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)
* '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.
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>
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>
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>
It looks like broken BIOSes controlling Rhine chips will remain in use in
significant numbers; such systems fail to come up via PXE after they have
been put into D3 (power-saving) mode.
This patch adds a module option for disabling the call that puts the chip
to sleep.
Signed-off-by: Roger Luethi <rl@hellgate.ch>
Cc: Jeff Garzik <jgarzik@pobox.com>
Cc: Joerg Bashir <brak@archive.org>
Cc: Tim Phipps <tim@phipps-hutton.freeserve.co.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
The poll_enable should be in init_registers before enabling interrupts, not
in tx_timeout. Thanks for spotting it Roger.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Cc: Roger Luethi <rl@hellgate.ch>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Add NAPI support to the via-rhine driver so that it can handle higher
speeds and doesn't get overloaded by interrupts as easily.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Roger Luethi <rl@hellgate.ch>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
When in-kernel net drivers branched from Donald Becker's vanilla driver
set, in the days before BitKeeper and git, a driver changelog was
maintained in the driver source code. These days, the kernel's
changelog is far superior and much more accurate, so the in-driver
changelogs are removed.
Another relic of the Becker/kernel split was version numbering, using
"foo-LKx.y.z" notation, resulting in weird version numbers like
"1.17b-LK1.1.9". These drivers are for older hardware, and see few
changes these days, so the version numbers were all bumped to something
more simple.
Finally, in xircom_tulip_cb specifically, an additional cleanup removes
the always-enabled CARDBUS cpp macro.
Signed-off-by: Jeff Garzik <jeff@garzik.org>
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>
Revert previous patch with subject "change mdelay to msleep and remove
from ISR path". This patch seems to have caused bigger problems than
it solved, and it didn't solve much of a problem to begin with...
Discussion about backing-out this patch can be found here:
http://marc.theaimsgroup.com/?l=linux-netdev&m=114321570402396&w=2
The git commit associated w/ the original patch is:
6ba98d311d
Signed-off-by: John W. Linville <linville@tuxdriver.com
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Fixes Rhine I cards disclosing fragments of previously transmitted frames
in new transmissions.
Before transmission, any socket buffer (skb) shorter than the ethernet
minimum length of 60 bytes was zero-padded. On Rhine I cards the data can
later be copied into an aligned transmission buffer without copying this
padding. This resulted in the transmission of the frame with the extra
bytes beyond the provided content leaking the previous contents of this
buffer on to the network.
Now zero-padding is repeated in the local aligned buffer if one is used.
Following a suggestion from the via-rhine maintainer, no attempt is made
here to avoid the duplicated effort of padding the skb if it is known that
an aligned buffer will definitely be used. This is to make the change
"obviously correct" and allow it to be applied to a stable kernel if
necessary. There is no change to the flow of control and the changes are
only to the Rhine I code path.
The patch has run on an in-service Rhine-I host without incident. Frames
shorter than 60 bytes are now correctly zero-padded when captured on a
separate host. I see no unusual stats reported by ifconfig, and no unusual
log messages.
Signed-off-by: Craig Brind <craigbrind@gmail.com>
Signed-off-by: Roger Luethi <rl@hellgate.ch>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Roger Luethi <rl@hellgate.ch>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Problems with link state detection have been reported several times in the
past months.
Denis Vlasenko did all the work tracking it down. Jeff Garzik suggested the
proper place for the fix.
When using the mii library, the driver needs to check mii->force_media
and set dev->state accordingly.
Signed-off-by: Roger Luethi <rl@hellgate.ch>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Get rid of the mdelay call in rhine_disable_linkmon. The function
is called from the via-rhine versions of mdio_read and mdio_write.
Those functions are indirectly called from rhine_check_media and
rhine_tx_timeout, both of which can be called in interrupt context.
So, create tx_timeout_task and check_media_task as instances of struct
work_struct inside of rhine_private. Then, change rhine_tx_timeout to
invoke schedule_work for tx_timeout_task (i.e. rhine_tx_timeout_task),
moving the work to process context. Also, change rhine_error (invoked
from rhine_interrupt) to invoke schedule_work for check_media_task
(i.e. rhine_check_media_task), which simply calls rhine_check media
in process context. Finally, add a call to flush_scheduled_work in
rhine_close to avoid any resource conflicts with pending work items.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
Add support for ETHTOOL_GPERMADDR to via-rhine.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>
Many drivers use skb->tail unnecessarily.
In these situations, the code roughly looks like:
dev = dev_alloc_skb(...);
[optional] skb_reserve(skb, ...);
... skb->tail ...
But even if the skb_reserve() happens, skb->data equals
skb->tail. So it doesn't make any sense to use anything
other than skb->data in these cases.
Another case was the s2io.c driver directly mucking with
the skb->data and skb->tail pointers. It really just wanted
to do an skb_reserve(), so that's what the code was changed
to do instead.
Another reason I'm making this change as it allows some SKB
cleanups I have planned simpler to merge. In those cleanups,
skb->head, skb->tail, and skb->end pointers are removed, and
replaced with skb->head_room and skb->tail_room integers.
Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: Jeff Garzik <jgarzik@pobox.com>
Now we can change the pci core to always set this pointer, as pci drivers
should use it, not the driver core callback.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
--Boundary-00=_F5lsC5eH1wGW5o9
Content-Type: text/plain;
charset="koi8-r"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Hi Jeff,
In some messages in via-rhine.c there is a leading space
for no apparent reason. This patch removes it.
--
vda
--Boundary-00=_F5lsC5eH1wGW5o9
Content-Type: text/x-diff;
charset="koi8-r";
name="via-rhine.c.diff"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename="via-rhine.c.diff"
Use the DMA_{64,32}BIT_MASK constants from dma-mapping.h when calling
pci_set_dma_mask() or pci_set_consistent_dma_mask()
This patch includes dma-mapping.h explicitly because it caused errors
on some architectures otherwise.
See http://marc.theaimsgroup.com/?t=108001993000001&r=1&w=2 for details
Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch>
Signed-off-by: Domen Puncer <domen@coderock.org>
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
