* Re: [PATCH net-next] ipv6: Do not use this_cpu_ptr() in preemptible context
From: Eric Dumazet @ 2017-10-08 16:03 UTC (permalink / raw)
To: Ido Schimmel; +Cc: netdev, davem, weiwan, mlxsw
In-Reply-To: <20171008151839.24519-1-idosch@mellanox.com>
On Sun, 2017-10-08 at 18:18 +0300, Ido Schimmel wrote:
> Without the rwlock and with PREEMPT_RCU we're no longer guaranteed to be
> in non-preemptible context when performing a route lookup, so use
> raw_cpu_ptr() instead.
>
> Takes care of the following splat:
> [ 122.221814] BUG: using smp_processor_id() in preemptible [00000000] code: sshd/2672
> [ 122.221845] caller is debug_smp_processor_id+0x17/0x20
> [ 122.221866] CPU: 0 PID: 2672 Comm: sshd Not tainted 4.14.0-rc3-idosch-next-custom #639
> [ 122.221880] Hardware name: Mellanox Technologies Ltd. MSN2100-CB2FO/SA001017, BIOS 5.6.5 06/07/2016
> [ 122.221893] Call Trace:
> [ 122.221919] dump_stack+0xb1/0x10c
> [ 122.221946] ? _atomic_dec_and_lock+0x124/0x124
> [ 122.221974] ? ___ratelimit+0xfe/0x240
> [ 122.222020] check_preemption_disabled+0x173/0x1b0
> [ 122.222060] debug_smp_processor_id+0x17/0x20
> [ 122.222083] ip6_pol_route+0x1482/0x24a0
> ...
>
> Fixes: 66f5d6ce53e6 ("ipv6: replace rwlock with rcu and spinlock in fib6_table")
> Signed-off-by: Ido Schimmel <idosch@mellanox.com>
> ---
Thanks Ido for this patch.
IMO, we no longer play this read_lock() -> write_lock() game since
ip6_dst_gc() could be called from rt6_make_pcpu_route()
So we might simplify things quite a bit, by blocking BH (and thus
preventing preemption)
Something like :
net/ipv6/route.c | 26 ++++++--------------------
1 file changed, 6 insertions(+), 20 deletions(-)
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 399d1bceec4a6e6736c367e706dd2acbd4093d58..606e80325b21c0e10a02e9c7d5b3fcfbfc26a003 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -1136,15 +1136,7 @@ static struct rt6_info *rt6_make_pcpu_route(struct rt6_info *rt)
dst_hold(&pcpu_rt->dst);
p = this_cpu_ptr(rt->rt6i_pcpu);
prev = cmpxchg(p, NULL, pcpu_rt);
- if (prev) {
- /* If someone did it before us, return prev instead */
- /* release refcnt taken by ip6_rt_pcpu_alloc() */
- dst_release_immediate(&pcpu_rt->dst);
- /* release refcnt taken by above dst_hold() */
- dst_release_immediate(&pcpu_rt->dst);
- dst_hold(&prev->dst);
- pcpu_rt = prev;
- }
+ BUG_ON(prev);
rt6_dst_from_metrics_check(pcpu_rt);
return pcpu_rt;
@@ -1739,31 +1731,25 @@ struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table,
struct rt6_info *pcpu_rt;
dst_use_noref(&rt->dst, jiffies);
+ local_bh_disable();
pcpu_rt = rt6_get_pcpu_route(rt);
- if (pcpu_rt) {
- rcu_read_unlock();
- } else {
+ if (!pcpu_rt) {
/* atomic_inc_not_zero() is needed when using rcu */
if (atomic_inc_not_zero(&rt->rt6i_ref)) {
- /* We have to do the read_unlock first
- * because rt6_make_pcpu_route() may trigger
- * ip6_dst_gc() which will take the write_lock.
- *
- * No dst_hold() on rt is needed because grabbing
+ /* No dst_hold() on rt is needed because grabbing
* rt->rt6i_ref makes sure rt can't be released.
*/
- rcu_read_unlock();
pcpu_rt = rt6_make_pcpu_route(rt);
rt6_release(rt);
} else {
/* rt is already removed from tree */
- rcu_read_unlock();
pcpu_rt = net->ipv6.ip6_null_entry;
dst_hold(&pcpu_rt->dst);
}
}
-
+ local_bh_enable();
+ rcu_read_unlock();
trace_fib6_table_lookup(net, pcpu_rt, table->tb6_id, fl6);
return pcpu_rt;
}
^ permalink raw reply related
* [PATCH 00/12] radix-tree: split out struct radix_tree_root out to <linux/radix-tree-root.h>
From: Masahiro Yamada @ 2017-10-08 16:10 UTC (permalink / raw)
To: linux-kernel
Cc: Thomas Gleixner, Andrew Morton, Ian Abbott, Ingo Molnar,
Linus Torvalds, Masahiro Yamada, linux-cachefs, linux-sh,
Rodrigo Vivi, dri-devel, David Airlie, linux-rdma, Yoshinori Sato,
Tariq Toukan, Rich Felker, Leon Romanovsky, Jani Nikula,
J. Bruce Fields, David Howells, intel-gfx, Yishai Hadas
The motivation of this series is to cut down unnecessary header
dependency in terms of radix tree.
Sub-systems or drivers that use radix-tree for data management
typically embed struct radix_tree_root in their data structures,
like this:
struct foo {
...
struct radix_tree_root foo_tree;
...
};
So, <linux/foo.h> needs to include <linux/radix-tree.h>,
therefore, users of <linux/foo.h> include a lot of bloat
from <linux/radix-tree.h>.
If you see the definition of radix_tree_root,
struct radix_tree_root {
gfp_t gfp_mask;
struct radix_tree_node __rcu *rnode;
};
it is a very simple structure.
It only depends on <linux/types.h> for gfp_t and
<linux/compiler.h> for __rcu.
By splitting out the radix_tree_root definition,
we can reduce the header file dependency.
Reducing the header dependency will help for speeding the kernel
build, suppressing unnecessary recompile of objects during
git-bisect'ing, etc.
The patch 1 is a trivial clean-up; it is just here
to avoid conflict.
The patch 2 is the main part of this series;
split out struct radix_tree_root.
The rest of the series replace <linux/radix-tree.h>
with <linux/radix-tree-root.h> where appropriate.
Please review if the idea is OK.
If it is OK, I'd like to know how to apply the series.
Perhaps, the first two for v4.15. Then, rest of series
will be sent per-subsystem for v4.16?
Or, can somebody take care of the whole series?
I checked allmodconfig for x86 and arm64.
I am expecting 0 day testing will check it too.
Masahiro Yamada (12):
radix-tree: replace <linux/spinlock.h> with <linux/spinlock_types.h>
radix-tree: split struct radix_tree_root to <linux/radix-tree-root.h>
irqdomain: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
writeback: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
iocontext.h: replace <linux/radix-tree.h> with
<linux/radix-tree-root.h>
fs: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
blkcg: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
fscache: include <linux-radix-tree.h>
sh: intc: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
net/mlx4: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
net/mlx5: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
drm/i915: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
drivers/gpu/drm/i915/i915_gem.c | 1 +
drivers/gpu/drm/i915/i915_gem_context.c | 1 +
drivers/gpu/drm/i915/i915_gem_context.h | 2 +-
drivers/gpu/drm/i915/i915_gem_execbuffer.c | 1 +
drivers/gpu/drm/i915/i915_gem_object.h | 1 +
drivers/net/ethernet/mellanox/mlx4/cq.c | 1 +
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 2 +-
drivers/net/ethernet/mellanox/mlx4/qp.c | 1 +
drivers/net/ethernet/mellanox/mlx4/srq.c | 1 +
drivers/sh/intc/internals.h | 2 +-
include/linux/backing-dev-defs.h | 2 +-
include/linux/blk-cgroup.h | 2 +-
include/linux/fs.h | 2 +-
include/linux/fscache.h | 1 +
include/linux/iocontext.h | 2 +-
include/linux/irqdomain.h | 2 +-
include/linux/mlx4/device.h | 2 +-
include/linux/mlx4/qp.h | 1 +
include/linux/mlx5/driver.h | 2 +-
include/linux/mlx5/qp.h | 1 +
include/linux/radix-tree-root.h | 24 ++++++++++++++++++++++++
include/linux/radix-tree.h | 8 ++------
22 files changed, 46 insertions(+), 16 deletions(-)
create mode 100644 include/linux/radix-tree-root.h
--
2.7.4
^ permalink raw reply
* [PATCH 10/12] net/mlx4: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
From: Masahiro Yamada @ 2017-10-08 16:10 UTC (permalink / raw)
To: linux-kernel
Cc: Thomas Gleixner, Andrew Morton, Ian Abbott, Ingo Molnar,
Linus Torvalds, Masahiro Yamada, linux-rdma, Yishai Hadas,
Tariq Toukan, netdev
In-Reply-To: <1507479013-5207-1-git-send-email-yamada.masahiro@socionext.com>
The headers
- include/linux/mlx4/device.h
- drivers/net/ethernet/mellanox/mlx4/mlx4.h
require the definition of struct radix_tree_root, but do not need to
know anything about other radix tree stuff.
Include <linux/radix-tree-root.h> instead of <linux/radix-tree.h> to
reduce the header dependency.
While we are here, let's add missing <linux/radix-tree.h> where
radix tree accessors are used.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---
drivers/net/ethernet/mellanox/mlx4/cq.c | 1 +
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 2 +-
drivers/net/ethernet/mellanox/mlx4/qp.c | 1 +
drivers/net/ethernet/mellanox/mlx4/srq.c | 1 +
include/linux/mlx4/device.h | 2 +-
include/linux/mlx4/qp.h | 1 +
6 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/cq.c b/drivers/net/ethernet/mellanox/mlx4/cq.c
index 72eb50c..4cbe65c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cq.c
@@ -36,6 +36,7 @@
#include <linux/hardirq.h>
#include <linux/export.h>
+#include <linux/radix-tree.h>
#include <linux/mlx4/cmd.h>
#include <linux/mlx4/cq.h>
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index c68da19..975ef70 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -38,7 +38,7 @@
#define MLX4_H
#include <linux/mutex.h>
-#include <linux/radix-tree.h>
+#include <linux/radix-tree-root.h>
#include <linux/rbtree.h>
#include <linux/timer.h>
#include <linux/semaphore.h>
diff --git a/drivers/net/ethernet/mellanox/mlx4/qp.c b/drivers/net/ethernet/mellanox/mlx4/qp.c
index 728a2fb..50cbc62 100644
--- a/drivers/net/ethernet/mellanox/mlx4/qp.c
+++ b/drivers/net/ethernet/mellanox/mlx4/qp.c
@@ -35,6 +35,7 @@
#include <linux/gfp.h>
#include <linux/export.h>
+#include <linux/radix-tree.h>
#include <linux/mlx4/cmd.h>
#include <linux/mlx4/qp.h>
diff --git a/drivers/net/ethernet/mellanox/mlx4/srq.c b/drivers/net/ethernet/mellanox/mlx4/srq.c
index bedf521..4201a46 100644
--- a/drivers/net/ethernet/mellanox/mlx4/srq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/srq.c
@@ -36,6 +36,7 @@
#include <linux/mlx4/srq.h>
#include <linux/export.h>
#include <linux/gfp.h>
+#include <linux/radix-tree.h>
#include "mlx4.h"
#include "icm.h"
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index b0a57e0..75eac23 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -36,7 +36,7 @@
#include <linux/if_ether.h>
#include <linux/pci.h>
#include <linux/completion.h>
-#include <linux/radix-tree.h>
+#include <linux/radix-tree-root.h>
#include <linux/cpu_rmap.h>
#include <linux/crash_dump.h>
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index 8e2828d..dfa7d8e 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -35,6 +35,7 @@
#include <linux/types.h>
#include <linux/if_ether.h>
+#include <linux/radix-tree.h>
#include <linux/mlx4/device.h>
--
2.7.4
^ permalink raw reply related
* [PATCH 11/12] net/mlx5: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
From: Masahiro Yamada @ 2017-10-08 16:10 UTC (permalink / raw)
To: linux-kernel
Cc: Thomas Gleixner, Andrew Morton, Ian Abbott, Ingo Molnar,
Linus Torvalds, Masahiro Yamada, Matan Barak, linux-rdma, netdev,
Saeed Mahameed, Leon Romanovsky
In-Reply-To: <1507479013-5207-1-git-send-email-yamada.masahiro@socionext.com>
The header include/linux/mlx5/driver.h requires the definition of
struct radix_tree_root, but does not need to know anything about
other radix tree stuff.
Include <linux/radix-tree-root.h> instead of <linux/radix-tree.h> to
reduce the number of included header files.
Also, add <linux/radix-tree.h> to include/linux/mlx5/gp.h where radix
tree accessors are used.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---
include/linux/mlx5/driver.h | 2 +-
include/linux/mlx5/qp.h | 1 +
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index 401c897..0aea568 100644
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -40,7 +40,7 @@
#include <linux/semaphore.h>
#include <linux/slab.h>
#include <linux/vmalloc.h>
-#include <linux/radix-tree.h>
+#include <linux/radix-tree-root.h>
#include <linux/workqueue.h>
#include <linux/mempool.h>
#include <linux/interrupt.h>
diff --git a/include/linux/mlx5/qp.h b/include/linux/mlx5/qp.h
index 66d19b6..a90996f 100644
--- a/include/linux/mlx5/qp.h
+++ b/include/linux/mlx5/qp.h
@@ -35,6 +35,7 @@
#include <linux/mlx5/device.h>
#include <linux/mlx5/driver.h>
+#include <linux/radix-tree.h>
#define MLX5_INVALID_LKEY 0x100
#define MLX5_SIG_WQE_SIZE (MLX5_SEND_WQE_BB * 5)
--
2.7.4
^ permalink raw reply related
* Re: [PATCH net-next] ipv6: Do not use this_cpu_ptr() in preemptible context
From: Ido Schimmel @ 2017-10-08 16:54 UTC (permalink / raw)
To: Eric Dumazet; +Cc: Ido Schimmel, netdev, davem, weiwan, mlxsw
In-Reply-To: <1507478633.14419.30.camel@edumazet-glaptop3.roam.corp.google.com>
Hi Eric,
On Sun, Oct 08, 2017 at 09:03:53AM -0700, Eric Dumazet wrote:
> Thanks Ido for this patch.
>
> IMO, we no longer play this read_lock() -> write_lock() game since
> ip6_dst_gc() could be called from rt6_make_pcpu_route()
Right, cause we can't deadlock anymore as with the rwlock.
>
> So we might simplify things quite a bit, by blocking BH (and thus
> preventing preemption)
>
> Something like :
>
> net/ipv6/route.c | 26 ++++++--------------------
> 1 file changed, 6 insertions(+), 20 deletions(-)
>
> diff --git a/net/ipv6/route.c b/net/ipv6/route.c
> index 399d1bceec4a6e6736c367e706dd2acbd4093d58..606e80325b21c0e10a02e9c7d5b3fcfbfc26a003 100644
> --- a/net/ipv6/route.c
> +++ b/net/ipv6/route.c
> @@ -1136,15 +1136,7 @@ static struct rt6_info *rt6_make_pcpu_route(struct rt6_info *rt)
> dst_hold(&pcpu_rt->dst);
> p = this_cpu_ptr(rt->rt6i_pcpu);
> prev = cmpxchg(p, NULL, pcpu_rt);
> - if (prev) {
> - /* If someone did it before us, return prev instead */
> - /* release refcnt taken by ip6_rt_pcpu_alloc() */
> - dst_release_immediate(&pcpu_rt->dst);
> - /* release refcnt taken by above dst_hold() */
> - dst_release_immediate(&pcpu_rt->dst);
> - dst_hold(&prev->dst);
> - pcpu_rt = prev;
> - }
> + BUG_ON(prev);
Is this BUG_ON() now valid because of the local_bh_disable() in
ip6_pol_route()?
>
> rt6_dst_from_metrics_check(pcpu_rt);
> return pcpu_rt;
> @@ -1739,31 +1731,25 @@ struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table,
> struct rt6_info *pcpu_rt;
>
> dst_use_noref(&rt->dst, jiffies);
> + local_bh_disable();
> pcpu_rt = rt6_get_pcpu_route(rt);
>
> - if (pcpu_rt) {
> - rcu_read_unlock();
> - } else {
> + if (!pcpu_rt) {
> /* atomic_inc_not_zero() is needed when using rcu */
> if (atomic_inc_not_zero(&rt->rt6i_ref)) {
> - /* We have to do the read_unlock first
> - * because rt6_make_pcpu_route() may trigger
> - * ip6_dst_gc() which will take the write_lock.
> - *
> - * No dst_hold() on rt is needed because grabbing
> + /* No dst_hold() on rt is needed because grabbing
> * rt->rt6i_ref makes sure rt can't be released.
> */
> - rcu_read_unlock();
> pcpu_rt = rt6_make_pcpu_route(rt);
> rt6_release(rt);
> } else {
> /* rt is already removed from tree */
> - rcu_read_unlock();
> pcpu_rt = net->ipv6.ip6_null_entry;
> dst_hold(&pcpu_rt->dst);
> }
> }
> -
> + local_bh_enable();
> + rcu_read_unlock();
> trace_fib6_table_lookup(net, pcpu_rt, table->tb6_id, fl6);
> return pcpu_rt;
> }
I replaced my patch with yours and I don't trigger the bug anymore. Feel
free to add my tag:
Tested-by: Ido Schimmel <idosch@mellanox.com>
Thanks!
^ permalink raw reply
* Re: [PATCH 10/12] net/mlx4: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
From: David Miller @ 2017-10-08 17:00 UTC (permalink / raw)
To: yamada.masahiro
Cc: linux-kernel, tglx, akpm, abbotti, mingo, torvalds, linux-rdma,
yishaih, tariqt, netdev
In-Reply-To: <1507479013-5207-11-git-send-email-yamada.masahiro@socionext.com>
From: Masahiro Yamada <yamada.masahiro@socionext.com>
Date: Mon, 9 Oct 2017 01:10:11 +0900
> The headers
> - include/linux/mlx4/device.h
> - drivers/net/ethernet/mellanox/mlx4/mlx4.h
> require the definition of struct radix_tree_root, but do not need to
> know anything about other radix tree stuff.
>
> Include <linux/radix-tree-root.h> instead of <linux/radix-tree.h> to
> reduce the header dependency.
>
> While we are here, let's add missing <linux/radix-tree.h> where
> radix tree accessors are used.
>
> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Honestly this makes things more complicated.
The driver was trying to consolidate all of the header needs
by including them all in one place, the main driver header.
Now you're including headers in several different files.
I really don't like the results of this change and would
ask you to reconsider.
Just add both radix-tree-root.h _and_ radix-tree.h to mlx4.h
and leave the rest of the driver alone.
^ permalink raw reply
* Re: [PATCH 11/12] net/mlx5: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
From: David Miller @ 2017-10-08 17:00 UTC (permalink / raw)
To: yamada.masahiro
Cc: linux-kernel, tglx, akpm, abbotti, mingo, torvalds, matanb,
linux-rdma, netdev, saeedm, leonro
In-Reply-To: <1507479013-5207-12-git-send-email-yamada.masahiro@socionext.com>
From: Masahiro Yamada <yamada.masahiro@socionext.com>
Date: Mon, 9 Oct 2017 01:10:12 +0900
> The header include/linux/mlx5/driver.h requires the definition of
> struct radix_tree_root, but does not need to know anything about
> other radix tree stuff.
>
> Include <linux/radix-tree-root.h> instead of <linux/radix-tree.h> to
> reduce the number of included header files.
>
> Also, add <linux/radix-tree.h> to include/linux/mlx5/gp.h where radix
> tree accessors are used.
>
> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Same objections as the mlx4 changes, just include both headers in
driver.h
^ permalink raw reply
* Re: [PATCH 3/3] batman-adv: Add missing kerneldoc for extack
From: David Miller @ 2017-10-08 17:04 UTC (permalink / raw)
To: sven; +Cc: b.a.t.m.a.n, netdev, dsahern
In-Reply-To: <1655154.zJsG0lsWS3@sven-edge>
From: Sven Eckelmann <sven@narfation.org>
Date: Sun, 08 Oct 2017 09:29:17 +0200
> Are you expected to apply this change:
> =====================================
>
> I was hoping that Simon is picking the patch up and forwards it do you in a
> proper pull request. But I would doubt that he has a big problem with you
> applying this single line kernel-doc change. At least it is less work for him
> and less extra noise on both mailing lists...
>
> But feel free to tell us your preferred solution.
Ok I'll wait to get this from the next batman-adv pull request then.
Thanks.
^ permalink raw reply
* Re: [patch net] mlxsw: spectrum_router: Avoid expensive lookup during route removal
From: David Miller @ 2017-10-08 17:06 UTC (permalink / raw)
To: jiri; +Cc: netdev, idosch, mlxsw, dsa
In-Reply-To: <20171008095326.1238-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@resnulli.us>
Date: Sun, 8 Oct 2017 11:53:26 +0200
> From: Ido Schimmel <idosch@mellanox.com>
>
> In commit fc922bb0dd94 ("mlxsw: spectrum_router: Use one LPM tree for
> all virtual routers") I increased the scale of supported VRFs by having
> all of them share the same LPM tree.
>
> In order to avoid look-ups for prefix lengths that don't exist, each
> route removal would trigger an aggregation across all the active virtual
> routers to see which prefix lengths are in use and which aren't and
> structure the tree accordingly.
>
> With the way the data structures are currently laid out, this is a very
> expensive operation. When preformed repeatedly - due to the invocation
> of the abort mechanism - and with enough VRFs, this can result in a hung
> task.
>
> For now, avoid this optimization until it can be properly re-added in
> net-next.
>
> Fixes: fc922bb0dd94 ("mlxsw: spectrum_router: Use one LPM tree for all virtual routers")
> Signed-off-by: Ido Schimmel <idosch@mellanox.com>
> Reported-by: David Ahern <dsa@cumulusnetworks.com>
> Tested-by: David Ahern <dsa@cumulusnetworks.com>
> Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Applied.
^ permalink raw reply
* Re: [patch net-next repost 0/2] mlxsw: Add more extack error reporting
From: David Miller @ 2017-10-08 17:07 UTC (permalink / raw)
To: jiri; +Cc: netdev, idosch, mlxsw, dsahern
In-Reply-To: <20171008095756.2139-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@resnulli.us>
Date: Sun, 8 Oct 2017 11:57:54 +0200
> From: Jiri Pirko <jiri@mellanox.com>
>
> Ido says:
>
> Add error messages to VLAN and bridge enslavements to help users
> understand why the enslavement failed.
Series applied, thanks for the repost.
^ permalink raw reply
* Re: [PATCH net-next,0/3] hv_netvsc: support changing TCP hash level
From: David Miller @ 2017-10-08 17:11 UTC (permalink / raw)
To: haiyangz, haiyangz; +Cc: netdev, kys, sthemmin, olaf, vkuznets, linux-kernel
In-Reply-To: <20171006153359.8400-1-haiyangz@exchange.microsoft.com>
From: Haiyang Zhang <haiyangz@exchange.microsoft.com>
Date: Fri, 6 Oct 2017 08:33:56 -0700
> From: Haiyang Zhang <haiyangz@microsoft.com>
>
> The patch set simplifies the existing hash level switching code for
> UDP. It also adds the support for changing TCP hash level. So users
> can switch between L3 an L4 hash levels for TCP and UDP.
Series applied.
^ permalink raw reply
* Re: [PATCH v2] gso: fix payload length when gso_size is zero
From: David Miller @ 2017-10-08 17:13 UTC (permalink / raw)
To: alexey.kodanev; +Cc: netdev, steffen.klassert, alexander.h.duyck
In-Reply-To: <1507305755-14393-1-git-send-email-alexey.kodanev@oracle.com>
From: Alexey Kodanev <alexey.kodanev@oracle.com>
Date: Fri, 6 Oct 2017 19:02:35 +0300
> When gso_size reset to zero for the tail segment in skb_segment(), later
> in ipv6_gso_segment(), __skb_udp_tunnel_segment() and gre_gso_segment()
> we will get incorrect results (payload length, pcsum) for that segment.
> inet_gso_segment() already has a check for gso_size before calculating
> payload.
>
> The issue was found with LTP vxlan & gre tests over ixgbe NIC.
>
> Fixes: 07b26c9454a2 ("gso: Support partial splitting at the frag_list pointer")
> Signed-off-by: Alexey Kodanev <alexey.kodanev@oracle.com>
> ---
> v2: also added skb_is_gso to gre_gso_segment() and __skb_udp_tunnel_segment()
>
Applied and queued up for -stable, thanks.
^ permalink raw reply
* Re: [PATCH 10/12] net/mlx4: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
From: Masahiro Yamada @ 2017-10-08 17:29 UTC (permalink / raw)
To: David Miller
Cc: Linux Kernel Mailing List, Thomas Gleixner, Andrew Morton,
abbotti, Ingo Molnar, Linus Torvalds, linux-rdma, yishaih, tariqt,
netdev
In-Reply-To: <20171008.100013.2081721941420989813.davem@davemloft.net>
2017-10-09 2:00 GMT+09:00 David Miller <davem@davemloft.net>:
> From: Masahiro Yamada <yamada.masahiro@socionext.com>
> Date: Mon, 9 Oct 2017 01:10:11 +0900
>
>> The headers
>> - include/linux/mlx4/device.h
>> - drivers/net/ethernet/mellanox/mlx4/mlx4.h
>> require the definition of struct radix_tree_root, but do not need to
>> know anything about other radix tree stuff.
>>
>> Include <linux/radix-tree-root.h> instead of <linux/radix-tree.h> to
>> reduce the header dependency.
>>
>> While we are here, let's add missing <linux/radix-tree.h> where
>> radix tree accessors are used.
>>
>> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
>
> Honestly this makes things more complicated.
The idea is simple; include necessary headers explicitly.
Putting everything into one common header
means most of C files are forced to parse unnecessary headers.
> The driver was trying to consolidate all of the header needs
> by including them all in one place, the main driver header.
>
> Now you're including headers in several different files.
>
> I really don't like the results of this change and would
> ask you to reconsider.
>
> Just add both radix-tree-root.h _and_ radix-tree.h to mlx4.h
> and leave the rest of the driver alone.
If you do not like this, you can just throw it away.
<linux/radix-tree.h> includes <linux/radix-tree-root.h>.
You do not need to include both.
--
Best Regards
Masahiro Yamada
^ permalink raw reply
* Re: [PATCH 10/12] net/mlx4: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
From: Joe Perches @ 2017-10-08 17:32 UTC (permalink / raw)
To: Masahiro Yamada, David Miller
Cc: Linux Kernel Mailing List, Thomas Gleixner, Andrew Morton,
abbotti, Ingo Molnar, Linus Torvalds, linux-rdma, yishaih, tariqt,
netdev
In-Reply-To: <CAK7LNAQ1BeSb14poieRU4uMsnsEHPNL2zSNtWi1TGN=YrK_8jA@mail.gmail.com>
On Mon, 2017-10-09 at 02:29 +0900, Masahiro Yamada wrote:
> The idea is simple; include necessary headers explicitly.
Try that for kernel.h
There's a reason aggregation of #includes is useful.
^ permalink raw reply
* Re: [PATCH net-next] ipv6: Do not use this_cpu_ptr() in preemptible context
From: Eric Dumazet @ 2017-10-08 18:25 UTC (permalink / raw)
To: Ido Schimmel; +Cc: Ido Schimmel, netdev, davem, weiwan, mlxsw
In-Reply-To: <20171008165407.GA26635@shredder.mtl.com>
On Sun, 2017-10-08 at 19:54 +0300, Ido Schimmel wrote:
> Hi Eric,
> > prev = cmpxchg(p, NULL, pcpu_rt);
> > - if (prev) {
> > - /* If someone did it before us, return prev instead */
> > - /* release refcnt taken by ip6_rt_pcpu_alloc() */
> > - dst_release_immediate(&pcpu_rt->dst);
> > - /* release refcnt taken by above dst_hold() */
> > - dst_release_immediate(&pcpu_rt->dst);
> > - dst_hold(&prev->dst);
> > - pcpu_rt = prev;
> > - }
> > + BUG_ON(prev);
>
> Is this BUG_ON() now valid because of the local_bh_disable() in
> ip6_pol_route()?
Yes, this bug to trigger would need this code be re-entered from a hard
IRQ, and that would be wrong of course.
^ permalink raw reply
* Re: [PATCH v6 05/11] dt-bindings: net: dwmac-sun8i: update documentation about integrated PHY
From: Corentin Labbe @ 2017-10-08 18:33 UTC (permalink / raw)
To: Andrew Lunn, robh+dt, maxime.ripard, wens, f.fainelli
Cc: mark.rutland, linux, catalin.marinas, will.deacon,
peppe.cavallaro, alexandre.torgue, frowand.list, netdev,
devicetree, linux-arm-kernel, linux-kernel, linux-sunxi
In-Reply-To: <20170928073708.GB32676@Red>
On Thu, Sep 28, 2017 at 09:37:08AM +0200, Corentin Labbe wrote:
> On Wed, Sep 27, 2017 at 04:02:10PM +0200, Andrew Lunn wrote:
> > Hi Corentin
> >
> > > +Required properties for the mdio-mux node:
> > > + - compatible = "mdio-mux"
> >
> > This is too generic. Please add a more specific compatible for this
> > particular mux. You can keep "mdio-mux", since that is what the MDIO
> > subsystem will look for.
> >
>
> I will add allwinner,sun8i-h3-mdio-mux
>
> > > +Required properties of the integrated phy node:
> > > - clocks: a phandle to the reference clock for the EPHY
> > > - resets: a phandle to the reset control for the EPHY
> > > +- phy-is-integrated
> >
> > So the last thing you said is that the mux is not the problem
> > here. Something else is locking up. Did you discover what?
> >
> > I really would like phy-is-integrated to go away.
> >
>
> I have found the problem: by enabling ephy clk/reset the timeout does not occur anymore.
> So we could remove phy-is-integrated by:
> Moving internal phy clk/reset handling in mdio_mux_syscon_switch_fn()
> But this means:
> - getting internalphy node always by manually get internal_mdio/internal_phy (and not by the given phyhandle)
> - doing some unnecessary tasks (enable/scan/disable) when external_phy is needed
>
> Regards
Hello all
Below is the current patch, as you can read, it does not use anymore the phy-is-integrated property.
So now, the mdio-mux must always enable the internal mdio when switch_fn ask for it and so reset MAC and so need to enable ephy clk/reset.
But for this I need a reference to thoses clock and reset. (this is done in get_ephy_nodes)
The current version set those clock in mdio-mux node, and as you can see it is already ugly (lots of get next node),
if the clk/rst nodes were as it should be, in phy nodes, it will be more bad.
So, since the MAC have a dependency on thoses clk/rst nodes for doing reset(), I seek a proper way to get references on it.
OR do you agree that putting ephy clk/rst in emac is acceptable ?
thanks
regards
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-sun8i.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-sun8i.c
@@ -17,6 +17,7 @@
#include <linux/clk.h>
#include <linux/io.h>
#include <linux/iopoll.h>
+#include <linux/mdio-mux.h>
#include <linux/mfd/syscon.h>
#include <linux/module.h>
#include <linux/of_device.h>
@@ -41,14 +42,14 @@
* This value is used for disabling properly EMAC
* and used as a good starting value in case of the
* boot process(uboot) leave some stuff.
- * @internal_phy: Does the MAC embed an internal PHY
+ * @soc_has_internal_phy: Does the MAC embed an internal PHY
* @support_mii: Does the MAC handle MII
* @support_rmii: Does the MAC handle RMII
* @support_rgmii: Does the MAC handle RGMII
*/
struct emac_variant {
u32 default_syscon_value;
- int internal_phy;
+ bool soc_has_internal_phy;
bool support_mii;
bool support_rmii;
bool support_rgmii;
@@ -61,7 +62,7 @@ struct emac_variant {
* @rst_ephy: reference to the optional EPHY reset for the internal PHY
* @variant: reference to the current board variant
* @regmap: regmap for using the syscon
- * @use_internal_phy: Does the current PHY choice imply using the internal PHY
+ * @internal_phy_powered: Does the internal PHY is enabled
*/
struct sunxi_priv_data {
struct clk *tx_clk;
@@ -70,12 +71,13 @@ struct sunxi_priv_data {
struct reset_control *rst_ephy;
const struct emac_variant *variant;
struct regmap *regmap;
- bool use_internal_phy;
+ bool internal_phy_powered;
+ void *mux_handle;
};
static const struct emac_variant emac_variant_h3 = {
.default_syscon_value = 0x58000,
- .internal_phy = PHY_INTERFACE_MODE_MII,
+ .soc_has_internal_phy = true,
.support_mii = true,
.support_rmii = true,
.support_rgmii = true
@@ -83,20 +85,20 @@ static const struct emac_variant emac_variant_h3 = {
static const struct emac_variant emac_variant_v3s = {
.default_syscon_value = 0x38000,
- .internal_phy = PHY_INTERFACE_MODE_MII,
+ .soc_has_internal_phy = true,
.support_mii = true
};
static const struct emac_variant emac_variant_a83t = {
.default_syscon_value = 0,
- .internal_phy = 0,
+ .soc_has_internal_phy = false,
.support_mii = true,
.support_rgmii = true
};
static const struct emac_variant emac_variant_a64 = {
.default_syscon_value = 0,
- .internal_phy = 0,
+ .soc_has_internal_phy = false,
.support_mii = true,
.support_rmii = true,
.support_rgmii = true
@@ -195,6 +197,9 @@ static const struct emac_variant emac_variant_a64 = {
#define H3_EPHY_LED_POL BIT(17) /* 1: active low, 0: active high */
#define H3_EPHY_SHUTDOWN BIT(16) /* 1: shutdown, 0: power up */
#define H3_EPHY_SELECT BIT(15) /* 1: internal PHY, 0: external PHY */
+#define H3_EPHY_MUX_MASK (H3_EPHY_SHUTDOWN | H3_EPHY_SELECT)
+#define DWMAC_SUN8I_MDIO_MUX_INTERNAL_ID 1
+#define DWMAC_SUN8I_MDIO_MUX_EXTERNAL_ID 2
/* H3/A64 specific bits */
#define SYSCON_RMII_EN BIT(13) /* 1: enable RMII (overrides EPIT) */
@@ -634,6 +639,164 @@ static int sun8i_dwmac_reset(struct stmmac_priv *priv)
return 0;
}
+static int get_ephy_nodes(struct stmmac_priv *priv)
+{
+ struct sunxi_priv_data *gmac = priv->plat->bsp_priv;
+ struct device_node *mdio_mux;
+ struct device_node *mdio_internal;
+ int ret;
+
+ mdio_mux = of_get_child_by_name(priv->plat->mdio_node, "mdio-mux");
+ if (!mdio_mux) {
+ dev_err(priv->device, "Cannot get mdio-mux node\n");
+ return -ENODEV;
+ }
+
+ mdio_internal = of_find_compatible_node(mdio_mux, NULL,
+ "allwinner,sun8i-h3-mdio-internal");
+ if (!mdio_internal) {
+ dev_err(priv->device, "Cannot get internal_mdio node\n");
+ return -ENODEV;
+ }
+
+ /* insert here code to get child phynode */
+ gmac->ephy_clk = of_clk_get(mdio_internal, 0);
+ if (IS_ERR(gmac->ephy_clk)) {
+ ret = PTR_ERR(gmac->ephy_clk);
+ dev_err(priv->device, "Cannot get EPHY clock: %d\n", ret);
+ return -EINVAL;
+ }
+
+ /* Note that we cannot use devm_reset_control_get, since
+ * the reset is not in the device node.
