Commit graph

31 commits

Author SHA1 Message Date
David Howells
7d12e780e0 IRQ: Maintain regs pointer globally rather than passing to IRQ handlers
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)
2006-10-05 15:10:12 +01:00
Jeff Garzik
7282d491ec drivers/net: const-ify ethtool_ops declarations
Signed-off-by: Jeff Garzik <jeff@garzik.org>
2006-09-13 14:30:00 -04:00
Jeff Garzik
4e83b7fad8 Merge branch 'tmp' into upstream 2006-09-12 11:45:33 -04:00
Andy Gospodarek
d5b20697ca [PATCH] Remove more unnecessary driver printk's
As I promised last week, here is the first pass at removing all
unnecessary printk's that exist in network device drivers currently in
promiscuous mode.  The duplicate messages are not needed so they have
been removed.  Some of these drivers are quite old and might not need an
update, but I did them all anyway.

I am currently auditing the remaining conditional printk's and will send
out a patch for those soon.

Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
2006-09-12 11:45:19 -04:00
Jeff Garzik
299176206b drivers/net: Remove deprecated use of pci_module_init()
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>
2006-08-19 17:48:59 -04:00
Thomas Gleixner
1fb9df5d30 [PATCH] irq-flags: drivers/net: Use the new IRQF_ constants
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-07-02 13:58:51 -07:00
Herbert Xu
5b057c6b1a [NET]: Avoid allocating skb in skb_pad
First of all it is unnecessary to allocate a new skb in skb_pad since
the existing one is not shared.  More importantly, our hard_start_xmit
interface does not allow a new skb to be allocated since that breaks
requeueing.

This patch uses pskb_expand_head to expand the existing skb and linearize
it if needed.  Actually, someone should sift through every instance of
skb_pad on a non-linear skb as they do not fit the reasons why this was
originally created.

Incidentally, this fixes a minor bug when the skb is cloned (tcpdump,
TCP, etc.).  As it is skb_pad will simply write over a cloned skb.  Because
of the position of the write it is unlikely to cause problems but still
it's best if we don't do it.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-23 02:06:41 -07:00
Arjan van de Ven
f71e130966 Massive net driver const-ification. 2006-03-03 21:33:57 -05:00
Francois Romieu
10487fbd74 sis190: early setting of the pci driver private data
Below this point, the error path will proceed through
sis190_release_board(). It will happily oops if
pci_set_drvdata() has not been issued.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
2006-02-16 22:17:00 +01:00
Jesper Juhl
3c6bee1d40 [PATCH] turn "const static" into "static const"
ICC likes to complain about storage class not being first, GCC doesn't
care much (except for cases like "inline static").
have a hard time seeing how it could break anything.

Thanks to Gabriel A. Devenyi for pointing out
http://linuxicc.sourceforge.net/ which is what made me create this patch.

Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 08:01:55 -08:00
Aurelien Jarno
8fee5f51a5 [PATCH] sis190.c: fix multicast MAC filter
Here is a patch that changes the way the MAC filter is computed for the
multicast addresses. The computation is taken from the SiS GPL driver.

This patch is necessary to get IPv6 working.

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-10-28 16:14:11 -04:00
Francois Romieu
e797637f4c [PATCH] sis190: basic sis191 support
The sis191 is the gigabit brother of the sis190. SiS's driver suggests
that the register set is backward compatible: this should hopefully
give a basic driver.

The device should allow the usual features from a modern ethernet
adapter (802.1q, SG, Jumbo frames, TSO, checksum offload). So far
the relevant register layout is not documented. SiS's driver does
not provide these features either (at least not for Linux).

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-05 18:06:53 -04:00
Francois Romieu
c3d6f1f24c [PATCH] sis190: RGMII Tx internal delay fiddling
Don't ask.
The patch is based on SiS's GPLed driver.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-05 18:06:53 -04:00
Francois Romieu
6614a6dc6e [PATCH] sis190: make 10Mbps the default when handling the StationControl register
This patch does three things:
- widen the access to the StationControl register (note the SIS_W16
  versus SIS_W32 change);
- default to 10Mbps half duplex when the LPA can not be evaluated
  (reg31->ctl is identical for both). It can be argued that it makes
  sense as the lowest common denominator when everything else failed.
  Btw it works better than the current code. :o)
- remove some enums: they do not document anymore.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-05 18:06:53 -04:00
Francois Romieu
900eb9d692 [PATCH] sis190: recent chipsets from SiS include a RGMII
Extracted from SiS's GPLed driver. From the few pdf available at SiS's,
it seems that the 965 and the 966 south bridge include this interface
whereas the 965L (and anything below) does not. It is expected to be a
sis191 related feature and should not hurt the existing sis190 driver.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-05 18:06:52 -04:00
Francois Romieu
21461380d3 [PATCH] sis190: unmask the link change events
link changes reporting does not work when the driver masks its irq event

Signed-off-by: Arnaud Patard <apatard@mandriva.com>
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-09-05 18:06:52 -04:00
Francois Romieu
3690b6c124 [PATCH] sis190: complete the mii probe before registering the netdevice
The userspace must not be able to issue ethtool command and manage the
mii before it is completely initialized. Avoid some pesky "eth%d" messages.

Signed-off-by: Arnaud Patard <apatard@mandriva.com>
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-08-26 00:52:11 -04:00
Francois Romieu
8348b4db5f [PATCH] sis190: compare the lpa to the local advertisement
The station control register must depend on both the advertisement and the lpa

The link partner ability has better be intersected with the current
advertised value before it is feed to the station control register.

Sight-catched-by: Lars Vahlenberg <lars.vahlenberg@gmail.com>
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:01 -04:00
Francois Romieu
560d3d521d [PATCH] sis190: PHY identifier for the K8S-MX motherboard.
Added PHY identifier for the Asus K8S-MX motherboard

Note: the same ID appears in the sis900 driver.

Signed-off-by: Lars Vahlenberg <lars.vahlenberg@gmail.com>
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:01 -04:00
Francois Romieu
fcb9821d3d [PATCH] sis190: new PHY detection code.
New PHY detection code.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:00 -04:00
Francois Romieu
fc10c39d79 [PATCH] sis190: dummy read is required by the status register
Add a dummy read before accessing the status register

SiS driver suggests it.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:00 -04:00
Francois Romieu
9ede109bbe [PATCH] sis190: allow a non-hardcoded ID for the PHY.
Allow a non-hardcoded ID for the PHY

This is the first step before the driver probes for the PHY address.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:00 -04:00
Francois Romieu
3cec93c712 [PATCH] sis190: add endian annotations.
Add endian annotations.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:00 -04:00
Francois Romieu
bcad5e5378 [PATCH] sis190: extract bits definition from SiS driver.
extract bits definition from SiS driver

- fix the Rx stats;
- minor pieces of documentation.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:00 -04:00
Francois Romieu
8b5641d4f1 [PATCH] sis190: the size of the Rx buffer is constrained
Add a restriction to the size of the Rx buffer

SiS driver forces the size of any Rx buffer to be a multiple of 64 bit.
I would not be surprized that it goes along with some alignment issues
which have been experienced before. So far it does not make much of a
difference (both drivers use 1536 bytes buffer).

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:00 -04:00
Francois Romieu
830fb7d232 [PATCH] sis190: initialisation of MAC address.
Extract some mac addr code from SiS's driver.

Some magic may hide beyond the isa bridge. The Rx mac control
register is now set without condition.

Note: good or bad, this part of the code is quite close to sis900.c.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:00 -04:00
Francois Romieu
40292fb0f0 [PATCH] sis190: remove hardcoded constants.
Replace hardcoded constants by enumerated values in sis190_read_eeprom

The names of the enumerated values have been extracted from SiS'official
driver (v1.00.00 published on 2005/07/11).

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:00 -04:00
Francois Romieu
188f23ba94 [PATCH] sis190: merge some register related information from SiS driver.
Merge some register related information from SiS driver.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:21:00 -04:00
Francois Romieu
43afb949a9 [PATCH] sis190: ethtool/mii support.
ethtool/mii support

Bug: disabling autonegotiation and setting the link parameters at the
same time does not provide the expected result. More investigation is
needed.

Note: past the initial probe/open time, the link is managed from user-space
or accessed through sis190_phy_task, i.e. in a usermode context. Whence the
very limited locking needs.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:20:59 -04:00
Francois Romieu
4405d3b5ef [PATCH] sis190: netconsole support.
netconsole support.

This stuff should be factored out of every driver.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:20:59 -04:00
Francois Romieu
890e8d0a3d [PATCH] sis190: resurrection
Raise the sis190 driver from the dead

The driver handles the integrated network device found on SiS 965L
chipset. It follows the classical (non-napi) interrupt-driven model
and provides minimal ethtool support.

The code comes from a heavy cleanup/rewrite of the original code
which was removed from the kernel on 14/04/2004. Since the r8169
driver does not work too bad and there will probably be (at least)
a few months of improvements/testing/fixing, I made the code as
close as possible to the r8169 one.

Pascal Chapperon <pascal.chapperon@wanadoo.fr> deserves some special
credit for testing and bug-catching. Many thanks to Lars Vahlenberg
as well.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
2005-07-30 18:20:59 -04:00