slowed down, did you
oprofile? AIM forks a lot... The write side fast path generating the
overscheduling I gue
On Wed, Apr 02, 2008 at 02:54:52PM -0700, Christoph Lameter wrote:
> On Wed, 2 Apr 2008, Andrea Arcangeli wrote:
>
> > > Hmmm... Okay that is one solution that would just require a BUG_ON in the
> > > registration methods.
> >
> > Perhaps you didn't notice that this solution can't work if you ca
On Thu, 3 Apr 2008, Andrea Arcangeli wrote:
> That would work for #v10 if I remove the invalidate_range_start from
> try_to_unmap_cluster, it can't work for EMM because you've
> emm_invalidate_start firing anywhere outside the context of the
> current task (even regular rmap code, not just nonline
On Wed, 2 Apr 2008, Andrea Arcangeli wrote:
> diff --git a/mm/memory.c b/mm/memory.c
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -1626,9 +1626,10 @@
>*/
> page_table = pte_offset_map_lock(mm, pmd, address,
>
On Wed, Apr 02, 2008 at 02:05:28PM -0700, Christoph Lameter wrote:
> Here is a patch to require single threaded execution during emm_register.
> This also allows an easy implementation of an unregister function and gets
> rid of the races that Andrea worried about.
That would work for #v10 if I r
On Wed, 2 Apr 2008, Andrea Arcangeli wrote:
> paging), hence the slowdown. What you benchmarked is the write side,
> which is also the fast path when the system is heavily CPU bound. I've
> to say aim is a great benchmark to test this regression.
I am a bit surprised that brk performance is that
On Wed, 2 Apr 2008, Andrea Arcangeli wrote:
> > Hmmm... Okay that is one solution that would just require a BUG_ON in the
> > registration methods.
>
> Perhaps you didn't notice that this solution can't work if you call
> range_begin/end not in the "current" context and try_to_unmap_cluster
> do
On Wed, Apr 02, 2008 at 11:15:26AM -0700, Christoph Lameter wrote:
> On Wed, 2 Apr 2008, Andrea Arcangeli wrote:
>
> > On Tue, Apr 01, 2008 at 01:55:36PM -0700, Christoph Lameter wrote:
> > > This results in f.e. the Aim9 brk performance test to got down by
> > > 10-15%.
> >
> > I guess it's m
On Wed, Apr 02, 2008 at 10:59:50AM -0700, Christoph Lameter wrote:
> Did I see #v10? Could you start a new subject when you post please? Do
> not respond to some old message otherwise the threading will be wrong.
I wasn't clear enough, #v10 was in the works... I was thinking about
the last two is
# HG changeset patch
# User Andrea Arcangeli <[EMAIL PROTECTED]>
# Date 1207159058 -7200
# Node ID dd918e267ce1d054e8364a53adcecf3c7439cff4
# Parent 316e5b1e4bf388ef0198c91b3067ed1e4171d7f6
Convert the anon_vma spinlock to a rw semaphore. This allows concurrent
traversal of reverse maps for try_to
# HG changeset patch
# User Andrea Arcangeli <[EMAIL PROTECTED]>
# Date 1207159059 -7200
# Node ID f3f119118b0abd9c4624263ef388dc7230d937fe
# Parent 31fc23193bd039cc595fba1ca149a9715f7d0fb2
This patch adds a lock ordering rule to avoid a potential deadlock when
multiple mmap_sems need to be locked
# HG changeset patch
# User Andrea Arcangeli <[EMAIL PROTECTED]>
# Date 1207159059 -7200
# Node ID 31fc23193bd039cc595fba1ca149a9715f7d0fb2
# Parent dd918e267ce1d054e8364a53adcecf3c7439cff4
XPMEM would have used sys_madvise() except that madvise_dontneed()
returns an -EINVAL if VM_PFNMAP is set, w
# HG changeset patch
# User Andrea Arcangeli <[EMAIL PROTECTED]>
# Date 1207159055 -7200
# Node ID 316e5b1e4bf388ef0198c91b3067ed1e4171d7f6
# Parent 3c3787c496cab1fc590ba3f97e7904bdfaab5375
We no longer abort unmapping in unmap vmas because we can reschedule while
unmapping since we are holding a
# HG changeset patch
# User Andrea Arcangeli <[EMAIL PROTECTED]>
# Date 1207159011 -7200
# Node ID 3c3787c496cab1fc590ba3f97e7904bdfaab5375
# Parent d880c227ddf345f5d577839d36d150c37b653bfd
The conversion to a rwsem allows callbacks during rmap traversal
for files in a non atomic context. A rw sty
# HG changeset patch
# User Andrea Arcangeli <[EMAIL PROTECTED]>
# Date 1207159010 -7200
# Node ID d880c227ddf345f5d577839d36d150c37b653bfd
# Parent fe00cb9deeb31467396370c835cb808f4b85209a
Move the tlb flushing into free_pgtables. The conversion of the locks
taken for reverse map scanning would r
# HG changeset patch
# User Andrea Arcangeli <[EMAIL PROTECTED]>
# Date 1207159010 -7200
# Node ID fe00cb9deeb31467396370c835cb808f4b85209a
# Parent a406c0cc686d0ca94a4d890d661cdfa48cfba09f
Moves all mmu notifier methods outside the PT lock (first and not last
step to make them sleep capable).
Si
# HG changeset patch
# User Andrea Arcangeli <[EMAIL PROTECTED]>
# Date 1207158873 -7200
# Node ID a406c0cc686d0ca94a4d890d661cdfa48cfba09f
# Parent 249e077dc932a5322e04ac1d69326622ea4023b8
Core of mmu notifiers.
Signed-off-by: Andrea Arcangeli <[EMAIL PROTECTED]>
Signed-off-by: Nick Piggin <[EMA
Hello,
this is the mmu notifier #v10. Patches 1 and 2 are the only difference between
this and EMM V2. The rest is the same as with Christoph's patches.
I think maximum priority should be given in merging patch 1 and 2 into -mm and
ASAP in mainline.
Patches from 3 to 8 can go in -mm for testing
On Wed, 2 Apr 2008, Andrea Arcangeli wrote:
> but anyway it's silly to be hardwired to such an interface that worst
> of all requires switch statements instead of proper pointer to
> functions and a fixed set of parameters and retval semantics for all
> methods.
The EMM API with a single callback
On Wed, Apr 02, 2008 at 12:03:50PM -0700, Christoph Lameter wrote:
> + /*
> + * Callback may return a positive value to indicate a
> count
> + * or a negative error code. We keep the first error
> code
> + * but co
Here is a patch to require single threaded execution during emm_register.
This also allows an easy implementation of an unregister function and gets
rid of the races that Andrea worried about.
The approach here is similar to what was used in selinux for security
context changes (see selinux_setpr
On Wed, 2 Apr 2008, Christoph Lameter wrote:
> Here f.e. We can add a special emm_age() function that iterates
> differently and does the | for you.
Well maybe not really necessary. How about this fix? Its likely a problem
to stop callbacks if one callback returned an error.
Subject: EMM: Fix
On Wed, 2 Apr 2008, Andrea Arcangeli wrote:
> On Tue, Apr 01, 2008 at 01:55:36PM -0700, Christoph Lameter wrote:
> > This results in f.e. the Aim9 brk performance test to got down by 10-15%.
>
> I guess it's more likely because of overscheduling for small crtitical
> sections, did you counted t
On Wed, 2 Apr 2008, Andrea Arcangeli wrote:
> There are much bigger issues besides the rcu safety in this patch,
> proper aging of the secondary mmu through access bits set by hardware
> is unfixable with this model (you would need to do age |=
> e->callback), which is the proof of why this isn't
On Tue, Apr 01, 2008 at 01:55:36PM -0700, Christoph Lameter wrote:
> This results in f.e. the Aim9 brk performance test to got down by 10-15%.
I guess it's more likely because of overscheduling for small crtitical
sections, did you counted the total number of context switches? I
guess there will
Since I didn't hear any other votes, I commented out (turned off) the
old openib lists. I can reinstate them if needed. I hope this helps. Thanks.
-jeff
Or Gerlitz wrote:
Jeff Becker wrote:
Hi. Since valid stuff is being rejected, I reset the SPAM filter for
general to its previous setting. A
At 03:39 AM 4/2/2008, Dotan Barak wrote:
>If the value of device_attributes.num_mw is more than zero => the device
>supports memory windows, so i think this flag
>can be safely removed.
I agree with removing the flag, but if you mean "max_mw", looking at the
tree, there are a few problems with th
Yes, it's an iWARP NIC, and the OFED 1.3 perftest ib_rdma_lat program is
working.
Scott
> -Original Message-
> From: Richard Frank [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, April 02, 2008 11:04 AM
> To: Scott Weitzenkamp (sweitzen)
> Cc: [EMAIL PROTECTED]; [ofa_general]
> Subject: Re
RDS does not run over regular 10G NICs - that appear as simple NICS -
this was disabled in 1.3.
For now we are interested in RDS over IWARP NICS - configured as
accessible via the verbs interfaces.
Richard Frank wrote:
is the rds driver loaded (modprobe rds)
Scott Weitzenkamp (sweitzen) wr
I'd like to see netperf comparisions of UDP_STREAM/UDP_RR vs
RDS_STREAM/RDS_RR, does anyone have a patch that will apply cleanly to a
recent netperf?
Scott Weitzenkamp
SQA and Release Manager
Data Center Access Engineering
Cisco Systems
> -Original Message-
> From: Richard Frank [mail
OK - and the conversation was about using NetPerf to compare performance
of RDS to UDP relative to suitability for Oracle use ... so I think
those statements still illustrate my points...
1) NetPerf does not do what Oracle does - and hence is not useful from
Oracle's perspective in comparing U
Just a reminder that we are going to have a BOF for OpenSM Monday the 7th at
6:30pm; room is TBA.
Please come and share your use, experience and desires for OpenSM. Or if you
have yet to try OpenSM, listen in on what others are doing with it.
Thanks,
Ira Weiny
Comp Sci./Math Prog.
Lawrence Liver
Yes, it's loaded, and dmesg says this:
Registered RDS/ib transport
Registered RDS/tcp transport
NET: Registered protocol family 28
Scott
> -Original Message-
> From: Richard Frank [mailto:[EMAIL PROTECTED]
> Sent: Wednesday, April 02, 2008 10:44 AM
> To: Scott Weitzenkamp (sweitzen)
>
Rich,
On Nov 1, 2007, you wrote this to rds-devel:
"Netperf is too simplistic in that all it seems to do is stream data
in a
simple loop. This is not how Oracle uses the IPC and again does not
reflect what it would take to make UDP reliable.
For this reason we are not interested in hav
is the rds driver loaded (modprobe rds)
Scott Weitzenkamp (sweitzen) wrote:
Does't appear to work with Chelsio and OFED 1.3:
[EMAIL PROTECTED] counters]# ethtool -i eth2
driver: cxgb3
version: 1.0-ofed
firmware-version: T 5.0.0 TP 1.1.0
bus-info: :0b:00.0
[EMAIL PROTECTED] counters]# ifconf
Does't appear to work with Chelsio and OFED 1.3:
[EMAIL PROTECTED] counters]# ethtool -i eth2
driver: cxgb3
version: 1.0-ofed
firmware-version: T 5.0.0 TP 1.1.0
bus-info: :0b:00.0
[EMAIL PROTECTED] counters]# ifconfig eth2
eth2 Link encap:Ethernet HWaddr 00:07:43:05:43:9F
inet
I believe there is a patch for NetPerf which supports RDS - although it
may need to be updated - and submitted.
The only prior discussion I can think of - was whether or not NetPerf
exercises RDS as Oracle would.
I'm not proposing that we should enhance NetPerf to do that (but that's
OK with
> Can the maintainer submit RDS patch for mainline kernel, in 2.6.26 or
> 2.6.27 window? It's hard for Distros pick this feature without mainline
> kernel acceptance.
At least as a first order approximation, there is no chance of RDS being
merged for 2.6.26 even if patches appear right this sec
> got it...
Can the maintainer submit RDS patch for mainline kernel, in 2.6.26 or
2.6.27 window? It's hard for Distros pick this feature without mainline
kernel acceptance.
Thanks
Shirley___
general mailing list
general@lists.openfabrics.org
http://
We'd appreciate some feed back on your experience and would like to sort
out any issues ASAP.
Rick
___
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http:/
> WRT to merging RDS into the kernel - our current plans are to wait to
> see RDS adopted by more than Oracle - before approaching the kernel
> community about inclusion of RDS.
I've seen statements before from someone from Oracle that RDS was only
for Oracle's use, for example, that person did
got it...
Roland Dreier wrote:
> Yes, I see this is for pushing RDS upstream - but what about running
> RDS as is over IWARP NICs - that should just work right ?
No idea. It depends on whether you took into account the differences
between IB and iWARP. Anyway that's not really what this thr
> Yes, I see this is for pushing RDS upstream - but what about running
> RDS as is over IWARP NICs - that should just work right ?
No idea. It depends on whether you took into account the differences
between IB and iWARP. Anyway that's not really what this thread was about.
___
> Why do you need the flag IBV_DEVICE_MEM_WINDOW?
> If the value of device_attributes.num_mw is more than zero => the
> device supports memory windows, so i think this flag
> can be safely removed.
OK, I'll delete it from the libibverbs changes. I guess we can kill it
on the kernel side too.
WRT to merging RDS into the kernel - our current plans are to wait to
see RDS adopted by more than Oracle - before approaching the kernel
community about inclusion of RDS.
Roland Dreier wrote:
> What is the work we need to do here - I was thinking RDS should just work ?
Stuff doesn't get mer
> Since send with immediate and send with invalidate are mutually
> exclusive, wouldn't it make sense to use a union for both the immediate
> value and the invalidated rkey?
maybe although that would be hard to do in libibverbs without changing
the API. I'm not a big fan of anonymous unions an
> We want to add send with invalidate & mask compare and swap.
> Eli will be able to send the patches next week and since they are
> small I think they can be in for 2.6.26
Send with invalidate should be OK. Let's see about the masked atomics
stuff -- we have a ton of new verbs and I think we
Yes, I see this is for pushing RDS upstream - but what about running RDS
as is over IWARP NICs - that should just work right ?
Roland Dreier wrote:
> What is the work we need to do here - I was thinking RDS should just work ?
Stuff doesn't get merged into the kernel on its own. If you want R
> What is the work we need to do here - I was thinking RDS should just work ?
Stuff doesn't get merged into the kernel on its own. If you want RDS
upstream then the first step is to post patches in a form suitable for
reviewing. Then respond to the review comments.
The files Documentation/Subm
What is the work we need to do here - I was thinking RDS should just work ?
Roland Dreier wrote:
> What's the status of RDS?
I've never seen any patches. I guess ask the RDS guys if/when they want
to start working on getting RDS merged.
- R.
___
g
> - if (unlikely(wr->wr.ud.hlen) > 60)
> + if (unlikely(halign > 64))
heh, just noticed that we used to have
unlikely(wr->wr.ud.hlen)
compared to 60. So the annotation was messed up :)
- R.
___
general mailing list
general@list
> - unsigned halign = ALIGN(wr->wr.ud.hlen, 16);
> + unsigned halign = ALIGN(wr->wr.ud.hlen + sizeof *wqe, 16);
>
>
> - if (unlikely(wr->wr.ud.hlen) > 60)
> + if (unlikely(halign > 64))
Sure, makes sense.
___
general maili
On Wed, 2008-04-02 at 08:26 -0700, Roland Dreier wrote:
> Not sure I follow. Given that we have
>
> struct mlx4_lso_seg {
> __be32 mss_hdr_size;
> __be32 header[0];
> };
>
> I don't see much difference between my proposal
>
> > halign = ALIGN(w
> What's the status of RDS?
I've never seen any patches. I guess ask the RDS guys if/when they want
to start working on getting RDS merged.
- R.
___
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listi
Not sure I follow. Given that we have
struct mlx4_lso_seg {
__be32 mss_hdr_size;
__be32 header[0];
};
I don't see much difference between my proposal
> halign = ALIGN(wr->wr.ud.hlen + sizeof *wqe, 16);
and yours
> halign = ALIGN(wr
> - (flags & MLX4_IB_QP_LSO) ? 64 : 0;
> + ((flags & MLX4_IB_QP_LSO) ? 64 : 0);
Ugh, thanks, I've rolled that up into the patch. Sorry for messing
things up...
___
general mailing list
general@lists.openfabr
What's the status of RDS?
Thanks
Shirley
___
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
I must have missed v10. Could you repost so I can build xpmem
against it to see how it operates? To help reduce confusion, you should
probably comandeer the patches from Christoph's set which you think are
needed to make it sleep.
Thanks,
Robin
On Wed, Apr 02, 2008 at 01:16:51PM +0200, Andrea
Control you sexual power.
Buy solution online! http://fenybelloweg.com
___
general mailing list
general@lists.openfabrics.org
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
To unsubscribe, please visit http://openib.org/mailman/listinfo
mlx4: make firmware diagnostic counters available via sysfs.
Developed by: Gabi Liron of Mellanox.
Signed-off-by: Jack Morgenstein <[EMAIL PROTECTED]>
---
Roland,
Please queue this up for kernel 2.6.26.
Thanks!
Jack
Index: infiniband/drivers/net/mlx4/fw.c
==
Roland Dreier wrote:
Core:
- I did a bunch of cleanups all over drivers/infiniband and the
gcc and sparse warning noise is down to a pretty reasonable level.
Further cleanups welcome of course.
We want to add send with invalidate & mask compare and swap.
Eli will be able to send the p
WRT fb9fbf7cc5301a914e099d95d8f9a46a34e58aee
Since send with immediate and send with invalidate are mutually
exclusive, wouldn't it make sense to use a union for both the immediate
value and the invalidated rkey?
Also it seems like this commit touches code in both ib core and in hw
drivers.
__
On Tue, 2008-04-01 at 12:59 -0700, Roland Dreier wrote:
> > + halign = ALIGN(wr->wr.ud.hlen, 16);
>
> This doesn't seem connected to the problem I see, but is this correct?
> Suppose hlen is 48... then halign will be 48 but it really should be
> 64 I think. Do we really want
>
> halign
Gold and Silver prices have been skyrocketing
The recent fall has given you the perfect time to get in and ride gold to $2 for the past week as the undiscovered gem of Gold Production
Now is the time to get in
Symbol : GSML
The move has started and this stock is definately dollar plus bound
Thi
1 - 100 of 106 matches
Mail list logo
