* Re: [PATCH] net: sh_eth: add support for SH7757's ETHER
From: David Miller @ 2010-07-07 22:45 UTC (permalink / raw)
To: yoshihiro.shimoda.uh; +Cc: netdev, linux-sh
In-Reply-To: <4C32B1F2.4040106@renesas.com>
From: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Date: Tue, 06 Jul 2010 13:32:50 +0900
> The SH7757 has 2 Fast Ethernet controller (ETHER) and 2 Gigabit Ethernet
> Controller (GETHER). This patch supports 2 ETHER only.
>
> Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Applied to net-next-2.6, thanks.
^ permalink raw reply
* Re: [net-next PATCH] bnx2x: Set RXHASH for LRO packets
From: David Miller @ 2010-07-07 22:45 UTC (permalink / raw)
To: vladz; +Cc: netdev, eilong
In-Reply-To: <1278425383.22257.4.camel@lb-tlvb-vladz>
From: "Vladislav Zolotarov" <vladz@broadcom.com>
Date: Tue, 6 Jul 2010 17:09:43 +0300
> Set Toeplitz hash both for LRO and none-LRO skbs.
> The first CQE (TPA_START) will contain a hash for an LRO packet.
>
> Current code sets skb->rx_hash for none-LRO skbs only.
>
> Signed-off-by: Vladislav Zolotarov <vladz@broadcom.com>
> Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Applied.
^ permalink raw reply
* Re: [PATCH net-next-2.6 V2] net: fix 64 bit counters on 32 bit arches
From: David Miller @ 2010-07-07 22:41 UTC (permalink / raw)
To: eric.dumazet; +Cc: bhutchings, shemminger, arnd, netdev, linux-net-drivers
In-Reply-To: <20100707.143645.135512197.davem@davemloft.net>
From: David Miller <davem@davemloft.net>
Date: Wed, 07 Jul 2010 14:36:45 -0700 (PDT)
> From: Eric Dumazet <eric.dumazet@gmail.com>
> Date: Mon, 05 Jul 2010 22:05:22 +0200
>
>> [PATCH net-next-2.6 V2] net: fix 64 bit counters on 32 bit arches
Some dev_get_stats() conversions were missing, such as parisc/led,
some s390 code, usb rndis, etc.
I took care of this when integrating this patch.
^ permalink raw reply
* Re: [PATCH v3 0/9] atm: propagate atm_dev signal carrier to LOWER_UP of netdevice
From: David Miller @ 2010-07-07 22:07 UTC (permalink / raw)
To: karl; +Cc: linux-atm-general, netdev, chas
In-Reply-To: <1278492636-11094-1-git-send-email-karl@hiramoto.org>
From: Karl Hiramoto <karl@hiramoto.org>
Date: Wed, 7 Jul 2010 10:50:27 +0200
> Changes from v2:
> * use atomic instead of blocking notifier
> * use read_lock_irq() instead of read_lock() in atm/br2684
> * clean up comments
> * remove unused variable. I feel really bad about missing that last time.
>
> Changes from v1:
> Use atm_dev notifier chain instead of callback function pointer in struct vcc.
> In drivers/usb/atm call atm_dev_signal_change().
>
> In userspace it's helpful to know if a network device has a carrier signal.
> Often it is monitored via netlink. This patchset allows a way for the
> struct atm_dev drivers to pass carrier on/off to the netdevice.
>
> For DSL, carrier is on when the line has reached showtime state.
>
> Currently this patchset only propagates the changes to br2684 vccs,
> as this is the only type of hardware I have to test.
>
> If you prefer git you can pull from:
> git://github.com/karlhiramoto/linux-2.6.git atm-v3
I think the locking still needs another tweak.
By using read_lock_irq() you are assuming that you are invoked
from a context where irqs are disabled.
That's not necessarily the case, in fact some of your notifier call
sites in the drivers are in interrupt handlers where interrupts may or
may not be disabled.
So you'll likely need to use read_lock_irqsave() and read_lock_irqrestore().
Next, please format comments:
/* Like
* this.
*/
Thanks.
^ permalink raw reply
* Re: NET: SB1250: Initialize .owner
From: David Miller @ 2010-07-07 22:01 UTC (permalink / raw)
To: ralf; +Cc: netdev, linux-mips, macro
In-Reply-To: <20100706151811.GA4829@linux-mips.org>
From: Ralf Baechle <ralf@linux-mips.org>
Date: Tue, 6 Jul 2010 16:18:11 +0100
> Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Applied, thanks Ralf.
^ permalink raw reply
* Re: [PATCH] vxge: show startup message with KERN_INFO
From: David Miller @ 2010-07-07 21:52 UTC (permalink / raw)
To: jon.mason; +Cc: fengguang.wu, Sreenivasa.Honnur, netdev
In-Reply-To: <20100706160824.GA31635@exar.com>
From: Jon Mason <jon.mason@exar.com>
Date: Tue, 6 Jul 2010 11:08:25 -0500
> On Tue, Jul 06, 2010 at 06:02:03AM -0700, Wu Fengguang wrote:
>> The original KERN_CRIT will mess up terminals.
>
> Looks reasonable to me.
...
>> CC: Sreenivasa Honnur <Sreenivasa.Honnur@neterion.com>
>> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Applied, thanks.
^ permalink raw reply
* Re: [PATCH kernel 2.6.35-rc3-git7] axnet_cs: local_irq_save before calling ei_irq_wrapper
From: David Miller @ 2010-07-07 21:51 UTC (permalink / raw)
To: ken_kawasaki; +Cc: netdev
In-Reply-To: <20100706202243.95319ab4.ken_kawasaki@spring.nifty.jp>
From: Ken Kawasaki <ken_kawasaki@spring.nifty.jp>
Date: Tue, 6 Jul 2010 20:22:43 +0900
>
> axnet_cs:
> local_irq_save before calling ei_irq_wrapper.
>
>
> Signed-off-by: Ken Kawasaki <ken_kawasaki@spring.nifty.jp>
An interrupt handler may not assume that interrupts are disabled when
it is invoked.
The axnet_interrupt() code needs to be fixed to use
spin_lock_irqsave() et al.
Then, this change of your's is no longer needed.
^ permalink raw reply
* Re: [PATCH -net-2.6] ll_temac: Fix missing iounmaps
From: David Miller @ 2010-07-07 21:48 UTC (permalink / raw)
To: dkirjanov; +Cc: john.linn, brian.hill, netdev
In-Reply-To: <20100706074420.GA15451@hera.kernel.org>
From: Denis Kirjanov <dkirjanov@kernel.org>
Date: Tue, 6 Jul 2010 07:44:20 +0000
> Fix missing iounmaps.
>
> Signed-off-by: Denis Kirjanov <dkirjanov@kernel.org>
Applied, thanks.
^ permalink raw reply
* Re: Possible bug in net/ipv4/route.c?
From: David Miller @ 2010-07-07 21:46 UTC (permalink / raw)
To: herbert
Cc: gjin, shemminger, skavy, linux-kernel, gren, msezgin, silgen,
netdev
In-Reply-To: <20100706072928.GB14612@gondor.apana.org.au>
From: Herbert Xu <herbert@gondor.apana.org.au>
Date: Tue, 6 Jul 2010 15:29:28 +0800
> bridge: Clear IPCB before possible entry into IP stack
>
> The bridge protocol lives dangerously by having incestuous relations
> with the IP stack. In this instance an abomination has been created
> where a bogus IPCB area from a bridged packet leads to a crash in
> the IP stack because it's interpreted as IP options.
>
> This patch papers over the problem by clearing the IPCB area in that
> particular spot. To fix this properly we'd also need to parse any
> IP options if present but I'm way too lazy for that.
>
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Applied, thanks a lot!
^ permalink raw reply
* [PATCH 14/18] irda/irnet: use noop_llseek
From: Arnd Bergmann @ 2010-07-07 21:40 UTC (permalink / raw)
To: linux-kernel
Cc: John Kacur, Frederic Weisbecker, Arnd Bergmann, Christoph Hellwig,
Samuel Ortiz, netdev
In-Reply-To: <1278538820-1392-1-git-send-email-arnd@arndb.de>
There may be applications trying to seek
on the irnet character device, so we should
use noop_llseek to avoid returning an error
when the default llseek changes to no_llseek.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Samuel Ortiz <samuel@sortiz.org>
Cc: netdev@vger.kernel.org
---
net/irda/irnet/irnet_ppp.h | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/net/irda/irnet/irnet_ppp.h b/net/irda/irnet/irnet_ppp.h
index b5df241..7b9949e 100644
--- a/net/irda/irnet/irnet_ppp.h
+++ b/net/irda/irnet/irnet_ppp.h
@@ -104,6 +104,7 @@ static const struct file_operations irnet_device_fops =
.unlocked_ioctl = dev_irnet_ioctl,
.open = dev_irnet_open,
.release = dev_irnet_close
+ .llseek = noop_llseek,
/* Also : llseek, readdir, mmap, flush, fsync, fasync, lock, readv, writev */
};
--
1.7.1
^ permalink raw reply related
* [PATCH 02/18] net/wireless: use generic_file_llseek in debugfs
From: Arnd Bergmann @ 2010-07-07 21:40 UTC (permalink / raw)
To: linux-kernel
Cc: John Kacur, Frederic Weisbecker, Arnd Bergmann, Christoph Hellwig,
John W. Linville, linux-wireless, netdev
In-Reply-To: <1278538820-1392-1-git-send-email-arnd@arndb.de>
The default llseek operation is changing from
default_llseek to no_llseek, so all code relying on
the current behaviour needs to make that explicit.
The wireless driver infrastructure and some of the drivers
make use of generated debugfs files, so they cannot
be converted by our script that automatically determines
the right operation.
All these files use debugfs and they typically rely
on simple_read_from_buffer, so the best llseek operation
here is generic_file_llseek.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: "John W. Linville" <linville@tuxdriver.com>
Cc: linux-wireless@vger.kernel.org
Cc: netdev@vger.kernel.org
---
drivers/misc/iwmc3200top/debugfs.c | 3 +++
drivers/net/wireless/b43/debugfs.c | 1 +
drivers/net/wireless/b43legacy/debugfs.c | 1 +
drivers/net/wireless/iwlwifi/iwl-debugfs.c | 3 +++
drivers/net/wireless/libertas/debugfs.c | 1 +
drivers/net/wireless/rt2x00/rt2x00debug.c | 1 +
drivers/net/wireless/wl12xx/wl1251_debugfs.c | 2 ++
drivers/net/wireless/wl12xx/wl1271_debugfs.c | 2 ++
net/mac80211/debugfs.c | 2 ++
net/mac80211/debugfs_key.c | 2 ++
net/mac80211/debugfs_netdev.c | 1 +
net/mac80211/debugfs_sta.c | 2 ++
net/wireless/debugfs.c | 1 +
13 files changed, 22 insertions(+), 0 deletions(-)
diff --git a/drivers/misc/iwmc3200top/debugfs.c b/drivers/misc/iwmc3200top/debugfs.c
index e9eda47..62fbaec 100644
--- a/drivers/misc/iwmc3200top/debugfs.c
+++ b/drivers/misc/iwmc3200top/debugfs.c
@@ -71,6 +71,7 @@ ssize_t iwmct_dbgfs_##name##_write(struct file *file, \
static const struct file_operations iwmct_dbgfs_##name##_ops = { \
.read = iwmct_dbgfs_##name##_read, \
.open = iwmct_dbgfs_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
#define DEBUGFS_WRITE_FILE_OPS(name) \
@@ -78,6 +79,7 @@ ssize_t iwmct_dbgfs_##name##_write(struct file *file, \
static const struct file_operations iwmct_dbgfs_##name##_ops = { \
.write = iwmct_dbgfs_##name##_write, \
.open = iwmct_dbgfs_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
#define DEBUGFS_READ_WRITE_FILE_OPS(name) \
@@ -87,6 +89,7 @@ ssize_t iwmct_dbgfs_##name##_write(struct file *file, \
.write = iwmct_dbgfs_##name##_write, \
.read = iwmct_dbgfs_##name##_read, \
.open = iwmct_dbgfs_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
diff --git a/drivers/net/wireless/b43/debugfs.c b/drivers/net/wireless/b43/debugfs.c
index 80b19a4..59f59fa 100644
--- a/drivers/net/wireless/b43/debugfs.c
+++ b/drivers/net/wireless/b43/debugfs.c
@@ -627,6 +627,7 @@ out_unlock:
.open = b43_debugfs_open, \
.read = b43_debugfs_read, \
.write = b43_debugfs_write, \
+ .llseek = generic_file_llseek, \
}, \
.file_struct_offset = offsetof(struct b43_dfsentry, \
file_##name), \
diff --git a/drivers/net/wireless/b43legacy/debugfs.c b/drivers/net/wireless/b43legacy/debugfs.c
index 1f85ac5..f232618 100644
--- a/drivers/net/wireless/b43legacy/debugfs.c
+++ b/drivers/net/wireless/b43legacy/debugfs.c
@@ -334,6 +334,7 @@ out_unlock:
.open = b43legacy_debugfs_open, \
.read = b43legacy_debugfs_read, \
.write = b43legacy_debugfs_write, \
+ .llseek = generic_file_llseek, \
}, \
.file_struct_offset = offsetof(struct b43legacy_dfsentry, \
file_##name), \
diff --git a/drivers/net/wireless/iwlwifi/iwl-debugfs.c b/drivers/net/wireless/iwlwifi/iwl-debugfs.c
index 9659c5d..0d03cf2 100644
--- a/drivers/net/wireless/iwlwifi/iwl-debugfs.c
+++ b/drivers/net/wireless/iwlwifi/iwl-debugfs.c
@@ -87,6 +87,7 @@ static int iwl_dbgfs_open_file_generic(struct inode *inode, struct file *file)
static const struct file_operations iwl_dbgfs_##name##_ops = { \
.read = iwl_dbgfs_##name##_read, \
.open = iwl_dbgfs_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
#define DEBUGFS_WRITE_FILE_OPS(name) \
@@ -94,6 +95,7 @@ static const struct file_operations iwl_dbgfs_##name##_ops = { \
static const struct file_operations iwl_dbgfs_##name##_ops = { \
.write = iwl_dbgfs_##name##_write, \
.open = iwl_dbgfs_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
@@ -104,6 +106,7 @@ static const struct file_operations iwl_dbgfs_##name##_ops = { \
.write = iwl_dbgfs_##name##_write, \
.read = iwl_dbgfs_##name##_read, \
.open = iwl_dbgfs_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
int iwl_dbgfs_statistics_flag(struct iwl_priv *priv, char *buf, int bufsz)
diff --git a/drivers/net/wireless/libertas/debugfs.c b/drivers/net/wireless/libertas/debugfs.c
index de2caac..94f8f99 100644
--- a/drivers/net/wireless/libertas/debugfs.c
+++ b/drivers/net/wireless/libertas/debugfs.c
@@ -713,6 +713,7 @@ out_unlock:
.open = open_file_generic, \
.read = (fread), \
.write = (fwrite), \
+ .llseek = generic_file_llseek, \
}
struct lbs_debugfs_files {
diff --git a/drivers/net/wireless/rt2x00/rt2x00debug.c b/drivers/net/wireless/rt2x00/rt2x00debug.c
index e9fe93f..fe92500 100644
--- a/drivers/net/wireless/rt2x00/rt2x00debug.c
+++ b/drivers/net/wireless/rt2x00/rt2x00debug.c
@@ -508,6 +508,7 @@ static const struct file_operations rt2x00debug_fop_##__name = {\
.write = rt2x00debug_write_##__name, \
.open = rt2x00debug_file_open, \
.release = rt2x00debug_file_release, \
+ .llseek = generic_file_llseek, \
