During an oprofile session of linux-2.6.20 on a dual opteron system, I noticed
an expensive divide was done in tg3_poll().
I am using gcc-4.1.1, so the following comment from drivers/net/tg3.c seems
over-optimistic :
/* Do not place this n-ring entries value into the tp struct itself,
* we really want to expose these constants to GCC so that modulo et
* al. operations are done with shifts and masks instead of with
* hw multiply/modulo instructions. Another solution would be to
* replace things like '% foo' with '& (foo - 1)'.
*/
#define TG3_RX_RCB_RING_SIZE(tp) \
((tp->tg3_flags2 & TG3_FLG2_5705_PLUS) ? 512 : 1024)
Assembly code before patch :
(oprofile results included)
6434 0.0088 :ffffffff803684b9: mov 0x6f0(%r15),%eax
587 8.0e-04 :ffffffff803684c0: and $0x40000,%eax
2170 0.0030 :ffffffff803684c5: cmp $0x1,%eax
:ffffffff803684c8: lea 0x1(%r13),%eax
:ffffffff803684cc: sbb %ecx,%ecx
2051 0.0028 :ffffffff803684ce: xor %edx,%edx
:ffffffff803684d0: and $0x200,%ecx
20 2.7e-05 :ffffffff803684d6: add $0x200,%ecx
1986 0.0027 :ffffffff803684dc: div %ecx
103427 0.1410 :ffffffff803684de: cmp %edx,0xffffffffffffff7c(%rbp)
Assembly code after the suggested patch :
ffffffff803684b9: mov 0x6f0(%r15),%eax
ffffffff803684c0: and $0x40000,%eax
ffffffff803684c5: cmp $0x1,%eax
ffffffff803684c8: sbb %eax,%eax
ffffffff803684ca: inc %r13d
ffffffff803684cd: and $0x200,%eax
ffffffff803684d2: add $0x1ff,%eax
ffffffff803684d7: and %eax,%r13d
ffffffff803684da: cmp %r13d,0xffffffffffffff7c(%rbp)
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Acked-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove the NETIF_F_TSO #ifdef-ery in drivers/net; this was
for old-old-2.4 compat (even current 2.4 has NETIF_F_TSO)
but it's time to get rid of it by now.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Some PHY trim values need to be fine-tuned on 5755M to be
IEEE-compliant.
Update version to 3.72.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The 5906 PHY requires a special register bit to power down and up the
PHY.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Hot-plug scripts can call tg3_open() as soon as register_netdev() is
called in tg3_init_one(). We need to call pci_set_drvdata() before
register_netdev(), and netif_carrier_off() needs to be moved to
tg3_open() to avoid race conditions.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tp->link_config.orig_* values must be assigned during
tg3_set_settings() because these values will be used to setup the
link speed during tg3_open(). Without these assignments, the link
speed settings will be all messed by if tg3_set_settings() is called
when the device is down.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace kmalloc+memset with kzalloc
Signed-off-by: Yan Burman <burman.yan@gmail.com>
Acked-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change the message to more clearly identify Serdes devices.
Update version to 3.70.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change some udelay() in some eeprom functions to msleep(). Eeprom
related functions are always called from sleepable context.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use netif_msg_* to turn on or off some messages.
Based on Stephen Hemminger's initial patch.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Honor the advertisement bitmask from ethtool. We used to always
advertise the full capability when autoneg was set to on.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add Tg3_FLG2_IS_NIC flag to unambiguously determine whether the
device is NIC or onboard. Previously, the EEPROM_WRITE_PROT flag was
overloaded to also mean onboard. With the separation, we can
support some devices that are onboard but do not use eeprom write
protect.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Phy loopback on most 10/100 devices need to be run in 1Gbps mode in
GMII mode.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sparse noticed a locking imbalance in tg3_open(). This patch adds an
unlock to one of the error paths, so that tg3_open() always exits
without the lock held.
Signed-off-by: Ira W. Snyder <kernel@irasnyder.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Due to hardware errata, TSO must be disabled if the PCI Express clock
request is enabled on 5906. The chip may hang when transmitting TSO
frames if CLKREQ is enabled.
Update version to 3.69.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Newer 5906 bootcode needs about 7ms to finish resetting so the poll
firmware loop was changed to maximum 20ms.