+ */
+ gmac->rst_ephy = of_reset_control_get_exclusive(mdio_internal, NULL);
+ if (IS_ERR(gmac->rst_ephy)) {
+ ret = PTR_ERR(gmac->rst_ephy);
+ if (ret == -EPROBE_DEFER)
+ return ret;
+ dev_err(priv->device, "No EPHY reset control found %d\n", ret);
+ return -EINVAL;
+ }
+ return 0;
+}
+
+static int sun8i_dwmac_power_internal_phy(struct stmmac_priv *priv)
+{
+ struct sunxi_priv_data *gmac = priv->plat->bsp_priv;
+ int ret;
+
+ if (gmac->internal_phy_powered) {
+ dev_warn(priv->device, "Internal PHY already powered\n");
+ return 0;
+ }
+
+ dev_info(priv->device, "Powering internal PHY\n");
+ ret = clk_prepare_enable(gmac->ephy_clk);
+ if (ret) {
+ dev_err(priv->device, "Cannot enable internal PHY\n");
+ return ret;
+ }
+
+ /* Make sure the EPHY is properly reseted, as U-Boot may leave
+ * it at deasserted state, and thus it may fail to reset EMAC.
+ */
+ reset_control_assert(gmac->rst_ephy);
+
+ ret = reset_control_deassert(gmac->rst_ephy);
+ if (ret) {
+ dev_err(priv->device, "Cannot deassert internal phy\n");
+ clk_disable_unprepare(gmac->ephy_clk);
+ return ret;
+ }
+
+ gmac->internal_phy_powered = true;
+
+ return 0;
+}
+
+static int sun8i_dwmac_unpower_internal_phy(struct sunxi_priv_data *gmac)
+{
+ if (!gmac->internal_phy_powered)
+ return 0;
+
+ clk_disable_unprepare(gmac->ephy_clk);
+ reset_control_assert(gmac->rst_ephy);
+ gmac->internal_phy_powered = false;
+ return 0;
+}
+
+/* MDIO multiplexing switch function
+ * This function is called by the mdio-mux layer when it thinks the mdio bus
+ * multiplexer needs to switch.
+ * 'current_child' is the current value of the mux register
+ * 'desired_child' is the value of the 'reg' property of the target child MDIO
+ * node.
+ * The first time this function is called, current_child == -1.
+ * If current_child == desired_child, then the mux is already set to the
+ * correct bus.
+ *
+ * Note that we do not use reg/mask like mdio-mux-mmioreg because we need to
+ * know easily which bus is used (reset must be done only for desired bus).
+ */
+static int mdio_mux_syscon_switch_fn(int current_child, int desired_child,
+ void *data)
+{
+ struct stmmac_priv *priv = data;
+ struct sunxi_priv_data *gmac = priv->plat->bsp_priv;
+ u32 reg, val;
+ int ret = 0;
+ bool need_power_ephy = false;
+
+ if (current_child ^ desired_child) {
+ regmap_read(gmac->regmap, SYSCON_EMAC_REG, ®);
+ switch (desired_child) {
+ case DWMAC_SUN8I_MDIO_MUX_INTERNAL_ID:
+ dev_info(priv->device, "Switch mux to internal PHY");
+ val = (reg & ~H3_EPHY_MUX_MASK) | H3_EPHY_SELECT;
+
+ need_power_ephy = true;
+ break;
+ case DWMAC_SUN8I_MDIO_MUX_EXTERNAL_ID:
+ dev_info(priv->device, "Switch mux to external PHY");
+ val = (reg & ~H3_EPHY_MUX_MASK) | H3_EPHY_SHUTDOWN;
+ need_power_ephy = false;
+ break;
+ default:
+ dev_err(priv->device, "Invalid child ID %x\n",
+ desired_child);
+ return -EINVAL;
+ }
+ regmap_write(gmac->regmap, SYSCON_EMAC_REG, val);
+ if (need_power_ephy) {
+ ret = sun8i_dwmac_power_internal_phy(priv);
+ if (ret)
+ return ret;
+ } else {
+ sun8i_dwmac_unpower_internal_phy(gmac);
+ }
+ /* After changing syscon value, the MAC need reset or it will
+ * use the last value (and so the last PHY set).
+ */
+ ret = sun8i_dwmac_reset(priv);
+ }
+ return ret;
+}
+
+static int sun8i_dwmac_register_mdio_mux(struct stmmac_priv *priv)
+{
+ int ret;
+ struct device_node *mdio_mux;
+ struct sunxi_priv_data *gmac = priv->plat->bsp_priv;
+
+ mdio_mux = of_get_child_by_name(priv->plat->mdio_node, "mdio-mux");
+ if (!mdio_mux)
+ return -ENODEV;
+
+ ret = mdio_mux_init(priv->device, mdio_mux, mdio_mux_syscon_switch_fn,
+ &gmac->mux_handle, priv, priv->mii);
+ return ret;
+}
+
static int sun8i_dwmac_set_syscon(struct stmmac_priv *priv)
{
struct sunxi_priv_data *gmac = priv->plat->bsp_priv;
@@ -648,35 +811,25 @@ static int sun8i_dwmac_set_syscon(struct stmmac_priv *priv)
"Current syscon value is not the default %x (expect %x)\n",
val, reg);
- if (gmac->variant->internal_phy) {
- if (!gmac->use_internal_phy) {
- /* switch to external PHY interface */
- reg &= ~H3_EPHY_SELECT;
- } else {
- reg |= H3_EPHY_SELECT;
- reg &= ~H3_EPHY_SHUTDOWN;
- dev_dbg(priv->device, "Select internal_phy %x\n", reg);
-
- if (of_property_read_bool(priv->plat->phy_node,
- "allwinner,leds-active-low"))
- reg |= H3_EPHY_LED_POL;
- else
- reg &= ~H3_EPHY_LED_POL;
-
- /* Force EPHY xtal frequency to 24MHz. */
- reg |= H3_EPHY_CLK_SEL;
-
- ret = of_mdio_parse_addr(priv->device,
- priv->plat->phy_node);
- if (ret < 0) {
- dev_err(priv->device, "Could not parse MDIO addr\n");
- return ret;
- }
- /* of_mdio_parse_addr returns a valid (0 ~ 31) PHY
- * address. No need to mask it again.
- */
- reg |= ret << H3_EPHY_ADDR_SHIFT;
+ if (gmac->variant->soc_has_internal_phy) {
+ if (of_property_read_bool(priv->plat->phy_node,
+ "allwinner,leds-active-low"))
+ reg |= H3_EPHY_LED_POL;
+ else
+ reg &= ~H3_EPHY_LED_POL;
+
+ /* Force EPHY xtal frequency to 24MHz. */
+ reg |= H3_EPHY_CLK_SEL;
+
+ ret = of_mdio_parse_addr(priv->device, priv->plat->phy_node);
+ if (ret < 0) {
+ dev_err(priv->device, "Could not parse MDIO addr\n");
+ return ret;
}
+ /* of_mdio_parse_addr returns a valid (0 ~ 31) PHY
+ * address. No need to mask it again.
+ */
+ reg |= 1 << H3_EPHY_ADDR_SHIFT;
}
if (!of_property_read_u32(node, "allwinner,tx-delay-ps", &val)) {
@@ -746,81 +899,21 @@ static void sun8i_dwmac_unset_syscon(struct sunxi_priv_data *gmac)
regmap_write(gmac->regmap, SYSCON_EMAC_REG, reg);
}
-static int sun8i_dwmac_power_internal_phy(struct stmmac_priv *priv)
+static void sun8i_dwmac_exit(struct platform_device *pdev, void *priv)
{
- struct sunxi_priv_data *gmac = priv->plat->bsp_priv;
- int ret;
-
- if (!gmac->use_internal_phy)
- return 0;
-
- ret = clk_prepare_enable(gmac->ephy_clk);
- if (ret) {
- dev_err(priv->device, "Cannot enable ephy\n");
- return ret;
- }
-
- /* Make sure the EPHY is properly reseted, as U-Boot may leave
- * it at deasserted state, and thus it may fail to reset EMAC.
- */
- reset_control_assert(gmac->rst_ephy);
+ struct sunxi_priv_data *gmac = priv;
- ret = reset_control_deassert(gmac->rst_ephy);
- if (ret) {
- dev_err(priv->device, "Cannot deassert ephy\n");
- clk_disable_unprepare(gmac->ephy_clk);
- return ret;
+ if (gmac->variant->soc_has_internal_phy) {
+ /* sun8i_dwmac_exit could be called with mdiomux uninit */
+ if (gmac->mux_handle)
+ mdio_mux_uninit(gmac->mux_handle);
+ if (gmac->internal_phy_powered)
+ sun8i_dwmac_unpower_internal_phy(gmac);
}
- return 0;
-}
-
-static int sun8i_dwmac_unpower_internal_phy(struct sunxi_priv_data *gmac)
-{
- if (!gmac->use_internal_phy)
- return 0;
-
- clk_disable_unprepare(gmac->ephy_clk);
- reset_control_assert(gmac->rst_ephy);
- return 0;
-}
-
-/* sun8i_power_phy() - Activate the PHY:
- * In case of error, no need to call sun8i_unpower_phy(),
- * it will be called anyway by sun8i_dwmac_exit()
- */
-static int sun8i_power_phy(struct stmmac_priv *priv)
-{
- int ret;
-
- ret = sun8i_dwmac_power_internal_phy(priv);
- if (ret)
- return ret;
-
- ret = sun8i_dwmac_set_syscon(priv);
- if (ret)
- return ret;
-
- /* After changing syscon value, the MAC need reset or it will use
- * the last value (and so the last PHY set.
- */
- ret = sun8i_dwmac_reset(priv);
- if (ret)
- return ret;
- return 0;
-}
-
-static void sun8i_unpower_phy(struct sunxi_priv_data *gmac)
-{
sun8i_dwmac_unset_syscon(gmac);
- sun8i_dwmac_unpower_internal_phy(gmac);
-}
-static void sun8i_dwmac_exit(struct platform_device *pdev, void *priv)
-{
- struct sunxi_priv_data *gmac = priv;
-
- sun8i_unpower_phy(gmac);
+ reset_control_put(gmac->rst_ephy);
clk_disable_unprepare(gmac->tx_clk);
@@ -849,7 +942,7 @@ static struct mac_device_info *sun8i_dwmac_setup(void *ppriv)
if (!mac)
return NULL;
- ret = sun8i_power_phy(priv);
+ ret = sun8i_dwmac_set_syscon(priv);
if (ret)
return NULL;
@@ -889,6 +982,8 @@ static int sun8i_dwmac_probe(struct platform_device *pdev)
struct sunxi_priv_data *gmac;
struct device *dev = &pdev->dev;
int ret;
+ struct stmmac_priv *priv;
+ struct net_device *ndev;
ret = stmmac_get_platform_resources(pdev, &stmmac_res);
if (ret)
@@ -932,29 +1027,6 @@ static int sun8i_dwmac_probe(struct platform_device *pdev)
}
plat_dat->interface = of_get_phy_mode(dev->of_node);
- if (plat_dat->interface == gmac->variant->internal_phy) {
- dev_info(&pdev->dev, "Will use internal PHY\n");
- gmac->use_internal_phy = true;
- gmac->ephy_clk = of_clk_get(plat_dat->phy_node, 0);
- if (IS_ERR(gmac->ephy_clk)) {
- ret = PTR_ERR(gmac->ephy_clk);
- dev_err(&pdev->dev, "Cannot get EPHY clock: %d\n", ret);
- return -EINVAL;
- }
-
- gmac->rst_ephy = of_reset_control_get(plat_dat->phy_node, NULL);
- if (IS_ERR(gmac->rst_ephy)) {
- ret = PTR_ERR(gmac->rst_ephy);
- if (ret == -EPROBE_DEFER)
- return ret;
- dev_err(&pdev->dev, "No EPHY reset control found %d\n",
- ret);
- return -EINVAL;
- }
- } else {
- dev_info(&pdev->dev, "Will use external PHY\n");
- gmac->use_internal_phy = false;
- }
/* platform data specifying hardware features and callbacks.
* hardware features were copied from Allwinner drivers.
@@ -973,9 +1045,34 @@ static int sun8i_dwmac_probe(struct platform_device *pdev)
ret = stmmac_dvr_probe(&pdev->dev, plat_dat, &stmmac_res);
if (ret)
- sun8i_dwmac_exit(pdev, plat_dat->bsp_priv);
+ goto dwmac_exit;
+
+ ndev = dev_get_drvdata(&pdev->dev);
+ priv = netdev_priv(ndev);
+ /* The mux must be registered after parent MDIO
+ * so after stmmac_dvr_probe()
+ */
+ if (gmac->variant->soc_has_internal_phy) {
+ ret = get_ephy_nodes(priv);
+ if (ret)
+ return ret;
+ ret = sun8i_dwmac_register_mdio_mux(priv);
+ if (ret) {
+ dev_err(&pdev->dev, "Failed to register mux\n");
+ goto dwmac_mux;
+ }
+ } else {
+ ret = sun8i_dwmac_reset(priv);
+ if (ret)
+ goto dwmac_exit;
+ }
return ret;
+dwmac_mux:
+ sun8i_dwmac_unset_syscon(gmac);
+dwmac_exit:
+ sun8i_dwmac_exit(pdev, plat_dat->bsp_priv);
+return ret;
}
static const struct of_device_id sun8i_dwmac_match[] = {
--
2.13.6
^ permalink raw reply
* Re: [PATCH 00/12] radix-tree: split out struct radix_tree_root out to <linux/radix-tree-root.h>
From: Leon Romanovsky @ 2017-10-08 18:52 UTC (permalink / raw)
To: Masahiro Yamada
Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA, Thomas Gleixner,
Andrew Morton, Ian Abbott, Ingo Molnar, Linus Torvalds,
linux-cachefs-H+wXaHxf7aLQT0dZR+AlfA,
linux-sh-u79uwXL29TY76Z2rM5mHXA, Rodrigo Vivi,
dri-devel-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW, David Airlie,
linux-rdma-u79uwXL29TY76Z2rM5mHXA, Yoshinori Sato, Tariq Toukan,
Rich Felker, Jani Nikula, J. Bruce Fields, David Howells,
intel-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW, Yishai Hadas,
Joonas Lahtinen
In-Reply-To: <1507479013-5207-1-git-send-email-yamada.masahiro-uWyLwvC0a2jby3iVrkZq2A@public.gmane.org>
[-- Attachment #1: Type: text/plain, Size: 2655 bytes --]
On Mon, Oct 09, 2017 at 01:10:01AM +0900, Masahiro Yamada wrote:
<...>
>
> By splitting out the radix_tree_root definition,
> we can reduce the header file dependency.
>
> Reducing the header dependency will help for speeding the kernel
> build, suppressing unnecessary recompile of objects during
> git-bisect'ing, etc.
If we judge by the diffstat of this series, there won't be any
visible change in anything mentioned above.
<...>
>
> Masahiro Yamada (12):
> radix-tree: replace <linux/spinlock.h> with <linux/spinlock_types.h>
> radix-tree: split struct radix_tree_root to <linux/radix-tree-root.h>
> irqdomain: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
> writeback: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
> iocontext.h: replace <linux/radix-tree.h> with
> <linux/radix-tree-root.h>
> fs: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
> blkcg: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
> fscache: include <linux-radix-tree.h>
> sh: intc: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
> net/mlx4: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
> net/mlx5: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
> drm/i915: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
>
> drivers/gpu/drm/i915/i915_gem.c | 1 +
> drivers/gpu/drm/i915/i915_gem_context.c | 1 +
> drivers/gpu/drm/i915/i915_gem_context.h | 2 +-
> drivers/gpu/drm/i915/i915_gem_execbuffer.c | 1 +
> drivers/gpu/drm/i915/i915_gem_object.h | 1 +
> drivers/net/ethernet/mellanox/mlx4/cq.c | 1 +
> drivers/net/ethernet/mellanox/mlx4/mlx4.h | 2 +-
> drivers/net/ethernet/mellanox/mlx4/qp.c | 1 +
> drivers/net/ethernet/mellanox/mlx4/srq.c | 1 +
> drivers/sh/intc/internals.h | 2 +-
> include/linux/backing-dev-defs.h | 2 +-
> include/linux/blk-cgroup.h | 2 +-
> include/linux/fs.h | 2 +-
> include/linux/fscache.h | 1 +
> include/linux/iocontext.h | 2 +-
> include/linux/irqdomain.h | 2 +-
> include/linux/mlx4/device.h | 2 +-
> include/linux/mlx4/qp.h | 1 +
> include/linux/mlx5/driver.h | 2 +-
> include/linux/mlx5/qp.h | 1 +
> include/linux/radix-tree-root.h | 24 ++++++++++++++++++++++++
> include/linux/radix-tree.h | 8 ++------
> 22 files changed, 46 insertions(+), 16 deletions(-)
> create mode 100644 include/linux/radix-tree-root.h
>
> --
> 2.7.4
>
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply
* Re: [PATCH 10/12] net/mlx4: replace <linux/radix-tree.h> with <linux/radix-tree-root.h>
From: Leon Romanovsky @ 2017-10-08 18:55 UTC (permalink / raw)
To: Masahiro Yamada
Cc: David Miller, Linux Kernel Mailing List, Thomas Gleixner,
Andrew Morton, abbotti, Ingo Molnar, Linus Torvalds, linux-rdma,
yishaih, tariqt, netdev
In-Reply-To: <CAK7LNAQ1BeSb14poieRU4uMsnsEHPNL2zSNtWi1TGN=YrK_8jA@mail.gmail.com>
[-- Attachment #1: Type: text/plain, Size: 1951 bytes --]
On Mon, Oct 09, 2017 at 02:29:15AM +0900, Masahiro Yamada wrote:
> 2017-10-09 2:00 GMT+09:00 David Miller <davem@davemloft.net>:
> > From: Masahiro Yamada <yamada.masahiro@socionext.com>
> > Date: Mon, 9 Oct 2017 01:10:11 +0900
> >
> >> The headers
> >> - include/linux/mlx4/device.h
> >> - drivers/net/ethernet/mellanox/mlx4/mlx4.h
> >> require the definition of struct radix_tree_root, but do not need to
> >> know anything about other radix tree stuff.
> >>
> >> Include <linux/radix-tree-root.h> instead of <linux/radix-tree.h> to
> >> reduce the header dependency.
> >>
> >> While we are here, let's add missing <linux/radix-tree.h> where
> >> radix tree accessors are used.
> >>
> >> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
> >
> > Honestly this makes things more complicated.
>
>
> The idea is simple; include necessary headers explicitly.
>
> Putting everything into one common header
> means most of C files are forced to parse unnecessary headers.
It is neglected, only first caller will actually parse that header file,
other callers will check the #ifndef pragma without need to reparse the
whole file.
>
>
>
> > The driver was trying to consolidate all of the header needs
> > by including them all in one place, the main driver header.
> >
> > Now you're including headers in several different files.
> >
> > I really don't like the results of this change and would
> > ask you to reconsider.
> >
> > Just add both radix-tree-root.h _and_ radix-tree.h to mlx4.h
> > and leave the rest of the driver alone.
>
>
> If you do not like this, you can just throw it away.
>
> <linux/radix-tree.h> includes <linux/radix-tree-root.h>.
> You do not need to include both.
>
>
>
>
> --
> Best Regards
> Masahiro Yamada
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
^ permalink raw reply
* [PATCH] net: ethernet: stmmac: Clean up dead code
From: Christos Gkekas @ 2017-10-08 19:13 UTC (permalink / raw)
To: Giuseppe Cavallaro, Alexandre Torgue, netdev, linux-kernel
Cc: Christos Gkekas
Many macros in dwmac-ipq806x are unused and should be removed.
Moreover gmac->id is an unsigned variable and therefore checking
whether it is less than zero is redundant.
Signed-off-by: Christos Gkekas <chris.gekas@gmail.com>
---
drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c | 14 +-------------
1 file changed, 1 insertion(+), 13 deletions(-)
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c
index 866444b..2c6d7c6 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-ipq806x.c
@@ -51,15 +51,11 @@
#define NSS_COMMON_CLK_SRC_CTRL_RGMII(x) 1
#define NSS_COMMON_CLK_SRC_CTRL_SGMII(x) ((x >= 2) ? 1 : 0)
-#define NSS_COMMON_MACSEC_CTL 0x28
-#define NSS_COMMON_MACSEC_CTL_EXT_BYPASS_EN(x) (1 << x)
-
#define NSS_COMMON_GMAC_CTL(x) (0x30 + (x * 4))
#define NSS_COMMON_GMAC_CTL_CSYS_REQ BIT(19)
#define NSS_COMMON_GMAC_CTL_PHY_IFACE_SEL BIT(16)
#define NSS_COMMON_GMAC_CTL_IFG_LIMIT_OFFSET 8
#define NSS_COMMON_GMAC_CTL_IFG_OFFSET 0
-#define NSS_COMMON_GMAC_CTL_IFG_MASK 0x3f
#define NSS_COMMON_CLK_DIV_RGMII_1000 1
#define NSS_COMMON_CLK_DIV_RGMII_100 9
@@ -68,9 +64,6 @@
#define NSS_COMMON_CLK_DIV_SGMII_100 4
#define NSS_COMMON_CLK_DIV_SGMII_10 49
-#define QSGMII_PCS_MODE_CTL 0x68
-#define QSGMII_PCS_MODE_CTL_AUTONEG_EN(x) BIT((x * 8) + 7)
-
#define QSGMII_PCS_CAL_LCKDT_CTL 0x120
#define QSGMII_PCS_CAL_LCKDT_CTL_RST BIT(19)
@@ -83,15 +76,10 @@
#define QSGMII_PHY_TX_DRIVER_EN BIT(3)
#define QSGMII_PHY_QSGMII_EN BIT(7)
#define QSGMII_PHY_PHASE_LOOP_GAIN_OFFSET 12
-#define QSGMII_PHY_PHASE_LOOP_GAIN_MASK 0x7
#define QSGMII_PHY_RX_DC_BIAS_OFFSET 18
-#define QSGMII_PHY_RX_DC_BIAS_MASK 0x3
#define QSGMII_PHY_RX_INPUT_EQU_OFFSET 20
-#define QSGMII_PHY_RX_INPUT_EQU_MASK 0x3
#define QSGMII_PHY_CDR_PI_SLEW_OFFSET 22
-#define QSGMII_PHY_CDR_PI_SLEW_MASK 0x3
#define QSGMII_PHY_TX_DRV_AMP_OFFSET 28
-#define QSGMII_PHY_TX_DRV_AMP_MASK 0xf
struct ipq806x_gmac {
struct platform_device *pdev;
@@ -217,7 +205,7 @@ static int ipq806x_gmac_of_parse(struct ipq806x_gmac *gmac)
* code and keep it consistent with the Linux convention, we'll number
* them from 0 to 3 here.
*/
- if (gmac->id < 0 || gmac->id > 3) {
+ if (gmac->id > 3) {
dev_err(dev, "invalid gmac id\n");
return -EINVAL;
}
--
2.7.4
^ permalink raw reply related
* [PATCH] net/core: Fix BUG to BUG_ON conditionals.
From: Tim Hansen @ 2017-10-08 19:17 UTC (permalink / raw)
To: davem
Cc: willemb, edumazet, soheil, elena.reshetova, pabeni, tom, Jason,
fw, netdev, linux-kernel, alexander.levin, devtimhansen
Fix BUG() calls to use BUG_ON(conditional) macros.
This was found using make coccicheck M=net/core on linux next
tag next-20170929.