};
RT2X00DEBUGFS_OPS(csr, "0x%.8x\n", u32);
diff --git a/drivers/net/wireless/wl12xx/wl1251_debugfs.c b/drivers/net/wireless/wl12xx/wl1251_debugfs.c
index 5e4465a..a4ae7c4 100644
--- a/drivers/net/wireless/wl12xx/wl1251_debugfs.c
+++ b/drivers/net/wireless/wl12xx/wl1251_debugfs.c
@@ -50,6 +50,7 @@ static ssize_t name## _read(struct file *file, char __user *userbuf, \
static const struct file_operations name## _ops = { \
.read = name## _read, \
.open = wl1251_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
#define DEBUGFS_ADD(name, parent) \
@@ -86,6 +87,7 @@ static ssize_t sub## _ ##name## _read(struct file *file, \
static const struct file_operations sub## _ ##name## _ops = { \
.read = sub## _ ##name## _read, \
.open = wl1251_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
#define DEBUGFS_FWSTATS_ADD(sub, name) \
diff --git a/drivers/net/wireless/wl12xx/wl1271_debugfs.c b/drivers/net/wireless/wl12xx/wl1271_debugfs.c
index c239ef4..6e25303 100644
--- a/drivers/net/wireless/wl12xx/wl1271_debugfs.c
+++ b/drivers/net/wireless/wl12xx/wl1271_debugfs.c
@@ -51,6 +51,7 @@ static ssize_t name## _read(struct file *file, char __user *userbuf, \
static const struct file_operations name## _ops = { \
.read = name## _read, \
.open = wl1271_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
#define DEBUGFS_ADD(name, parent) \
@@ -87,6 +88,7 @@ static ssize_t sub## _ ##name## _read(struct file *file, \
static const struct file_operations sub## _ ##name## _ops = { \
.read = sub## _ ##name## _read, \
.open = wl1271_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
#define DEBUGFS_FWSTATS_ADD(sub, name) \
diff --git a/net/mac80211/debugfs.c b/net/mac80211/debugfs.c
index 637929b..31b6b89 100644
--- a/net/mac80211/debugfs.c
+++ b/net/mac80211/debugfs.c
@@ -36,6 +36,7 @@ static ssize_t name## _read(struct file *file, char __user *userbuf, \
static const struct file_operations name## _ops = { \
.read = name## _read, \
.open = mac80211_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
#define DEBUGFS_ADD(name) \
@@ -349,6 +350,7 @@ static ssize_t stats_ ##name## _read(struct file *file, \
static const struct file_operations stats_ ##name## _ops = { \
.read = stats_ ##name## _read, \
.open = mac80211_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
#define DEBUGFS_STATS_ADD(name) \
diff --git a/net/mac80211/debugfs_key.c b/net/mac80211/debugfs_key.c
index 97c9e46..1bc39ac 100644
--- a/net/mac80211/debugfs_key.c
+++ b/net/mac80211/debugfs_key.c
@@ -32,6 +32,7 @@ static ssize_t key_##name##_read(struct file *file, \
static const struct file_operations key_ ##name## _ops = { \
.read = key_##name##_read, \
.open = mac80211_open_file_generic, \
+ .llseek = generic_file_llseek, \
}
#define KEY_FILE(name, format) \
@@ -46,6 +47,7 @@ static const struct file_operations key_ ##name## _ops = { \
static const struct file_operations key_ ##name## _ops = { \
.read = key_conf_##name##_read, \
.open = mac80211_open_file_generic, \
+ .llseek = generic_file_llseek, \
}
#define KEY_CONF_FILE(name, format) \
diff --git a/net/mac80211/debugfs_netdev.c b/net/mac80211/debugfs_netdev.c
index 20b2998..8ad33ee 100644
--- a/net/mac80211/debugfs_netdev.c
+++ b/net/mac80211/debugfs_netdev.c
@@ -121,6 +121,7 @@ static const struct file_operations name##_ops = { \
.read = ieee80211_if_read_##name, \
.write = (_write), \
.open = mac80211_open_file_generic, \
+ .llseek = generic_file_llseek, \
}
#define __IEEE80211_IF_FILE_W(name) \
diff --git a/net/mac80211/debugfs_sta.c b/net/mac80211/debugfs_sta.c
index e763f15..73f8f36 100644
--- a/net/mac80211/debugfs_sta.c
+++ b/net/mac80211/debugfs_sta.c
@@ -37,6 +37,7 @@ static ssize_t sta_ ##name## _read(struct file *file, \
static const struct file_operations sta_ ##name## _ops = { \
.read = sta_##name##_read, \
.open = mac80211_open_file_generic, \
+ .llseek = generic_file_llseek, \
}
#define STA_OPS_RW(name) \
@@ -44,6 +45,7 @@ static const struct file_operations sta_ ##name## _ops = { \
.read = sta_##name##_read, \
.write = sta_##name##_write, \
.open = mac80211_open_file_generic, \
+ .llseek = generic_file_llseek, \
}
#define STA_FILE(name, field, format) \
diff --git a/net/wireless/debugfs.c b/net/wireless/debugfs.c
index a4991a3..3f9a57e 100644
--- a/net/wireless/debugfs.c
+++ b/net/wireless/debugfs.c
@@ -34,6 +34,7 @@ static ssize_t name## _read(struct file *file, char __user *userbuf, \
static const struct file_operations name## _ops = { \
.read = name## _read, \
.open = cfg80211_open_file_generic, \
+ .llseek = generic_file_llseek, \
};
DEBUGFS_READONLY_FILE(rts_threshold, 20, "%d",
--
1.7.1
^ permalink raw reply related
* Re: [PATCH net-next-2.6 V2] net: fix 64 bit counters on 32 bit arches
From: David Miller @ 2010-07-07 21:36 UTC (permalink / raw)
To: eric.dumazet; +Cc: bhutchings, shemminger, arnd, netdev, linux-net-drivers
In-Reply-To: <1278360322.2466.90.camel@edumazet-laptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Mon, 05 Jul 2010 22:05:22 +0200
> [PATCH net-next-2.6 V2] net: fix 64 bit counters on 32 bit arches
>
> There is a small possibility that a reader gets incorrect values on 32
> bit arches. SNMP applications could catch incorrect counters when a
> 32bit high part is changed by another stats consumer/provider.
>
> One way to solve this is to add a rtnl_link_stats64 param to all
> ndo_get_stats64() methods, and also add such a parameter to
> dev_get_stats().
>
> Rule is that we are not allowed to use dev->stats64 as a temporary
> storage for 64bit stats, but a caller provided area (usually on stack)
>
> Old drivers (only providing get_stats() method) need no changes.
>
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Looks great, applied, thanks Eric.
^ permalink raw reply
* Re: [PATCHv3 2/2] sfc: Implement 64-bit net device statistics on all architectures
From: David Miller @ 2010-07-07 19:56 UTC (permalink / raw)
To: eric.dumazet; +Cc: bhutchings, shemminger, arnd, netdev, linux-net-drivers
In-Reply-To: <1278354682.2877.639.camel@edumazet-laptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Mon, 05 Jul 2010 20:31:22 +0200
> One other way would be to add a rtnl_link_stats64 param to
> ndo_get_stats64() method, and ask drivers to copy their stats in this
> zone, instead of returning &dev->stats64 or something...
>
> And also change dev_get_stats() with this new parameter.
>
> Each caller would use a private copy, with no risk of concurrent
> updates.
If this can be reliably be made at all call sites, I think such
an approach is preferable.
