7d12e780e0
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)
|
||
|---|---|---|
| .. | ||
| bus-osm.c | ||
| config-osm.c | ||
| core.h | ||
| debug.c | ||
| device.c | ||
| driver.c | ||
| exec-osm.c | ||
| i2o_block.c | ||
| i2o_block.h | ||
| i2o_config.c | ||
| i2o_lan.h | ||
| i2o_proc.c | ||
| i2o_scsi.c | ||
| iop.c | ||
| Kconfig | ||
| Makefile | ||
| pci.c | ||
| README | ||
| README.ioctl | ||
Linux I2O Support (c) Copyright 1999 Red Hat Software
and others.
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version
2 of the License, or (at your option) any later version.
AUTHORS (so far)
Alan Cox, Building Number Three Ltd.
Core code, SCSI and Block OSMs
Steve Ralston, LSI Logic Corp.
Debugging SCSI and Block OSM
Deepak Saxena, Intel Corp.
Various core/block extensions
/proc interface, bug fixes
Ioctl interfaces for control
Debugging LAN OSM
Philip Rumpf
Fixed assorted dumb SMP locking bugs
Juha Sievanen, University of Helsinki Finland
LAN OSM code
/proc interface to LAN class
Bug fixes
Core code extensions
Auvo Häkkinen, University of Helsinki Finland
LAN OSM code
/Proc interface to LAN class
Bug fixes
Core code extensions
Taneli Vähäkangas, University of Helsinki Finland
Fixes to i2o_config
CREDITS
This work was made possible by
Red Hat Software
Funding for the Building #3 part of the project
Symbios Logic (Now LSI)
Host adapters, hints, known to work platforms when I hit
compatibility problems
BoxHill Corporation
Loan of initial FibreChannel disk array used for development work.
European Comission
Funding the work done by the University of Helsinki
SysKonnect
Loan of FDDI and Gigabit Ethernet cards
ASUSTeK
Loan of I2O motherboard
STATUS:
o The core setup works within limits.
o The scsi layer seems to almost work.
I'm still chasing down the hang bug.
o The block OSM is mostly functional
o LAN OSM works with FDDI and Ethernet cards.
TO DO:
General:
o Provide hidden address space if asked
o Long term message flow control
o PCI IOP's without interrupts are not supported yet
o Push FAIL handling into the core
o DDM control interfaces for module load etc
o Add I2O 2.0 support (Deffered to 2.5 kernel)
Block:
o Multiple major numbers
o Read ahead and cache handling stuff. Talk to Ingo and people
o Power management
o Finish Media changers
SCSI:
o Find the right way to associate drives/luns/busses
Lan:
o Performance tuning
o Test Fibre Channel code
Tape:
o Anyone seen anything implementing this ?
(D.S: Will attempt to do so if spare cycles permit)