Signed-off-by: Gary Zambrano <zambrano@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Acked-by: Jeff Garzik <jeff@garzik.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use proper upper limits for the loops and check for all error
conditions.
The problem was noticed by Adrian Bunk.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This fixes a bug reported in:
http://bugzilla.kernel.org/show_bug.cgi?id=7438
tg3_close() turns off the PHY if WoL and ASF are both disabled. On
the next tg3_open(), some devices such as the 5752M will not be
brought up correctly without a PHY reset early in the reset sequence.
The PHY clock is needed for some internal MAC blocks to function
correctly.
This problem is fixed by always resetting the PHY early in
tg3_reset_hw() when it is called from tg3_open() or tg3_resume().
tg3_setup_phy() can then be called later in the sequence without the
reset_phy parameter set to 1, since the PHY reset is already done.
Update version to 3.68.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The minimum tx ring size must be greater than MAX_SKB_FRAGS or 3
times that on some chips with TSO bugs.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes the implementation of the ethtool set ring parameters for the
tg3 transmit ring. The size of tx_pending is taken into account
before doing a netif_wake_queue. This prevents the interface from
locking up when smaller transmit ring sizes are used.
Signed-off-by: Ranjit Manomohan <ranjitm@google.com>
Acked-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Maintain a per-CPU global "struct pt_regs *" variable which can be used instead
of passing regs around manually through all ~1800 interrupt handlers in the
Linux kernel.
The regs pointer is used in few places, but it potentially costs both stack
space and code to pass it around. On the FRV arch, removing the regs parameter
from all the genirq function results in a 20% speed up of the IRQ exit path
(ie: from leaving timer_interrupt() to leaving do_IRQ()).
Where appropriate, an arch may override the generic storage facility and do
something different with the variable. On FRV, for instance, the address is
maintained in GR28 at all times inside the kernel as part of general exception
handling.
Having looked over the code, it appears that the parameter may be handed down
through up to twenty or so layers of functions. Consider a USB character
device attached to a USB hub, attached to a USB controller that posts its
interrupts through a cascaded auxiliary interrupt controller. A character
device driver may want to pass regs to the sysrq handler through the input
layer which adds another few layers of parameter passing.
I've build this code with allyesconfig for x86_64 and i386. I've runtested the
main part of the code on FRV and i386, though I can't test most of the drivers.
I've also done partial conversion for powerpc and MIPS - these at least compile
with minimal configurations.
This will affect all archs. Mostly the changes should be relatively easy.
Take do_IRQ(), store the regs pointer at the beginning, saving the old one:
struct pt_regs *old_regs = set_irq_regs(regs);
And put the old one back at the end:
set_irq_regs(old_regs);
Don't pass regs through to generic_handle_irq() or __do_IRQ().
In timer_interrupt(), this sort of change will be necessary:
- update_process_times(user_mode(regs));
- profile_tick(CPU_PROFILING, regs);
+ update_process_times(user_mode(get_irq_regs()));
+ profile_tick(CPU_PROFILING);
I'd like to move update_process_times()'s use of get_irq_regs() into itself,
except that i386, alone of the archs, uses something other than user_mode().
Some notes on the interrupt handling in the drivers:
(*) input_dev() is now gone entirely. The regs pointer is no longer stored in
the input_dev struct.
(*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does
something different depending on whether it's been supplied with a regs
pointer or not.
(*) Various IRQ handler function pointers have been moved to type
irq_handler_t.
Signed-Off-By: David Howells <dhowells@redhat.com>
(cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
Add support for the new 5709 device. This is a new 10/100 Mbps chip.
The mailbox access and firmware interface are quite different from
all other tg3 chips.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Put the firmware polling logic into a separate function. This makes
the code cleaner.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some PHY related fixes:
1. Fix Serdes WoL.
2. Fix loopback test on 10/100 only devices.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change to a different ASF heartbeat message code to improve
reliability.
There were some reports of unintended resets on real time kernels
where the timer may be slow and cause the heartbeat to be late.
Netpoll will also have the same problem because the timer irq will
be unavailable.