Signed-off-by: Tim Hansen <devtimhansen@gmail.com>
---
net/core/skbuff.c | 15 ++++++---------
1 file changed, 6 insertions(+), 9 deletions(-)
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index d98c2e3ce2bf..461516f45b33 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -1350,8 +1350,7 @@ struct sk_buff *skb_copy(const struct sk_buff *skb, gfp_t gfp_mask)
/* Set the tail pointer and length */
skb_put(n, skb->len);
- if (skb_copy_bits(skb, -headerlen, n->head, headerlen + skb->len))
- BUG();
+ BUG_ON(skb_copy_bits(skb, -headerlen, n->head, headerlen + skb->len));
copy_skb_header(n, skb);
return n;
@@ -1449,8 +1448,7 @@ int pskb_expand_head(struct sk_buff *skb, int nhead, int ntail,
BUG_ON(nhead < 0);
- if (skb_shared(skb))
- BUG();
+ BUG_ON(skb_shared(skb));
size = SKB_DATA_ALIGN(size);
@@ -1595,9 +1593,8 @@ struct sk_buff *skb_copy_expand(const struct sk_buff *skb,
head_copy_off = newheadroom - head_copy_len;
/* Copy the linear header and data. */
- if (skb_copy_bits(skb, -head_copy_len, n->head + head_copy_off,
- skb->len + head_copy_len))
- BUG();
+ BUG_ON(skb_copy_bits(skb, -head_copy_len, n->head + head_copy_off,
+ skb->len + head_copy_len));
copy_skb_header(n, skb);
@@ -1878,8 +1875,8 @@ void *__pskb_pull_tail(struct sk_buff *skb, int delta)
return NULL;
}
- if (skb_copy_bits(skb, skb_headlen(skb), skb_tail_pointer(skb), delta))
- BUG();
+ BUG_ON(skb_copy_bits(skb, skb_headlen(skb),
+ skb_tail_pointer(skb), delta))
/* Optimization: no fragments, no reasons to preestimate
* size of pulled pages. Superb.
--
2.14.2
^ permalink raw reply related
* [PATCH] ath10k: spectral: Simplify error checking
From: Christos Gkekas @ 2017-10-08 19:30 UTC (permalink / raw)
To: Kalle Valo, ath10k, linux-wireless, netdev, linux-kernel; +Cc: Christos Gkekas
Variable val is unsigned so checking whether it is less than zero is
redundant.
Signed-off-by: Christos Gkekas <chris.gekas@gmail.com>
---
drivers/net/wireless/ath/ath10k/spectral.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/drivers/net/wireless/ath/ath10k/spectral.c b/drivers/net/wireless/ath/ath10k/spectral.c
index dd9cc09..1867937 100644
--- a/drivers/net/wireless/ath/ath10k/spectral.c
+++ b/drivers/net/wireless/ath/ath10k/spectral.c
@@ -403,10 +403,7 @@ static ssize_t write_file_spectral_count(struct file *file,
return -EFAULT;
buf[len] = '\0';
- if (kstrtoul(buf, 0, &val))
- return -EINVAL;
-
- if (val < 0 || val > 255)
+ if (kstrtoul(buf, 0, &val) || val > 255)
return -EINVAL;
mutex_lock(&ar->conf_mutex);
--
2.7.4
^ permalink raw reply related
* Re: [PATCH v2] net/core: Fix BUG to BUG_ON conditionals.
From: Tim Hansen @ 2017-10-08 20:03 UTC (permalink / raw)
To: davem
Cc: willemb, edumazet, soheil, elena.reshetova, pabeni, tom, Jason,
fw, netdev, linux-kernel, alexander.levin
In-Reply-To: <20171008191720.prewqllyazdstwkp@debian>
Mistakenly sent the patch previously with a missing semicolon.
Apologies.
Fix BUG() calls to use BUG_ON(conditional) macros.
This was found using make coccicheck M=net/core on linux next
tag next-20170929
Signed-off-by: Tim Hansen <devtimhansen@gmail.com>
---
net/core/skbuff.c | 15 ++++++---------
1 file changed, 6 insertions(+), 9 deletions(-)
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index d98c2e3ce2bf..34ce4c1a0f3c 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -1350,8 +1350,7 @@ struct sk_buff *skb_copy(const struct sk_buff *skb, gfp_t gfp_mask)
/* Set the tail pointer and length */
skb_put(n, skb->len);
- if (skb_copy_bits(skb, -headerlen, n->head, headerlen + skb->len))
- BUG();
+ BUG_ON(skb_copy_bits(skb, -headerlen, n->head, headerlen + skb->len));
copy_skb_header(n, skb);
return n;
@@ -1449,8 +1448,7 @@ int pskb_expand_head(struct sk_buff *skb, int nhead, int ntail,
BUG_ON(nhead < 0);
- if (skb_shared(skb))
- BUG();
+ BUG_ON(skb_shared(skb));
size = SKB_DATA_ALIGN(size);
@@ -1595,9 +1593,8 @@ struct sk_buff *skb_copy_expand(const struct sk_buff *skb,
head_copy_off = newheadroom - head_copy_len;
/* Copy the linear header and data. */
- if (skb_copy_bits(skb, -head_copy_len, n->head + head_copy_off,
- skb->len + head_copy_len))
- BUG();
+ BUG_ON(skb_copy_bits(skb, -head_copy_len, n->head + head_copy_off,
+ skb->len + head_copy_len));
copy_skb_header(n, skb);
@@ -1878,8 +1875,8 @@ void *__pskb_pull_tail(struct sk_buff *skb, int delta)
return NULL;
}
- if (skb_copy_bits(skb, skb_headlen(skb), skb_tail_pointer(skb), delta))
- BUG();
+ BUG_ON(skb_copy_bits(skb, skb_headlen(skb),
+ skb_tail_pointer(skb), delta));
/* Optimization: no fragments, no reasons to preestimate
* size of pulled pages. Superb.
--
2.14.2
^ permalink raw reply related
* RIF/VRF overflow in spectrum and reporting errors back to user
From: David Ahern @ 2017-10-08 20:10 UTC (permalink / raw)
To: Ido Schimmel, Jiri Pirko; +Cc: netdev@vger.kernel.org
Jiri / Ido:
I am looking at adding user messages for spectrum failures related to
RIF and VRF overflow coming from the inetaddr and inet6addr notifier
paths. The key is that if the notifiers fail the address add needs to
fail and an error reported to the user as to what happened.
Earlier this year 3ad7d2468f79f added in_validator_info and
in6_validator_info as a way for the notifiers to fail adding an address.
Adding support to spectrum for that notifier is complicated by the fact
that the validator notifier and address notifiers will come in back to
back for the NETDEV_UP case. Ignoring NETDEV_UP in
mlxsw_sp_inetaddr_event seems ok for IPv6 but not clear for IPv4 since
the NETDEV_UP case is emitted on an address delete that involves a
promotion. Handling the back to back NETDEV_UP is complicated since
functions invoked by __mlxsw_sp_inetaddr_event can take multiple
references. Specifically, in mlxsw_sp_port_vlan_router_join():
fid = rif->ops->fid_get(rif);
Can NETDEV_UP be ignored for the inetaddr notifier if it is handled by
the validator notitifer?
David
^ permalink raw reply
* [PATCH] ath9k: debug: Simplify error checking
From: Christos Gkekas @ 2017-10-08 21:29 UTC (permalink / raw)
To: QCA ath9k Development, Kalle Valo, linux-wireless, netdev,
linux-kernel
Cc: Christos Gkekas
Variable val is unsigned so checking whether it is less than zero is
redundant.
Signed-off-by: Christos Gkekas <chris.gekas@gmail.com>
---
drivers/net/wireless/ath/ath9k/debug.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/drivers/net/wireless/ath/ath9k/debug.c b/drivers/net/wireless/ath/ath9k/debug.c
index 01fa301..3b93c23 100644
--- a/drivers/net/wireless/ath/ath9k/debug.c
+++ b/drivers/net/wireless/ath/ath9k/debug.c
@@ -1164,10 +1164,7 @@ static ssize_t write_file_tpc(struct file *file, const char __user *user_buf,
return -EFAULT;
buf[len] = '\0';
- if (kstrtoul(buf, 0, &val))
- return -EINVAL;
-
- if (val < 0 || val > 1)
+ if (kstrtoul(buf, 0, &val) || val > 1)
return -EINVAL;
tpc_enabled = !!val;
--
2.7.4
^ permalink raw reply related
* [PATCH] qed: Delete redundant check on dcb_app priority
From: Christos Gkekas @ 2017-10-08 22:46 UTC (permalink / raw)
To: Ariel Elior, everest-linux-l2, netdev, linux-kernel; +Cc: Christos Gkekas
dcb_app priority is unsigned thus checking whether it is less than zero
is redundant.
Signed-off-by: Christos Gkekas <chris.gekas@gmail.com>
---
drivers/net/ethernet/qlogic/qed/qed_dcbx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qlogic/qed/qed_dcbx.c b/drivers/net/ethernet/qlogic/qed/qed_dcbx.c
index 8f6ccc0..6e15d3c 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_dcbx.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_dcbx.c
@@ -2308,7 +2308,7 @@ static int qed_dcbnl_ieee_setapp(struct qed_dev *cdev, struct dcb_app *app)
DP_VERBOSE(hwfn, QED_MSG_DCB, "selector = %d protocol = %d pri = %d\n",
app->selector, app->protocol, app->priority);
- if (app->priority < 0 || app->priority >= QED_MAX_PFC_PRIORITIES) {
+ if (app->priority >= QED_MAX_PFC_PRIORITIES) {
DP_INFO(hwfn, "Invalid priority %d\n", app->priority);
return -EINVAL;
}
--
2.7.4
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox