[Xen-devel] [SPAM] Re: [PATCH] netback for upstream

[Jeremy Fitzhardinge <jeremy@xxxxxxxx>]

On 12/17/2010 03:00 AM, Ian Campbell wrote:
> I have rebased the current xen/next-2.6.32 netback stuff onto 2.6.37-rc5
> in a new branch at:
>   git://xenbits.xen.org/people/ianc/linux-2.6.git 
> upstream/dom0/backend/netback
> This builds on my upstream/dom0/backend/core branch. (full pull request
> style details are at the end)
>
> I've added some proper commit messages to the early commits which were
> previously lacking them. In particular 8495d90cf5c3..5f7b85f2899e seem
> to have been Authored by Jeremy (according to the git metadata) but were
> not signed-off. Jeremy is it OK to add your sign-off to those three?

I'm pretty sure they were yours originally and I lost the authorship
when importing from mq -> git.

> Until those early commit messages are done the branch will be subject to
> rebasing.
>
> As it stands the branch won't work because it still needs PageForeign.
> My intention is to replace this dependency with the new m2p overlay
> grant table map stuff added for gntdev.
>
> In the meantime if you need something right now I do have a working (if
> somewhat over-engineered) version which removed the dependency of
> PageForeign by always copying. This is in the
> upstream/dom0/backend/netback-copying branch in the same repository. I
> intend to throw this away when I've got the new m2p overlay stuff ready.

Would it make sense to develop the m2p overlay version on top of the
copying branch and drop the foreign code altogether?  It makes logical
sense to go from simple -> more optimised within one history?

> If you wanted the old PageForeign style thing you could look in
> notupstream/foreign-memory while flagellating yourself.

Thanks, but no.

> So my plan going forward is to:
>       * switch to gnttab_{map,unmap}_refs + m2p overlay stuff.
>       * do a checkpatch.pl sweep over the complete files.
>       * do a "git mv" (or perhaps a suitable "git filter-branch") to
>         move from drivers/xen/netback to drivers/net/xen-netback.
>       * ensure netif.h has a reasonably current description of the ring
>         protocol, I suspect it is missing stuff like the GSO prefix
>         support.
>
> After that I intend to submit to linux-netdev as a single large patch
> for review. When it comes time to merge I'll ask if they are willing to
> take the branch with history.

Yep, seems reasonable.

> Ian.
>
>
> The following changes since commit a8d67d067346b924e4d12d585257e2c33251ea04:
>   Ian Campbell (1):
>         xen/xenbus: fixup checkpatch issues in xenbus_probe*
>
> are available in the git repository at:
>
>   git://xenbits.xen.org/people/ianc/linux-2.6.git 
> upstream/dom0/backend/netback
>
> Bastian Blank (1):
>       xen/netback: Fix null-pointer access in netback_uevent
>
> Christophe Saout (1):
>       xen/netback: use dev_name() instead of removed ->bus_id.
>
> Dongxiao Xu (6):
>       xen/netback: use smart polling instead of event notification
>       xen/netback: Move global/static variables into struct xen_netbk.
>       xen/netback: Introduce a new struct type page_ext.
>       xen/netback: Multiple tasklets support.
>       xen/netback: Use Kernel thread to replace the tasklet.
>       Netback: Set allocated memory to zero from vmalloc.

I think we should drop the smartpoll stuff.  It is still unfixed, and I
have never pushed it into upstream netfront.

Is the tasklet work is independent of that though?

> Ian Campbell (21):
>       netback: add ethtool stat to track copied skbs.
>       netback: make queue length parameter writeable in sysfs
>       netback: parent sysfs device should be set before registering.
>       xen: rename netbk module xen-netback.
>       xen: netback: remove unused xen_network_done code
>       xen: netback: factor disconnect from backend into new function.
>       xen: netback: wait for hotplug scripts to complete before signalling 
> connected to frontend
>       xen/netback: Always pull through PKT_PROT_LEN bytes into the linear 
> part of an skb.
>       xen/netback: Allow setting of large MTU before rings have connected.
>       xen/netback: correctly setup skb->ip_summed on receive
>       xen: netback: handle NET_SKBUFF_DATA_USES_OFFSET correctly
>       xen: netback: drop frag member from struct netbk_rx_meta
>       xen/netback: linearise SKBs as we copy them into guest memory on 
> guest-RX.
>       xen: netback: drop more relics of flipping mode
>       xen/rings: make protocol specific usage of shared sring explicit
>       xen/netback: make protocol specific usage of shared sring explicit
>       xen: netback: check if foreign pages are actually netback-created 
> foreign pages.
>       xen: netback: do not unleash netback threads until initialisation is 
> complete
>       xen: netback: save interrupt state in add_to_net_schedule_list_tail
>       xen: netback: increase size of rx_meta array.
>       xen: netback: take net_schedule_list_lock when removing entry from 
> net_schedule_list
>
> James Harper (1):
>       xen/netback: avoid null-pointer access in netback_uevent

I guess this is fixed to the extent that I haven't seen any complaints
lately, but I think there were still some unresolved issues.

> Jan Beulich (2):
>       linux/netback: unmap tx ring gref when mapping of rx ring gref failed
>       Provide a variant of __RING_SIZE() that is an integer constant 
> expression

This is already upstream.

> Jeremy Fitzhardinge (26):
>       xen/netback: Initial import of linux-2.6.18-xen.hg netback driver.
>       xen/netback: first cut at porting to upstream and cleaning up
>       xen: Add defines for multicall parameter offsets.

Pretty sure these three are yours.

>       xen/netback: don't include xen/evtchn.h
>       xen/netback: use mod_timer
>       xen/netback: use NET_SKB_PAD rather than "16"
>       xen/netback: completely drop flip support
>       xen/netback: demacro MASK_PEND_IDX
>       xen/netback: convert PEND_RING_IDX into a proper typedef name
>       xen/netback: rename NR_PENDING_REQS to nr_pending_reqs()
>       xen/netback: pre-initialize list and spinlocks; use empty list to 
> indicate not on list
>       xen/netback: remove CONFIG_XEN_NETDEV_PIPELINED_TRANSMITTER
>       xen/netback: make netif_get/put inlines
>       xen/netback: move code around
>       xen/netback: document PKT_PROT_LEN
>       netback: convert to net_device_ops
>       xen/netback: reinstate missing code
>       xen/netback: remove debug noise
>       xen/netback: don't screw around with packet gso state
>       xen/netback: use dev_get/set_drvdata() inteface
>       xen/netback: include linux/sched.h for TASK_* definitions
>       xen/netback: use get_sset_count rather than obsolete get_stats_count
>       netback: minor code formatting fixup
>       xen/netback: make sure all the group structures are initialized before 
> starting async code
>       Revert "xen/netback: make sure all the group structures are initialized 
> before starting async code"
>       netback: only initialize for PV domains

I can see a definite need for some folding there.

> Keir Fraser (1):
>       xen: netback: Fixes for delayed copy of tx network packets.
>
> Konrad Rzeszutek Wilk (1):
>       Fix compile warnings: ignoring return value of 
> 'xenbus_register_backend' ..
>
> Paul Durrant (3):
>       xen/netback: Fix basic indentation issue
>       xen/netback: Add a new style of passing GSO packets to frontends.
>       xen/netback: Make frontend features distinct from netback feature flags.

I guess it would be a good time to add Paul's other patches?

    J

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel