- 08 Mar, 2005 40 commits
-
-
Dave Hansen authored
Refactor the i386 default and CONFIG_DISCONTIG_MEM setup_memory() functions to share the common bootmem initialisation code. This code is intended to be identical, but there are currently some fixes applied to one and not the other. This patch extracts this common initialisation code. Signed-off-by:
Andy Whitcroft <apw@shadowen.org> Signed-off-by:
Dave Hansen <haveblue@us.ibm.com> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
Keith Mannthey authored
Dave Hanson mentioned I should send this patch to you. I posted it to linux-mm a while ago without complaint. This patch solves a simple problem related to the interpretation on the SRAT table and the info found in the e820. When a possible hot-add area is exposed on my box (IBM x445 with hot add enabled in the bios) my SRAT table correctly exposed a new node from the end of my physical memory to 64gb. In the present kernels the numa KVA areas (based on the SRAT) are calculated before find_max_pfn. The remap area is created for this large non-populated zone and the system dies a while later during bootup. I believe the correct things to do (as did the hot-plug community) the correct thing to do is the keep the node_start_end_pfn data structures focuses on memory that is in the system. That is all this patch does. It ignores any node data (correctly reported by the SRAT) that is above the e820 end of memory. Signed-off-by: <kmannth@us.ibm.com> Signed-off-by:
-
Paul Mundt authored
More random cleanup and build fixes. Signed-off-by:
Paul Mundt <lethal@linux-sh.org> Signed-off-by:
-
Paul Mundt authored
This updates the sh cpufreq driver for the cpumask changes. Signed-off-by:
-
Paul Mundt authored
hitfb and pvr2fb were both recently broken by mainline changes, this gets them working properly again. Signed-off-by:
-
Paul Mundt authored
This gets hp620 working again.. Signed-off-by:
Kristoffer Ericson <kristoffer@jlime.com> Signed-off-by:
-
Paul Mundt authored
Convert sh to use generic hardirqs. Signed-off-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
-
Paul Mundt authored
Nothing to see here, move along. Signed-off-by:
-
Paul Mundt authored
Mostly random bugfixes and some build fixes, as well as killing off some leftover cruft. Signed-off-by:
-
Paul Mundt authored
Somewhere along the line dcache disabling decided it wanted to stop itself from compiling. So we fix it. Signed-off-by:
-
Paul Mundt authored
There were a couple of cli()/sti() users left, so we get rid of them.. Hopefully this is the last of this mess. Signed-off-by:
James Nelson <james4765@cwazy.co.uk> Signed-off-by:
-
Paul Mundt authored
TMU0 initialization was broken when the timer was already started by someone else (for instance, a boot loader). This fixes it. Signed-off-by:
Andy Sturges <andy.sturges@st.com> Signed-off-by:
Richard Curnow <rc@rc0.org.uk> Signed-off-by:
-
Paul Mundt authored
This fixes up a few minor IDE issues on sh64. We also enable cayman on-board IDE in the SuperIO. Signed-off-by:
-
Paul Mundt authored
Convert sh64 to use generic hardirqs. Signed-off-by:
-
Paul Mundt authored
This adds support for modules. Signed-off-by:
-
Paul Mundt authored
This adds support for the iomap interface to sh64. As a result of this, we can also clean up a lot of the sh64 common I/O routines. We also add a board-specific ioport_map() for the cayman so we can use iomap generically. Signed-off-by:
-
Paul Mundt authored
Update for the cayman defconfig. We also drop the generic arch/sh64/defconfig as it has very little hope of being kept up to date, so we use the cayman one instead. Signed-off-by:
-
Paul Mundt authored
The BYTES_PER_WORD assumption doesn't work out on sh64 when we are using a 32-bit ABI. We want slab caches to be forced to a minimum alignment of 8-bytes, as it was before Anton's change. This was also already discussed at length with Manfred in this thread: http://marc.theaimsgroup.com/?l=linux-kernel&m=110227138116749&w=2Signed-off-by:
-
Paul Mundt authored
Fairly self explanatory.. Signed-off-by:
-
Paul Mundt authored
This provides a port of checkstack for sh64 for the simple frames allocated as an immediate with a single instruction. Stack frame creation on sh64 happens in a couple of different ways, when the frame size is less than 511 bytes an addi or addi.l is typically used, generally along the lines of something like: addi{,.l} r15, -IMM_FRAME_SIZE, r15 For larger frames, this ends up getting split up into a movi/sub pair: movi IMM_FRAME_SIZE, rX sub r15, rX, r15 We currently don't handle the split pair case, as basically any register can be used, and there is no easy way to determine what happens without scanning the prologue multiple times and using some sort of register cache (we already do something similar for the sh64 stack unwinder, but it would be preferable not to do this in perl..). This does have limited usefulness in that we are not easily able to check for huge frames without manual inspection, but this is still useful enough in the general case to be worth doing for the addi/addi.l case as long as people are aware of this caveat. It may be worth revisiting at a later point to try and catch the larger users though. Signed-off-by: -
Yoichi Yuasa authored
This patch adds spare timer initialization for NEC VR41xx. Signed-off-by:
Yoichi Yuasa <yuasa@hh.iij4u.or.jp> Signed-off-by:
-
Paul Mackerras authored
This patch is from John Rose <johnrose@austin.ibm.com>. EEH scans the system I/O adapters at boot for EEH-capabilities. If no EEH-capable adapters are found, the subsystem is marked disabled for the life of the system. EEH should allow dynamic enabling of the EEH subsystem when hotplug-adding an adapter. Signed-off-by:
John Rose <johnrose@austin.ibm.com> Signed-off-by:
Paul Mackerras <paulus@samba.org> Signed-off-by:
-
Paul Mackerras authored
This patch is from John Rose <johnrose@austin.ibm.com>. Upon DLPAR addition of a PCI Host Brige to a system with purely virtual I/O, set pci_io_base as necessary. Signed-off-by:
-
Paul Mackerras authored
This patch is from Ananth N Mavinakayanahalli <ananth@in.ibm.com>. While processing a kprobe, we were currently not handling all available trap variants available on PowerPC. This lead to the breakage of BUG() handling in ppc64. Signed-off-by:
-
Paul Mackerras authored
This patch is from Thomas Gleixner <tglx@linutronix.de>. Convert the initializers of hw_interrupt_type structures to C99 initializers. Signed-off-by:
Thomas Gleixner <tglx@linutronix.de> Signed-off-by:
-
Paul Mackerras authored
This patch is from Amos Waterland <apw@us.ibm.com>. In pmac_setup.c, the function init_boot_display as currently written only makes sense with CONFIG_BOOTX_TEXT enabled, and causes a link error if it is not enabled. Signed-off-by:
Amos Waterland <apw@us.ibm.com> Signed-off-by:
-
Paul Mackerras authored
This patch is from Jake Moilanen <moilanen@austin.ibm.com>, reformatted by me. The `assigned-addresses' property in the Open Firmware device tree nodes for PCI devices has 64 bits of PCI bus address, but we were only using 32. This patch fixes it so we use all 64. Signed-off-by:
Jake Moilanen <moilanen@austin.ibm.com> Signed-off-by:
-
Paul Mackerras authored
This patch is from Brian King <brking@us.ibm.com>. When working with a PCI-X Mode 2 adapter on a PCI-X Mode 1 PPC64 system, the current code used to determine the config space size of a device results in a PCI Master abort and an EEH error, resulting in the device being taken offline. This patch checks OF to see if the PCI bridge supports PCI-X Mode 2 and fails config accesses beyond 256 bytes if it does not. Signed-off-by:
Brian King <brking@us.ibm.com> Signed-off-by:
-
Colin Leroy authored
This patch lets therm_adt746x handle the latest powerbooks. In these ones, Apple doesn't put the i2c bus number in the "reg" property of the fan node. Instead, we can get the bus number from the fan node path, which looks like "/proc/device-tree/.../i2c-bus@1/.../fan". Here's a patch that handles both old and new form. Signed-off-by:
Colin Leroy <colin@colino.net> Signed-off-by:
-
Mark A. Greer authored
Radstone PPC7D are ppc7447A VME boards with Marvell Discovery-II, dual GigE, dual PMC, 6 serial ports, keyboard/mouse, USB and optional SCSI/VGA. This patch adds support for the PPC7D platform. Signed-off-by:
James Chapman <jchapman@katalix.com> Signed-off-by:
Mark A. Greer <mgreer@mvista.com> Signed-off-by:
-
Jeff Mahoney authored
This patch changes reiserfs to use the VFS level private inode flags, and eliminates the old reiserfs private inode flag. Signed-off-by:
Jeff Mahoney <jeffm@suse.com> Signed-off-by:
-
Jeff Mahoney authored
This patch moves the assignment of i_priv_object to a static inline. This is in preparation for selinux support in reiserfs. Signed-off-by:
-
Jeff Mahoney authored
This patch applies the IS_PRIVATE test to the selinux internal inode loop. Signed-off-by:
-
Jeff Mahoney authored
This patch series adds SELinux support to reiserfs. This patch adds an S_PRIVATE flag to inode->i_flags to mark an inode as filesystem-internal. As such, it should be excepted from the security infrastructure to allow the filesystem to perform its own access control. Signed-off-by:
Stephen Smalley <sds@tycho.nsa.gov> Acked-by:
Chris Wright <chrisw@osdl.org> Signed-off-by:
-
Nigel Cunningham authored
Here's a patch I've prepared which improves the speed at which memory is freed prior to suspend. It should be a big gain for swsusp. For suspend2, it isn't used much, but has shown big improvements when I set a very low image size limit and had memory quite full. 1GB P4, 2.6.11+Suspend2 2.1.8. Soft image size limit set to 2MB to emulate Pavel's implementation (eat as much memory as we can). Without patch: Freed 16545 pages in 4000 jiffies = 16.16 MB/s Freed 83281 pages in 14060 jiffies = 23.14 MB/s Freed 237754 pages in 41482 jiffies = 22.39 MB/s With patch: Freed 52257 pages in 6700 jiffies = 30.46 MB/s Freed 105693 pages in 11035 jiffies = 37.41 MB/s Freed 239007 pages in 18284 jiffies = 51.06 MB/s With a less aggressive image size limit (200MB): Without the patch: Freed 14600 pages in 1749 jiffies = 32.61 MB/s (Anomolous!) Freed 88563 pages in 14719 jiffies = 23.50 MB/s Freed 205734 pages in 32389 jiffies = 24.81 MB/s With the patch: Freed 68252 pages in 496 jiffies = 537.52 MB/s Freed 116464 pages in 569 jiffies = 798.54 MB/s Freed 209699 pages in 705 jiffies = 1161.89 MB/s The later pages take more work to get, which accounts for the slower MB/s with smaller numbers of pages to free. Without the patch, though, getting the easier pages also takes longer because we do a far greater number of invocations of shrink_all_memory in order to get the same number of pages. Signed-Off-By:
Nigel Cunningham <ncunningham@cyclades.com> Acked-By:
Pavel Machek <pavel@ucw.cz> Signed-off-by:
-
Christoph Hellwig authored
This way we actually shake dentries before inodes and thus mark more inodes reclaimable once we shake them. Signed-off-by:
-
Hugh Dickins authored
A small change to the tests for "Bad page state", to avoid one class of the page_remove_rmap BUG reports, giving more information while letting the system continue: check page_mapcount (_mapcount != -1) rather than page_mapped (_mapcount >= 0). And how does _mapcount go bad? In the case under study, it looks sure now that an overheating(?) Pentium III sometimes gets confused by a pair of instructions in the no-buddy-bitmap __free_pages_bulk, and clears the PG_private bit from the _mapcount field while buddying around - changing PG_private value changes the bit cleared from _mapcount. Bad page state mapcount:-4096 would have tracked this down much sooner, and will be recognizable if other cpus show the same aberrant reaction to 2.6.11. The page_remove_rmap BUG does need to be replaced by more permissive and informative handling, but I'm not yet ready to to finalize such a patch. Please admit Colin Harrison to the Order of the Iridescent Penguin, for his tireless testing. Signed-off-by:
Hugh Dickins <hugh@veritas.com> Signed-off-by:
-
Marcelo Tosatti authored
Use find_*_page helpers in swap code instead handcoding it. Signed-off-by:
Marcelo Tosatti <marcelo.tosatti@cyclades.com> Acked-by:
-
Oleg Nesterov authored
On top of "[PATCH 2/2] readahead: improve sequential read detection". Signed-off-by:
Oleg Nesterov <oleg@tv-sign.ru> Signed-off-by:
-
Oleg Nesterov authored
1. Current code can't always detect sequential reading, in case when read size is not PAGE_CACHE_SIZE aligned. If application reads the file by 4096+512 chunks, we have: 1st read: first read detected, prev_page = 2. 2nd read: offset == 2, the read is considered random. page_cache_readahead() should treat prev_page == offset as sequential access. In this case it is better to ++offset, because of blockable_page_cache_readahead(offset, size). 2. If application reads 4096 bytes with *ppos == 512, we have to read 2 pages, but req_size == 1 in do_generic_mapping_read(). Usually it's not a problem. But in random read case it results in unnecessary page cache misses. ~$ time dd conv=notrunc if=/tmp/GIG of=/tmp/dummy bs=$((4096+512)) 2.6.11-clean: real=370.35 user=0.16 sys=14.66 2.6.11-patched: real=234.49 user=0.19 sys=12.41 Signed-off-by:
-