^ permalink raw reply
* Re: [PATCH] bnx2x: add support for receive hashing
From: Tom Herbert @ 2010-07-07 19:17 UTC (permalink / raw)
To: Vladislav Zolotarov; +Cc: netdev@vger.kernel.org
In-Reply-To: <8628FE4E7912BF47A96AE7DD7BAC0AADDDE646FD52@SJEXCHCCR02.corp.ad.broadcom.com>
On Tue, Jul 6, 2010 at 12:16 AM, Vladislav Zolotarov <vladz@broadcom.com> wrote:
> Let me rephrase: the part of your patch below enables RSS flow in our FW even if there is only one HW queue and I wonder why?
> To refresh:
> 1) FW won't provide Toeplitz hash on CQE if RSS is not enabled.
> 2) There can be one HW queue in only 2 cases:
> - There is only one CPU in the system. In that case I wonder if u have anything to do with Toeplitz hash on the skb at all.
> - User has explicitly requested 1 HW queue with module parameter (num_queues=1). In that case if u r going to use the RSS hash on the skb means that u r actually going to do SW RSS instead of HW RSS, which sounds strange to me.
It might not be so strange. Previously, we has hit a firmware bug in
bnx2x that was make multi-queue not perform well under some loads, so
we had disabled it for a while... it is a valid configuration we have
used.
>
> So, Herbert, could u, pls., explain, what was your original idea about these code lines?
>
It is to enable the device to provide the RSS hash in RX descriptor.
The hash severs two purposes now, it's used internally in the device
to perform RSS table lookup and also value in RX descriptor. The
latter does not require multi-queue. Strictly speaking, on a single
processor system without multqueue, it would be true that enabling the
RX hash on bnx2x is currently superfluous (notwithstanding some other
use of the hash might be implemented).
> Thanks,
> vlad
>
>> -----Original Message-----
>> From: netdev-owner@vger.kernel.org [mailto:netdev-owner@vger.kernel.org] On
>> Behalf Of Vladislav Zolotarov
>> Sent: Sunday, July 04, 2010 7:36 PM
>> To: Tom Herbert
>> Cc: netdev@vger.kernel.org
>> Subject: RE: [PATCH] bnx2x: add support for receive hashing
>>
>> Tom, could u, pls., explain what did u mean by taking the (RSS) flags
>> configuration out of RSS "if"? To recall "if(is_multi(bp))" is true iff RSS
>> is enabled.
>>
>> Thanks,
>> vlad
>>
>> > @@ -5750,10 +5757,10 @@ static void bnx2x_init_internal_func(struct bnx2x
>> > *bp)
>> > u32 offset;
>> > u16 max_agg_size;
>> >
>> > - if (is_multi(bp)) {
>> > - tstorm_config.config_flags = MULTI_FLAGS(bp);
>> > + tstorm_config.config_flags = RSS_FLAGS(bp);
>> > +
>> > + if (is_multi(bp))
>> > tstorm_config.rss_result_mask = MULTI_MASK;
>> > - }
>> >
>> > /* Enable TPA if needed */
>> > if (bp->flags & TPA_ENABLE_FLAG)
>>
>>
>> > -----Original Message-----
>> > From: netdev-owner@vger.kernel.org [mailto:netdev-owner@vger.kernel.org] On
>> > Behalf Of Tom Herbert
>> > Sent: Friday, April 23, 2010 8:54 AM
>> > To: davem@davemloft.net; netdev@vger.kernel.org
>> > Subject: [PATCH] bnx2x: add support for receive hashing
>> >
>> > Add support to bnx2x to extract Toeplitz hash out of the receive descriptor
>> > for use in skb->rxhash.
>> >
>> > Signed-off-by: Tom Herbert <therbert@google.com>
>> > ---
>> > diff --git a/drivers/net/bnx2x.h b/drivers/net/bnx2x.h
>> > index 0819530..8bd2368 100644
>> > --- a/drivers/net/bnx2x.h
>> > +++ b/drivers/net/bnx2x.h
>> > @@ -1330,7 +1330,7 @@ static inline u32 reg_poll(struct bnx2x *bp, u32 reg,
>> > u32 expected, int ms,
>> > AEU_INPUTS_ATTN_BITS_MCP_LATCHED_UMP_TX_PARITY | \
>> > AEU_INPUTS_ATTN_BITS_MCP_LATCHED_SCPAD_PARITY)
>> >
>> > -#define MULTI_FLAGS(bp) \
>> > +#define RSS_FLAGS(bp) \
>> > (TSTORM_ETH_FUNCTION_COMMON_CONFIG_RSS_IPV4_CAPABILITY | \
>> > TSTORM_ETH_FUNCTION_COMMON_CONFIG_RSS_IPV4_TCP_CAPABILITY | \
>> > TSTORM_ETH_FUNCTION_COMMON_CONFIG_RSS_IPV6_CAPABILITY | \
>> > diff --git a/drivers/net/bnx2x_main.c b/drivers/net/bnx2x_main.c
>> > index 0c6dba2..613f727 100644
>> > --- a/drivers/net/bnx2x_main.c
>> > +++ b/drivers/net/bnx2x_main.c
>> > @@ -1582,7 +1582,7 @@ static int bnx2x_rx_int(struct bnx2x_fastpath *fp,
>> int
>> > budget)
>> > struct sw_rx_bd *rx_buf = NULL;
>> > struct sk_buff *skb;
>> > union eth_rx_cqe *cqe;
>> > - u8 cqe_fp_flags;
>> > + u8 cqe_fp_flags, cqe_fp_status_flags;
>> > u16 len, pad;
>> >
>> > comp_ring_cons = RCQ_BD(sw_comp_cons);
>> > @@ -1598,6 +1598,7 @@ static int bnx2x_rx_int(struct bnx2x_fastpath *fp,
>> int
>> > budget)
>> >
>> > cqe = &fp->rx_comp_ring[comp_ring_cons];
>> > cqe_fp_flags = cqe->fast_path_cqe.type_error_flags;
>> > + cqe_fp_status_flags = cqe->fast_path_cqe.status_flags;
>> >
>> > DP(NETIF_MSG_RX_STATUS, "CQE type %x err %x status %x"
>> > " queue %x vlan %x len %u\n", CQE_TYPE(cqe_fp_flags),
>> > @@ -1727,6 +1728,12 @@ reuse_rx:
>> >
>> > skb->protocol = eth_type_trans(skb, bp->dev);
>> >
>> > + if ((bp->dev->features & ETH_FLAG_RXHASH) &&
>> > + (cqe_fp_status_flags &
>> > + ETH_FAST_PATH_RX_CQE_RSS_HASH_FLG))
>> > + skb->rxhash = le32_to_cpu(
>> > + cqe->fast_path_cqe.rss_hash_result);
>> > +
>> > skb->ip_summed = CHECKSUM_NONE;
>> > if (bp->rx_csum) {
>> > if (likely(BNX2X_RX_CSUM_OK(cqe)))
>> > @@ -5750,10 +5757,10 @@ static void bnx2x_init_internal_func(struct bnx2x
>> > *bp)
>> > u32 offset;
>> > u16 max_agg_size;
>> >
>> > - if (is_multi(bp)) {
>> > - tstorm_config.config_flags = MULTI_FLAGS(bp);
>> > + tstorm_config.config_flags = RSS_FLAGS(bp);
>> > +
>> > + if (is_multi(bp))
>> > tstorm_config.rss_result_mask = MULTI_MASK;
>> > - }
>> >
>> > /* Enable TPA if needed */
>> > if (bp->flags & TPA_ENABLE_FLAG)
>> > @@ -6629,10 +6636,8 @@ static int bnx2x_init_common(struct bnx2x *bp)
>> > bnx2x_init_block(bp, PBF_BLOCK, COMMON_STAGE);
>> >
>> > REG_WR(bp, SRC_REG_SOFT_RST, 1);
>> > - for (i = SRC_REG_KEYRSS0_0; i <= SRC_REG_KEYRSS1_9; i += 4) {
>> > - REG_WR(bp, i, 0xc0cac01a);
>> > - /* TODO: replace with something meaningful */
>> > - }
>> > + for (i = SRC_REG_KEYRSS0_0; i <= SRC_REG_KEYRSS1_9; i += 4)
>> > + REG_WR(bp, i, random32());
>> > bnx2x_init_block(bp, SRCH_BLOCK, COMMON_STAGE);
>> > #ifdef BCM_CNIC
>> > REG_WR(bp, SRC_REG_KEYSEARCH_0, 0x63285672);
>> > @@ -11001,6 +11006,11 @@ static int bnx2x_set_flags(struct net_device *dev,
>> > u32 data)
>> > changed = 1;
>> > }
>> >
>> > + if (data & ETH_FLAG_RXHASH)
>> > + dev->features |= NETIF_F_RXHASH;
>> > + else
>> > + dev->features &= ~NETIF_F_RXHASH;
>> > +
>> > if (changed && netif_running(dev)) {
>> > bnx2x_nic_unload(bp, UNLOAD_NORMAL);
>> > rc = bnx2x_nic_load(bp, LOAD_NORMAL);
>> > --
>> > To unsubscribe from this list: send the line "unsubscribe netdev" in
>> > the body of a message to majordomo@vger.kernel.org
>> > More majordomo info at http://vger.kernel.org/majordomo-info.html
>>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe netdev" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
>
>
^ permalink raw reply
* Re: [PATCH 1/2] Export firmware assigned labels of network devices to sysfs
From: Domsch, Matt @ 2010-07-07 18:35 UTC (permalink / raw)
To: Greg KH
Cc: K, Narendra, netdev@vger.kernel.org,
linux-hotplug@vger.kernel.org, linux-pci@vger.kernel.org,
Hargrave, Jordan, Rose, Charles, Nijhawan, Vijay
In-Reply-To: <20100707181134.GB4293@kroah.com>
On Wed, Jul 07, 2010 at 01:11:34PM -0500, Greg KH wrote:
> > > Why do you need it? What is calling that function? What am I missing
> > > here?
> >
> > The function 'pci_create_smbiosname_file' below is calling the .test method.
> > For every pdev the function checks if it has a SMBIOS string associated
> > with it or not. If there is no string (and instance) associated, then the
> > attributes 'label' and 'instance' are not created for that pdev.
> > To check for the existance of the string, the .test method is needed and
> > it is not available in 'struct device_attribute'. It provides
> > .show and .store. We need a .show and .test. So we defined
>
> {sigh}
>
> So, you just reinvented the is_visible function in struct
> attribute_group? Please use the infrastructure already available to do
> this, it saves on code and debugging and review time.
I'll take the blame for this. I recommended Narendra use the .test
method, as this is what I did back in 2005 in drivers/firmware/edd.c
which was one of the earliest consumers of the new sysfs code. James
added the is_visible field to attribute groups in 2008, which I
missed (only 3 drivers make use of it, so it was easy to miss). Since
that's the "new" preferred way to do it, we can adjust this patch
accordingly.
Thanks,
Matt
--
Matt Domsch
Technology Strategist
Dell | Office of the CTO
^ permalink raw reply
* RE: [PATCH 2.6.35-rc1] net: vmxnet3 fixes [1/5] Spare skb to avoid starvation
From: Shreyas Bhatewara @ 2010-07-07 18:34 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: netdev, pv-drivers, stolarchuk, linux-kernel
Stephen,
Thanks for taking a look.
Comments inline.
> -----Original Message-----
> From: Stephen Hemminger [mailto:shemminger@vyatta.com]
> Sent: Wednesday, July 07, 2010 10:47 AM
> To: Shreyas Bhatewara
> Cc: netdev@vger.kernel.org; pv-drivers@vmware.com
> Subject: Re: [PATCH 2.6.35-rc1] net: vmxnet3 fixes [1/5] Spare skb to
> avoid starvation
>
> On Wed, 7 Jul 2010 02:21:55 -0700 (PDT)
> Shreyas Bhatewara <sbhatewara@vmware.com> wrote:
>
> >
> >
> > From: Shreyas Bhatewara <sbhatewara@vmware.com>
> >
> > skb_alloc() failure can cause the recv ring to loose all packet
> reception.
> > Avoid this by introducing a spare buffer.
> >
> > Signed-off-by: Michael Stolarchuk <stolarchuk@vmware.com>
> > Signed-off-by: Shreyas Bhatewara <sbhatewara@vmware.com>
> >
>
> I don't see how this fixes the problem, what happens when
> the spare buffer is used up? Better to design a flow control algorithm
> that holds off sender if allocation fails, and retry allocation later
> (for example with a work queue).
>
The spare skb is only used when the ring is "empty" and an skb allocation
failure would cause the ring to starve.
It is never handed up to netif_rx(), and instead, when the rx_interrupt
occurs, the skb is shuffled back to reuse.
A solution like the one below was considered,
vmxnet3_try_alloc_skb(adapter)
{
if(!vmxnet3_rq_alloc_rx_buf(..., ringsize - 1, adapter)) //This
will be scheduled only when ring is empty, so try allocating max
compat_schedule_delayed_work(&alloc_work, delay)
}
probe()
{
....
COMPAT_INIT_DELAYED_WORK(&alloc_work, vmxnet3_try_alloc_skb, adapter);
...
}
vmxnet3_rq_rx_complete()
{
.....
if (unlikely(num_to_alloc > VMXNET3_RX_ALLOC_THRESHOLD(rq,
ring_idx,
adapter))) {
if(!vmxnet3_rq_alloc_rx_buf(rq, ring_idx,
num_to_alloc, adapter) && num_to_alloc == ring_size-1)
compat_schedule_delayed_work(&alloc_work, delay);
....
}
But the value of delay, while scheduling the deferred work is critical. A
good value is the stdard deviation of packet inter-arrival timings (varies
a lot from low to high throughput rate). Instead of maintaining this value
explicitly, its much smarter to rely on packet receive interrupts as
skb-alloc-poll events, and use them to drive the skb allocation.
> > diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c
> b/drivers/net/vmxnet3/vmxnet3_drv.c
> > index 989b742..5a50d10 100644
> ...
>
> > /*
> > @@ -149,6 +149,13 @@ vmxnet3_cmd_ring_desc_avail(struct
> vmxnet3_cmd_ring *ring)
> > ring->next2comp - ring->next2fill - 1;
> > }
> >
> > +static inline bool
> > +vmxnet3_cmd_ring_desc_empty(struct vmxnet3_cmd_ring *ring)
> > +{
> > + return (ring->next2comp == ring->next2fill);
> > +}
>
> const is good practice on functions like this.
Reattaching the patch with const in the function.
---
skb_alloc() failure can cause the ring to loose all packet reception. Avoid
this by introducing a spare buffer. The spare skb is only used when the ring is
"empty" and an skb allocation failure would cause the ring to starve. It is
never handed up to netif_rx(), and instead, when the rx_interrupt occurs, the
skb is shuffled back to reuse. Further use the incoming receive packet
interrupts as events to poll for skb allocation.
Signed-off-by: Michael Stolarchuk <stolarchuk@vmware.com>
Signed-off-by: Shreyas Bhatewara <sbhatewara@vmware.com>
---
diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c b/drivers/net/vmxnet3/vmxnet3_drv.c
index 989b742..5a50d10 100644
--- a/drivers/net/vmxnet3/vmxnet3_drv.c
+++ b/drivers/net/vmxnet3/vmxnet3_drv.c
@@ -541,7 +541,12 @@ vmxnet3_rq_alloc_rx_buf(struct vmxnet3_rx_queue *rq, u32 ring_idx,
NET_IP_ALIGN);
if (unlikely(rbi->skb == NULL)) {
rq->stats.rx_buf_alloc_failure++;
- break;
+ /* starvation prevention */
+ if (vmxnet3_cmd_ring_desc_empty(
+ rq->rx_ring + ring_idx))
+ rbi->skb = rq->spare_skb;
+ else
+ break;
}
rbi->skb->dev = adapter->netdev;
@@ -611,6 +616,29 @@ vmxnet3_append_frag(struct sk_buff *skb, struct Vmxnet3_RxCompDesc *rcd,
}
+/*
+ * Free any pages which were attached to the frags of the spare skb. This can
+ * happen when the spare skb is attached to the rx ring to prevent starvation,
+ * but there was no issue with page allocation.
+ */
+
+static void
+vmxnet3_rx_spare_skb_free_frags(struct vmxnet3_adapter *adapter)
+{
+ struct sk_buff *skb = adapter->rx_queue.spare_skb;
+ int i;
+ for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
+ struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[i];
+ BUG_ON(frag->page != 0);
+ put_page(frag->page);
+ frag->page = 0;
+ frag->size = 0;
+ }
+ skb_shinfo(skb)->nr_frags = 0;
+ skb->data_len = 0;
+}
+
+
static void
vmxnet3_map_pkt(struct sk_buff *skb, struct vmxnet3_tx_ctx *ctx,
struct vmxnet3_tx_queue *tq, struct pci_dev *pdev,
@@ -1060,8 +1088,12 @@ vmxnet3_rx_error(struct vmxnet3_rx_queue *rq, struct Vmxnet3_RxCompDesc *rcd,
* ctx->skb may be NULL if this is the first and the only one
* desc for the pkt
*/
- if (ctx->skb)
- dev_kfree_skb_irq(ctx->skb);
+ if (ctx->skb) {
+ if (ctx->skb == rq->spare_skb)
+ vmxnet3_rx_spare_skb_free_frags(adapter);
+ else
+ dev_kfree_skb_irq(ctx->skb);
+ }
ctx->skb = NULL;
}
@@ -1159,6 +1191,12 @@ vmxnet3_rq_rx_complete(struct vmxnet3_rx_queue *rq,
skb = ctx->skb;
if (rcd->eop) {
+ if (skb == rq->spare_skb) {
+ rq->stats.drop_total++;
+ vmxnet3_rx_spare_skb_free_frags(adapter);
+ ctx->skb = NULL;
+ goto rcd_done;
+ }
skb->len += skb->data_len;
skb->truesize += skb->data_len;
@@ -1244,6 +1282,14 @@ vmxnet3_rq_cleanup(struct vmxnet3_rx_queue *rq,
rq->uncommitted[ring_idx] = 0;
}
+ /* free starvation prevention skb if allocated */
+ if (rq->spare_skb) {
+ vmxnet3_rx_spare_skb_free_frags(adapter);
+ dev_kfree_skb(rq->spare_skb);
+ rq->spare_skb = NULL;
+ }
+
+
rq->comp_ring.gen = VMXNET3_INIT_GEN;
rq->comp_ring.next2proc = 0;
}
@@ -1325,6 +1371,15 @@ vmxnet3_rq_init(struct vmxnet3_rx_queue *rq,
}
vmxnet3_rq_alloc_rx_buf(rq, 1, rq->rx_ring[1].size - 1, adapter);
+ /* allocate ring starvation protection */
+ rq->spare_skb = dev_alloc_skb(PAGE_SIZE);
+ if (rq->spare_skb == NULL) {
+ vmxnet3_rq_cleanup(rq, adapter);
+ return -ENOMEM;
+ }
+
+
+
/* reset the comp ring */
rq->comp_ring.next2proc = 0;
memset(rq->comp_ring.base, 0, rq->comp_ring.size *
diff --git a/drivers/net/vmxnet3/vmxnet3_int.h b/drivers/net/vmxnet3/vmxnet3_int.h
index 34f392f..7985ba4 100644
--- a/drivers/net/vmxnet3/vmxnet3_int.h
+++ b/drivers/net/vmxnet3/vmxnet3_int.h
@@ -68,10 +68,10 @@
/*
* Version numbers
*/
-#define VMXNET3_DRIVER_VERSION_STRING "1.0.5.0-k"
+#define VMXNET3_DRIVER_VERSION_STRING "1.0.6.0-k"
/* a 32-bit int, each byte encode a verion number in VMXNET3_DRIVER_VERSION */
-#define VMXNET3_DRIVER_VERSION_NUM 0x01000500
+#define VMXNET3_DRIVER_VERSION_NUM 0x01000600
/*
@@ -149,6 +149,13 @@ vmxnet3_cmd_ring_desc_avail(struct vmxnet3_cmd_ring *ring)
ring->next2comp - ring->next2fill - 1;
}
+static inline bool
+vmxnet3_cmd_ring_desc_empty(const struct vmxnet3_cmd_ring *ring)
+{
+ return (ring->next2comp == ring->next2fill);
+}
+
+
struct vmxnet3_comp_ring {
union Vmxnet3_GenericDesc *base;
u32 size;
@@ -266,9 +273,10 @@ struct vmxnet3_rx_queue {
u32 qid2; /* rqID in RCD for buffer from 2nd ring */
u32 uncommitted[2]; /* # of buffers allocated since last RXPROD
* update */
- struct vmxnet3_rx_buf_info *buf_info[2];
- struct Vmxnet3_RxQueueCtrl *shared;
+ struct vmxnet3_rx_buf_info *buf_info[2];
+ struct Vmxnet3_RxQueueCtrl *shared;
struct vmxnet3_rq_driver_stats stats;
+ struct sk_buff *spare_skb; /* starvation skb */
} __attribute__((__aligned__(SMP_CACHE_BYTES)));
#define VMXNET3_LINUX_MAX_MSIX_VECT 1
^ permalink raw reply related
* Re: [PATCH 1/2] Export firmware assigned labels of network devices to sysfs
From: Greg KH @ 2010-07-07 18:11 UTC (permalink / raw)
To: Narendra K
Cc: netdev, linux-hotplug, linux-pci, matt_domsch, jordan_hargrave,
charles_rose, vijay_nijhawan
In-Reply-To: <20100707174826.GA1046@auslistsprd01.us.dell.com>
On Wed, Jul 07, 2010 at 12:48:26PM -0500, Narendra K wrote:
> > -----Original Message-----
> > From: Greg KH [mailto:greg@kroah.com]
> > Sent: Wednesday, July 07, 2010 4:52 AM
> > To: K, Narendra
> > Cc: netdev@vger.kernel.org; linux-hotplug@vger.kernel.org;
> > linux-pci@vger.kernel.org; Domsch, Matt; Hargrave, Jordan; Rose,
> > Charles; Nijhawan, Vijay
> > Subject: Re: [PATCH 1/2] Export firmware assigned labels of network
> > devices to sysfs
> >
> > On Tue, Jul 06, 2010 at 01:52:18PM -0500, Narendra K wrote:
> > >
> > > 'device_create_file' takes 'struct device_attribute *' as a param
> > which
> > > we have not used here because 'struct device_attribute' does not have
> > .test
> > > member which we needed in this patch.
> >
> > Why do you need it? What is calling that function? What am I missing
> > here?
>
> The function 'pci_create_smbiosname_file' below is calling the .test method.
> For every pdev the function checks if it has a SMBIOS string associated
> with it or not. If there is no string (and instance) associated, then the
> attributes 'label' and 'instance' are not created for that pdev.
> To check for the existance of the string, the .test method is needed and
> it is not available in 'struct device_attribute'. It provides
> .show and .store. We need a .show and .test. So we defined
{sigh}
So, you just reinvented the is_visible function in struct
attribute_group? Please use the infrastructure already available to do
this, it saves on code and debugging and review time.
thanks,
greg k-h
^ permalink raw reply
* Re: [PATCH RFC 0/4] ARM/KGDB: Some fixes for SMP machines
From: Anton Vorontsov @ 2010-07-07 17:54 UTC (permalink / raw)
To: kgdb-bugreport
Cc: netdev, Russell King, David S. Miller, linux-arm-kernel,
Jason Wessel
In-Reply-To: <20100707171222.GA16448@oksana.dev.rtsoft.ru>
On Wed, Jul 07, 2010 at 09:12:22PM +0400, Anton Vorontsov wrote:
[...]
> 2. The patches are against a heavily patched kernel, and so far
> I didn't rebase them onto the 'debug core' rework as found
> in the very latest mainline kernels. I'll rebase the patches
> soon, so for now this is just an RFC.
BTW, I'm testing with another small fixup applied, I didn't send
it as a patch because this deadlock was already fixed in the
debug_core implementation (which KGDB is using nowadays). But
for the completeness, here it is:
(Don't deadlock if there's a wannabe-master CPUs, which entered
KGDB via exception, not NMI/IPI).
diff --git a/kernel/kgdb.c b/kernel/kgdb.c
index e7a2274..65bf75d 100644
--- a/kernel/kgdb.c
+++ b/kernel/kgdb.c
@@ -1522,7 +1522,9 @@ return_normal:
* from the debugger.
*/
for_each_online_cpu(i) {
- while (atomic_read(&cpu_in_kgdb[i]))
+ while (atomic_read(&cpu_in_kgdb[i]) &&
+ !(kgdb_info[i].exception_state &
+ DCPU_WANT_MASTER))
cpu_relax();
}
}
------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
^ permalink raw reply related
* Re: [PATCH 1/2] Export firmware assigned labels of network devices to sysfs
From: Narendra K @ 2010-07-07 17:48 UTC (permalink / raw)
To: greg
Cc: netdev, linux-hotplug, linux-pci, matt_domsch, jordan_hargrave,
charles_rose, vijay_nijhawan
In-Reply-To: <EDA0A4495861324DA2618B4C45DCB3EE612B27@blrx3m08.blr.amer.dell.com>
> -----Original Message-----
> From: Greg KH [mailto:greg@kroah.com]
> Sent: Wednesday, July 07, 2010 4:52 AM
> To: K, Narendra
> Cc: netdev@vger.kernel.org; linux-hotplug@vger.kernel.org;
> linux-pci@vger.kernel.org; Domsch, Matt; Hargrave, Jordan; Rose,
> Charles; Nijhawan, Vijay
> Subject: Re: [PATCH 1/2] Export firmware assigned labels of network
> devices to sysfs
>
> On Tue, Jul 06, 2010 at 01:52:18PM -0500, Narendra K wrote:
> >
> > 'device_create_file' takes 'struct device_attribute *' as a param
> which
> > we have not used here because 'struct device_attribute' does not have
> .test
> > member which we needed in this patch.
>
> Why do you need it? What is calling that function? What am I missing
> here?
The function 'pci_create_smbiosname_file' below is calling the .test method.
For every pdev the function checks if it has a SMBIOS string associated
with it or not. If there is no string (and instance) associated, then the
attributes 'label' and 'instance' are not created for that pdev.
To check for the existance of the string, the .test method is needed and
it is not available in 'struct device_attribute'. It provides
.show and .store. We need a .show and .test. So we defined
+struct smbios_attribute smbios_attr_label = {
+ .attr = {.name = "label", .mode = 0444, .owner = THIS_MODULE},
+ .show = smbiosname_show,
+ .test = smbios_instance_string_exist,
+};
'smbios_instance_string_exist' checks if the pdev has a 'string' and 'instance'.
+static int
+pci_create_smbiosname_file(struct pci_dev *pdev)
+{
+ if (smbios_attr_label.test && smbios_attr_label.test(&pdev->dev, NULL, NULL)) {
+ if (sysfs_create_file(&pdev->dev.kobj, &smbios_attr_label.attr))
+ return -1;
+ if (sysfs_create_file(&pdev->dev.kobj, &smbios_attr_instance.attr))
+ return -1;
+ return 0;
+ }
+ return -1;
+}
+int pci_create_firmware_label_files(struct pci_dev *pdev)
+{
+ if (!pci_create_smbiosname_file(pdev))
>
> Please always run your patches through scripts/checkpatch.pl and fix up
> the issues it finds before sending it out and having everyone else point
> them out to you :)
>
> Also, a new thread is nice at times for new versions of patches...
Thanks for the feedback. Sorry for missing this.I would address all the issues
and post the patch in a new thread.
With regards,
Narendra K
^ permalink raw reply
* Re: [PATCH 2.6.35-rc1] net: vmxnet3 fixes [1/5] Spare skb to avoid starvation
From: Stephen Hemminger @ 2010-07-07 17:46 UTC (permalink / raw)
To: Shreyas Bhatewara; +Cc: netdev, pv-drivers
In-Reply-To: <alpine.LRH.2.00.1007070214400.6105@localhost.localdomain>
On Wed, 7 Jul 2010 02:21:55 -0700 (PDT)
Shreyas Bhatewara <sbhatewara@vmware.com> wrote:
>
>
> From: Shreyas Bhatewara <sbhatewara@vmware.com>
>
> skb_alloc() failure can cause the recv ring to loose all packet reception.
> Avoid this by introducing a spare buffer.
>
> Signed-off-by: Michael Stolarchuk <stolarchuk@vmware.com>
> Signed-off-by: Shreyas Bhatewara <sbhatewara@vmware.com>
>
I don't see how this fixes the problem, what happens when
the spare buffer is used up? Better to design a flow control algorithm
that holds off sender if allocation fails, and retry allocation later
(for example with a work queue).
> diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c b/drivers/net/vmxnet3/vmxnet3_drv.c
> index 989b742..5a50d10 100644
...
> /*
> @@ -149,6 +149,13 @@ vmxnet3_cmd_ring_desc_avail(struct vmxnet3_cmd_ring *ring)
> ring->next2comp - ring->next2fill - 1;
> }
>
> +static inline bool
> +vmxnet3_cmd_ring_desc_empty(struct vmxnet3_cmd_ring *ring)
> +{
> + return (ring->next2comp == ring->next2fill);
> +}
const is good practice on functions like this.
^ permalink raw reply
* Re: [PATCH] vlan: allow TSO setting on vlan interfaces
From: Ben Hutchings @ 2010-07-07 17:25 UTC (permalink / raw)
To: Eric Dumazet; +Cc: David Miller, netdev, Patrick McHardy
In-Reply-To: <1278522874.5906.9.camel@edumazet-laptop>
On Wed, 2010-07-07 at 19:14 +0200, Eric Dumazet wrote:
> When we need to shape traffic with low speeds, we need to disable tso on
> network interface :
>
> ethtool -K eth0.2240 tso off
>
> It seems vlan interfaces miss the set_tso() ethtool method.
> Propagating tso changes from lower device is not always wanted, some
> vlan want TSO on, other want TSO off.
But it should not be possible to enable TSO if the underlying device
doesn't support it.
Ben.
> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
> ---
>
> diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
> index c6456cb..d827c2e 100644
> --- a/net/8021q/vlan_dev.c
> +++ b/net/8021q/vlan_dev.c
> @@ -844,6 +844,7 @@ static const struct ethtool_ops vlan_ethtool_ops = {
> .get_link = ethtool_op_get_link,
> .get_rx_csum = vlan_ethtool_get_rx_csum,
> .get_flags = vlan_ethtool_get_flags,
> + .set_tso = ethtool_op_set_tso,
> };
>
> static const struct net_device_ops vlan_netdev_ops = {
>
--
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply
* [PATCH] vlan: allow TSO setting on vlan interfaces
From: Eric Dumazet @ 2010-07-07 17:14 UTC (permalink / raw)
To: David Miller; +Cc: netdev, Patrick McHardy
When we need to shape traffic with low speeds, we need to disable tso on
network interface :
ethtool -K eth0.2240 tso off
It seems vlan interfaces miss the set_tso() ethtool method.
Propagating tso changes from lower device is not always wanted, some
vlan want TSO on, other want TSO off.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
index c6456cb..d827c2e 100644
--- a/net/8021q/vlan_dev.c
+++ b/net/8021q/vlan_dev.c
@@ -844,6 +844,7 @@ static const struct ethtool_ops vlan_ethtool_ops = {
.get_link = ethtool_op_get_link,
.get_rx_csum = vlan_ethtool_get_rx_csum,
.get_flags = vlan_ethtool_get_flags,
+ .set_tso = ethtool_op_set_tso,
};
static const struct net_device_ops vlan_netdev_ops = {
^ permalink raw reply related
* [PATCH RFC 2/4] ARM: kgdb: Disable preemption before re-enabling interrupts
From: Anton Vorontsov @ 2010-07-07 17:13 UTC (permalink / raw)
To: kgdb-bugreport
Cc: Russell King, David S. Miller, Jason Wessel, netdev,
linux-arm-kernel
In-Reply-To: <20100707171222.GA16448@oksana.dev.rtsoft.ru>
We have to disable preemption before enabling local IRQs because
local_irq_enable() makes it possible for the kernel to reschedule,
so the kernel might hit a breakpoint causing itself to re-enter
KGDB and die.
Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
---
arch/arm/kernel/kgdb.c | 8 ++++++++
1 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/arch/arm/kernel/kgdb.c b/arch/arm/kernel/kgdb.c
index 5c61100..6ece654 100644
--- a/arch/arm/kernel/kgdb.c
+++ b/arch/arm/kernel/kgdb.c
@@ -167,9 +167,17 @@ static void kgdb_call_nmi_hook(void *ignored)
void kgdb_roundup_cpus(unsigned long flags)
{
+ /*
+ * We have to disable preemption before enabling local
+ * IRQs because local_irq_enable() makes it possible for
+ * the kernel to reschedule, so the kernel might hit a
+ * breakpoint causing itself to re-enter KGDB and die.
+ */
+ preempt_disable();
local_irq_enable();
smp_call_function(kgdb_call_nmi_hook, NULL, 0);
local_irq_disable();
+ preempt_enable();
}
#ifdef CONFIG_SMP
--
1.7.0.5
^ permalink raw reply related
* [PATCH RFC 4/4] kgdb: Quiesce IO back-end before rounding up secondary CPUs
From: Anton Vorontsov @ 2010-07-07 17:13 UTC (permalink / raw)
To: kgdb-bugreport
Cc: Russell King, David S. Miller, Jason Wessel, netdev,
linux-arm-kernel
In-Reply-To: <20100707171222.GA16448@oksana.dev.rtsoft.ru>
KGDB may try to roundup secondary CPUs while they're in the IO
back-end code (e.g. NAPI is polling an eth device), so the
secondary CPU might stop processing IO at a random place, which
may cause the IO back-end to become unusable for KGDB itself.
This patch implements try_quiesce and activate KGDB IO callbacks,
so that we quiesce the IO back-end before rounding up the CPUs.
So far it's only implemented for KGDBoE driver.
Note that we have to quiesce the devices via _try mechanism since
we have to poll for IPIs during the time we wait for the other
CPUs.
Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
---
drivers/net/kgdboe.c | 21 +++++++++++++++++++++
include/linux/kgdb.h | 2 ++
kernel/kgdb.c | 6 ++++++
3 files changed, 29 insertions(+), 0 deletions(-)
diff --git a/drivers/net/kgdboe.c b/drivers/net/kgdboe.c
index 939797a..b46cc9a 100644
--- a/drivers/net/kgdboe.c
+++ b/drivers/net/kgdboe.c
@@ -270,8 +270,29 @@ static int param_set_kgdboe_var(const char *kmessage, struct kernel_param *kp)
return 0;
}
+static int eth_try_quiesce(void)
+{
+ struct napi_struct *napi;
+
+ list_for_each_entry(napi, &np.dev->napi_list, dev_list) {
+ if (!napi_try_disable(napi))
+ return 0;
+ }
+ return 1;
+}
+
+static void eth_activate(void)
+{
+ struct napi_struct *napi;
+
+ list_for_each_entry(napi, &np.dev->napi_list, dev_list)
+ napi_enable(napi);
+}
+
static struct kgdb_io local_kgdb_io_ops = {
.name = "kgdboe",
+ .try_quiesce = eth_try_quiesce,
+ .activate = eth_activate,
.read_char = eth_get_char,
.write_char = eth_put_char,
.flush = eth_flush_buf,
diff --git a/include/linux/kgdb.h b/include/linux/kgdb.h
index 4c80859..93cd305 100644
--- a/include/linux/kgdb.h
+++ b/include/linux/kgdb.h
@@ -250,6 +250,8 @@ struct kgdb_arch {
*/
struct kgdb_io {
const char *name;
+ int (*try_quiesce) (void);
+ void (*activate) (void);
int (*read_char) (void);
void (*write_char) (u8);
void (*flush) (void);
diff --git a/kernel/kgdb.c b/kernel/kgdb.c
index 97edb05..e7a2274 100644
--- a/kernel/kgdb.c
+++ b/kernel/kgdb.c
@@ -1475,6 +1475,9 @@ return_normal:
atomic_inc(&passive_cpu_wait[i]);
}
+ while (kgdb_io_ops->try_quiesce && !kgdb_io_ops->try_quiesce())
+ kgdb_arch_poll_ipi(regs);
+
#ifdef CONFIG_SMP
/* Signal the other CPUs to enter kgdb_wait() */
if ((!kgdb_single_step) && kgdb_do_roundup)
@@ -1489,6 +1492,9 @@ return_normal:
kgdb_arch_poll_ipi(regs);
}
+ if (kgdb_io_ops->activate)
+ kgdb_io_ops->activate();
+
/*
* At this point the primary processor is completely
* in the debugger and all secondary CPUs are quiescent
--
1.7.0.5
^ permalink raw reply related
* [PATCH RFC 3/4] net: Implement napi_try_disable()
From: Anton Vorontsov @ 2010-07-07 17:13 UTC (permalink / raw)
To: kgdb-bugreport
Cc: Russell King, David S. Miller, Jason Wessel, netdev,
linux-arm-kernel
In-Reply-To: <20100707171222.GA16448@oksana.dev.rtsoft.ru>
For KGDBoE we'll need to disable NAPI, so that we don't randomly
interrupt secondary CPUs while they process networking stuff.
We need to wait for NAPI to become disabled from an atomic context,
thus we can't use napi_disable() as it calls msleep().
Plus, in KGDB we have to always poll for IPIs during busy-waiting,
which means that we must implement the function in a such way that
we can do some caller specific work during the time we wait for
NAPI disable, so just changing msleep() to mdelay() won't work.
Signed-off-by: Anton Vorontsov <avorontsov@mvista.com>
---
include/linux/netdevice.h | 21 +++++++++++++++++++++
1 files changed, 21 insertions(+), 0 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 08a1dd8..4fc24eb 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -414,6 +414,27 @@ extern void __napi_complete(struct napi_struct *n);
extern void napi_complete(struct napi_struct *n);
/**
+ * napi_try_disable - try to prevent NAPI from scheduling
+ * @n: napi context
+ *
+ * This call is similar to napi_disable(), except that it doesn't
+ * wait till any outstanding processing completes, but returns
+ * whether napi was successfuly disabled.
+ *
+ * Note that you still must wait till napi is actually disabled,
+ * so the only benefit from using this call is that you can do
+ * some useful work while you wait.
+ */
+static inline int napi_try_disable(struct napi_struct *n)
+{
+ set_bit(NAPI_STATE_DISABLE, &n->state);
+ if (test_and_set_bit(NAPI_STATE_SCHED, &n->state))
+ return 0;
+ clear_bit(NAPI_STATE_DISABLE, &n->state);
+ return 1;
+}
+
+/**
* napi_disable - prevent NAPI from scheduling
* @n: napi context
*
--
1.7.0.5
^ 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