Using the new heartbeat code, the ASF firmware will also check the
ring condition before resetting the chip when the heartbeat is
expiring.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Improve 5704S autoneg logic by using a serdes_counter field to keep
track of the transient states. This eliminates a 200 msec busy
loop in the code. Autoneg will take its course without the driver
busy waiting for it to finish.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6: (217 commits)
net/ieee80211: fix more crypto-related build breakage
[PATCH] Spidernet: add ethtool -S (show statistics)
[NET] GT96100: Delete bitrotting ethernet driver
[PATCH] mv643xx_eth: restrict to 32-bit PPC_MULTIPLATFORM
[PATCH] Cirrus Logic ep93xx ethernet driver
r8169: the MMIO region of the 8167 stands behin BAR#1
e1000, ixgb: Remove pointless wrappers
[PATCH] Remove powerpc specific parts of 3c509 driver
[PATCH] s2io: Switch to pci_get_device
[PATCH] gt96100: move to pci_get_device API
[PATCH] ehea: bugfix for register access functions
[PATCH] e1000 disable device on PCI error
drivers/net/phy/fixed: #if 0 some incomplete code
drivers/net: const-ify ethtool_ops declarations
[PATCH] ethtool: allow const ethtool_ops
[PATCH] sky2: big endian
[PATCH] sky2: fiber support
[PATCH] sky2: tx pause bug fix
drivers/net: Trim trailing whitespace
[PATCH] ehea: IBM eHEA Ethernet Device Driver
...
Manually resolved conflicts in drivers/net/ixgb/ixgb_main.c and
drivers/net/sky2.c related to CHECKSUM_HW/CHECKSUM_PARTIAL changes by
commit 84fa7933a3 that just happened to be
next to unrelated changes in this update.
Convert the pci_device_ids to PCI_DEVICE() macro. Saves 1.5k in the
sourcefile.
Signed-off-by: Henrik Kretzschmar <henne@nachtwindheim.de>
Acked-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace CHECKSUM_HW by CHECKSUM_PARTIAL (for outgoing packets, whose
checksum still needs to be completed) and CHECKSUM_COMPLETE (for
incoming packets, device supplied full checksum).
Patch originally from Herbert Xu, updated by myself for 2.6.18-rc3.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
From: Michal Piotrowski <michal.k.k.piotrowski@gmail.com>
Signed-off-by: Michal Piotrowski <michal.k.k.piotrowski@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Fix a subtle race condition between tg3_start_xmit() and tg3_tx()
discovered by Herbert Xu <herbert@gondor.apana.org.au>:
CPU0 CPU1
tg3_start_xmit()
if (tx_ring_full) {
tx_lock
tg3_tx()
if (!netif_queue_stopped)
netif_stop_queue()
if (!tx_ring_full)
update_tx_ring
netif_wake_queue()
tx_unlock
}
Even though tx_ring is updated before the if statement in tg3_tx() in
program order, it can be re-ordered by the CPU as shown above. This
scenario can cause the tx queue to be stopped forever if tg3_tx() has
just freed up the entire tx_ring. The possibility of this happening
should be very rare though.
The following changes are made:
1. Add memory barrier to fix the above race condition.
2. Eliminate the private tx_lock altogether and rely solely on
netif_tx_lock. This eliminates one spinlock in tg3_start_xmit()
when the ring is full.
3. Because of 2, use netif_tx_lock in tg3_tx() before calling
netif_wake_queue().
4. Change TX_BUFFS_AVAIL to an inline function with a memory barrier.
Herbert and David suggested using the memory barrier instead of
volatile.
5. Check for the full wake queue condition before getting
netif_tx_lock in tg3_tx(). This reduces the number of unnecessary
spinlocks when the tx ring is full in a steady-state condition.
6. Update version to 3.65.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
All caller of netdev_alloc_skb need to assign skb->dev shortly
afterwards. Move it into common code.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Handle dev_alloc_skb() failures when initializing the RX rings.
Without proper handling, the driver will crash when using a partial
ring.
Thanks to Stephane Doyon <sdoyon@max-t.com> for reporting the bug and
providing the initial patch.
Howie Xu <howie@vmware.com> also reported the same issue.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add tg3_restart_hw() to handle failures when re-initializing the
device.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the AMD 8131 bridge to the list of chipsets that reorder writes.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Acked-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Enable ipv6 TSO feature on chips that support it.
Update version to 3.61.
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet
Patrick McHardy