- 11 Feb, 2007 40 commits
-
-
Jeff Dike authored
Make a couple of variables static. Signed-off-by:
Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Jeff Dike authored
Locking fixes. Locking was totally lacking for the mconsole_devices, which got a spin lock, and the unplugged pages data, which got a mutex. The locking of the mconsole console output code was confused. Now, the console_lock (renamed to client_lock) protects the clients list. Signed-off-by:
-
Jeff Dike authored
Whitespace and style fixes. Signed-off-by:
-
Jeff Dike authored
Comment the lack of locking and make a couple of variables static. Signed-off-by:
-
Jeff Dike authored
Whitespace and style fixes. Signed-off-by:
-
Jeff Dike authored
Replace BKL use with a spinlock. Also fix the control so that open doesn't return holding a lock. Signed-off-by:
-
Jeff Dike authored
Some whitespace and coding style cleanups in the network driver code. Signed-off-by:
Jeff Garzik <jeff@garzik.org> Signed-off-by:
-
Jeff Dike authored
The registration of host network transports needed some locking. The transport list itself is locked, but calls to the registration routines are not. This is compensated for by checking that a transport structure is not yet on any list. I also took the opportunity to const all fields in the transport structure except the list, which obviously can be modified. Signed-off-by:
-
Jeff Dike authored
Fix (i.e. add some) the locking around the irqs_to_free list. Signed-off-by:
-
Jeff Dike authored
Some comment and whitespace cleanups in the console and mconsole code. Signed-off-by:
-
Jeff Dike authored
I noticed that errors happening while hotplugging devices from the host were never returned back to the mconsole client. In some cases, success was returned instead of even an information-free error. This patch cleans that up by having the low-level configuration code pass back an error string along with an error code. At the top level, which knows whether it is early boot time or responding to an mconsole request, the string is printk'd or returned to the mconsole client. There are also whitespace and trivial code cleanups in the surrounding code. Signed-off-by:
-
Jeff Dike authored
Clean up the console driver locking. There are various problems here, including sleeping under a spinlock and spinlock recursion, some of which are fixed here. This patch deals with the locking involved with opens and closes. The problem is that an mconsole request to change a console's configuration can race with an open. Changing a configuration should only be done when a console isn't opened. Also, an open must be looking at a stable configuration. In addition, a get configuration request must observe the same locking since it must also see a stable configuration. With the old locking, it was possible for this to hang indefinitely in some cases because open would block for a long time waiting for a connection from the host while holding the lock needed by the mconsole request. As explained in the long comment, this is fixed by adding a spinlock for the use count and configuration and a mutex for the actual open and close. Signed-off-by:
-
Jiri Kosina authored
TLB handling for CRIS contains local_irq_disable() after local_save_flags(). Turn this into local_irq_save(). Signed-off-by:
Jiri Kosina <jkosina@suse.cz> Cc: Mikael Starvik <starvik@axis.com> Signed-off-by:
-
Ahmed S. Darwish authored
Use ARRAY_SIZE macro already defined in linux/kernel.h Signed-off-by:
Ahmed S. Darwish <darwish.07@gmail.com> Cc: Mikael Starvik <starvik@axis.com> Signed-off-by:
-
Jiri Kosina authored
Various headers for CRIS architecture contain local_irq_disable() after local_save_flags(). Turn it into local_irq_save(). Signed-off-by:
-
Jiri Kosina authored
arch/cris/arch-v10/kernel/time.c::get_ns_in_jiffie() contains local_irq_disable() call after local_irq_save(). This looks redundant. arch/cris/kernel/time.c::do_gettimeofday() contains local_irq_disable() call after local_irq_save(). This looks redundant. Signed-off-by:
-
Mike Frysinger authored
We don't actually use anything from asm-m68k/page.h in asm-m68k/user.h, so don't bother including it Signed-off-by:
Mike Frysinger <vapier@gentoo.org> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Roman Zippel <zippel@linux-m68k.org> Signed-off-by:
-
Ahmed S. Darwish authored
Use ARRAY_SIZE macro already defined in linux/kernel.h Signed-off-by:
-
Ahmed S. Darwish authored
Use ARRAY_SIZE macro already defined in linux/kernel.h Signed-off-by:
Greg Ungerer <gerg@uclinux.org> Signed-off-by:
-
Mathieu Desnoyers authored
kernel/time/clocksource.c needs struct task_struct on m68k. Because it uses spin_unlock_irq(), which, on m68k, uses hardirq_count(), which uses preempt_count(), which needs to dereference struct task_struct, we have to include sched.h. Because it would cause a loop inclusion, we cannot include sched.h in any other of asm-m68k/system.h, linux/thread_info.h, linux/hardirq.h, which leaves this ugly include in a C file as the only simple solution. Signed-off-by:
Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Cc: Ingo Molnar <mingo@elte.hu> Cc: Roman Zippel <zippel@linux-m68k.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: <stable@kernel.org> Signed-off-by:
-
Al Viro authored
Recent as(1) doesn't think that . terminates a macro name, so getuser.l is _not_ treated as invoking getuser with .l as the first argument. arch/m68k/math-emu relies on old behaviour, so it gets a lot of undefined macros with more or less current binutils. Note that this behaviour remains in all recent versions and is unrelated to another binutils problems we used to have for a while (having (%a0)+ parsed as two arguments). This one is there to stay; it's an intentional and documented change. .irp <identifier> <words> [text] .endr expands to a copy of text per each word, with \<identifier> replaced with corresponding word. Again, what happens depends on whether gas_ident.x is treated as one or as two tokens; in the former case we'll get old_gas incremented once, in the latter - twice. The rest is obvious. Unlike .macro argument list _anything_ is explicitly allowed after .irp <identifier>; here we are on very safe ground. And yes, it does work with all gas variants I've got here (including vanilla 2.15, 2.16, 2.16.1 and 2.17, plus debian and FC binutils). Signed-off-by:
Al Viro <viro@zeniv.linux.org.uk> Cc: Roman Zippel <zippel@linux-m68k.org> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by:
-
Hirokazu Takata authored
Cosmetic updates and trivial fixes of m32r arch-dependent files. - Remove RCS ID strings and trailing white lines - Other misc. cosmetic updates Signed-off-by:
Hirokazu Takata <takata@linux-m32r.org> Signed-off-by:
-
Hirokazu Takata authored
This patch fixes the kernel entry point address of vmlinux. The m32r kernel entry address is 0x08002000 (physical). But, so far, the ENTRY point written in vmlinux.lds.S was not point the correct kernel entry address. (before fix) $ objdump -x vmlinux vmlinux: file format elf32-m32r-linux vmlinux architecture: m32r2, flags 0x00000112: EXEC_P, HAS_SYMS, D_PAGED start address 0x88002090 /* NG */ : Sections: Idx Name Size VMA LMA File off Algn 0 .empty_zero_page 00001000 88001000 88001000 00001000 2**12 CONTENTS, ALLOC, LOAD, DATA 1 .boot 0000008c 88002000 88002000 00002000 2**2 CONTENTS, ALLOC, LOAD, READONLY, CODE 2 .text 001ab694 88002090 88002090 00002090 2**4 CONTENTS, ALLOC, LOAD, READONLY, CODE : (after fix) $ objdump -x vmlinux vmlinux: file format elf32-m32r-linux vmlinux architecture: m32r2, flags 0x00000112: EXEC_P, HAS_SYMS, D_PAGED start address 0x08002000 /* OK */ : This fix also remedies the following GDB error message (of gdb-6.4 or after) at the first operation of kernel debugging: "Previous frame identical to this frame (corrupt stack?)". Signed-off-by: -
Hirokazu Takata authored
This patch upgrades defconfig files for all m32r platforms. Signed-off-by:
-
Hirokazu Takata authored
Fix do_page_fault and update_mmu_cache. * Fix do_page_fault (vmalloc_fault:) to pass error_code correctly to update_mmu_cache by using a thread-fault code for all m32r chips. * Fix update_mmu_cache for OPSP chip - #ifdef CONFIG_CHIP_OPSP portion is a workaround of OPSP; Add a notfound-case operation to update_mmu_cache for OPSP like other m32r chip. - Fix pte_data that was not initialized if no entry found. Signed-off-by:Kazuhiro Inaoka <inaoka@linux-m32r.org> Signed-off-by:
-
Hirokazu Takata authored
Additional fixes for processors without ISA_DSP_LEVEL2. sigcontext_t does not have dummy_acc1h, dummy_acc1l members any longer. Signed-off-by:
-
Rafael J. Wysocki authored
Make the userland interface of swsusp call pm_ops->finish() after enable_nonboot_cpus() and before resume_device(), as indicated by the recent discussion on Linux-PM (cf. http://lists.osdl.org/pipermail/linux-pm/2006-November/004164.html). This patch changes the SNAPSHOT_PMOPS ioctl so that its first function, PMOPS_PREPARE, only sets a switch turning the platform suspend mode on, and its last function, PMOPS_FINISH, only checks if the platform mode is enabled. This should allow the older userland tools to work with new kernels without any modifications. The changes here only affect the userland interface of swsusp. Signed-off-by:
Rafael J. Wysocki <rjw@sisk.pl> Acked-by:
Pavel Machek <pavel@ucw.cz> Cc: Greg KH <greg@kroah.com> Cc: Nigel Cunningham <nigel@suspend2.net> Cc: Patrick Mochel <mochel@digitalimplant.org> Signed-off-by:
-
Andrew Morton authored
The compiler will do that. And if it doesn't, we don't want to either ;) Cc: Rafael J. Wysocki <rjw@sisk.pl> Cc: Pavel Machek <pavel@ucw.cz> Cc: Greg KH <greg@kroah.com> Cc: Nigel Cunningham <nigel@suspend2.net> Cc: Patrick Mochel <mochel@digitalimplant.org> Signed-off-by:
-
Rafael J. Wysocki authored
Change the ordering of code in kernel/power/user.c so that device_suspend() is called before disable_nonboot_cpus() and device_resume() is called after enable_nonboot_cpus(). This is needed to make the userland suspend call pm_ops->finish() after enable_nonboot_cpus() and before device_resume(), as indicated by the recent discussion on Linux-PM (cf. http://lists.osdl.org/pipermail/linux-pm/2006-November/004164.html). The changes here only affect the userland interface of swsusp. Signed-off-by:
-
Rafael J. Wysocki authored
Change the ordering of code in kernel/power/disk.c so that device_suspend() is called before disable_nonboot_cpus() and platform_finish() is called after enable_nonboot_cpus() and before device_resume(), as indicated by the recent discussion on Linux-PM (cf. http://lists.osdl.org/pipermail/linux-pm/2006-November/004164.html). The changes here only affect the built-in swsusp. [alexey.y.starikovskiy@linux.intel.com: fix LED blinking during image load] Signed-off-by:
-
Rafael J. Wysocki authored
As indicated in a recent thread on Linux-PM, it's necessary to call pm_ops->finish() before devce_resume(), but enable_nonboot_cpus() has to be called before pm_ops->finish() (cf. http://lists.osdl.org/pipermail/linux-pm/2006-November/004164.html). For consistency, it seems reasonable to call disable_nonboot_cpus() after device_suspend(). This way the suspend code will remain symmetrical with respect to the resume code and it may allow us to speed up things in the future by suspending and resuming devices and/or saving the suspend image in many threads. The following series of patches reorders the suspend and resume code so that nonboot CPUs are disabled after devices have been suspended and enabled before the devices are resumed. It also causes pm_ops->finish() to be called after enable_nonboot_cpus() wherever necessary. This patch: Change the ordering of code in kernel/power/main.c so that device_suspend() is called before disable_nonboot_cpus() and pm_ops->finish() is called after enable_nonboot_cpus() and before device_resume(), as indicated by recent discussion on Linux-PM (cf. http://lists.osdl.org/pipermail/linux-pm/2006-November/004164.html). Signed-off-by:
-
Ahmed S. Darwish authored
Use ARRAY_SIZE macro already defined in linux/kernel.h Signed-off-by:
Ian Molton <spyro@f2s.com> Signed-off-by:
-
Aneesh Kumar K.V authored
Module loading on Alpha was failing with error "Could not allocate 8 bytes percpu data". Looking at dmesg we have the below error "No per-cpu room for modules." Increase the PERCPU_ENOUGH_ROOM in a similar way as x86_64 Signed-off-by:
Aneesh Kumar K.V <aneesh.kumar@gmail.com> Cc: <Jay.Estabrook@hp.com> Cc: Richard Henderson <rth@twiddle.net> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Signed-off-by:
-
Eric Paris authored
Reading /proc/sys/kernel/cap-bound requires CAP_SYS_MODULE. (see proc_dointvec_bset in kernel/sysctl.c) sysctl appears to drive all over proc reading everything it can get it's hands on and is complaining when it is being denied access to read cap-bound. Clearly writing to cap-bound should be a sensitive operation but requiring CAP_SYS_MODULE to read cap-bound seems a bit to strong. I believe the information could with reasonable certainty be obtained by looking at a bunch of the output of /proc/pid/status which has very low security protection, so at best we are just getting a little obfuscation of information. Currently SELinux policy has to 'dontaudit' capability checks for CAP_SYS_MODULE for things like sysctl which just want to read cap-bound. In doing so we also as a byproduct have to hide warnings of potential exploits such as if at some time that sysctl actually tried to load a module. I wondered if anyone would have a problem opening cap-bound up to read from anyone? Acked-by:
Chris Wright <chrisw@sous-sol.org> Cc: Stephen Smalley <sds@tycho.nsa.gov> Cc: James Morris <jmorris@namei.org> Signed-off-by:
-
Ken Chen authored
When kernel unmaps an address range, it needs to transfer PTE state into page struct. Currently, kernel transfer access bit via mark_page_accessed(). The call to mark_page_accessed in the unmap path doesn't look logically correct. At unmap time, calling mark_page_accessed will causes page LRU state to be bumped up one step closer to more recently used state. It is causing quite a bit headache in a scenario when a process creates a shmem segment, touch a whole bunch of pages, then unmaps it. The unmapping takes a long time because mark_page_accessed() will start moving pages from inactive to active list. I'm not too much concerned with moving the page from one list to another in LRU. Sooner or later it might be moved because of multiple mappings from various processes. But it just doesn't look logical that when user asks a range to be unmapped, it's his intention that the process is no longer interested in these pages. Moving those pages to active list (or bumping up a state towards more active) seems to be an over reaction. It also prolongs unmapping latency which is the core issue I'm trying to solve. As suggested by Peter, we should still preserve the info on pte young pages, but not more. Signed-off-by:
Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by:
Ken Chen <kenchen@google.com> Cc: Hugh Dickins <hugh@veritas.com> Signed-off-by:
-
Ken Chen authored
As pointed out by Hugh, ramfs would also benefit from using the new set_page_dirty aop method for memory backed file systems. Signed-off-by:
-
Ken Chen authored
shmem backed file does not have page writeback, nor it participates in backing device's dirty or writeback accounting. So using generic __set_page_dirty_nobuffers() for its .set_page_dirty aops method is a bit overkill. It unnecessarily prolongs shm unmap latency. For example, on a densely populated large shm segment (sevearl GBs), the unmapping operation becomes painfully long. Because at unmap, kernel transfers dirty bit in PTE into page struct and to the radix tree tag. The operation of tagging the radix tree is particularly expensive because it has to traverse the tree from the root to the leaf node on every dirty page. What's bothering is that radix tree tag is used for page write back. However, shmem is memory backed and there is no page write back for such file system. And in the end, we spend all that time tagging radix tree and none of that fancy tagging will be used. So let's simplify it by introduce a new aops __set_page_dirty_no_writeback and this will speed up shm unmap. Signed-off-by:
-
Andy Whitcroft authored
Currently if we have a non-zero ZONES_SHIFT we assume we are able to rely on that as the bottom edge of the ZONEID, if not then we use the NODES_PGOFF as the right end of either NODES _or_ SECTION. This latter is more luck than judgement and would be incorrect if we reordered the SECTION,NODE,ZONE options in the fields space. Really what we want is the lower of the right hand end of the two fields we are using (either NODE,ZONE or SECTION,ZONE). Codify that explicitly. As always allow for there being no bits in either of the fields, such as might be valid in a non-numa machine with only a zone NORMAL. I have checked that the compiler is still able to constant fold all of this away correctly. Signed-off-by:
Andy Whitcroft <apw@shadowen.org> Acked-by:
Christoph Lameter <clameter@sgi.com> Signed-off-by:
-
Christoph Lameter authored
As Andi pointed out: CONFIG_GENERIC_ISA_DMA only disables the ISA DMA channel management. Other functionality may still expect GFP_DMA to provide memory below 16M. So we need to make sure that CONFIG_ZONE_DMA is set independent of CONFIG_GENERIC_ISA_DMA. Undo the modifications to mm/Kconfig where we made ZONE_DMA dependent on GENERIC_ISA_DMA and set theses explicitly in each arches Kconfig. Reviews must occur for each arch in order to determine if ZONE_DMA can be switched off. It can only be switched off if we know that all devices supported by a platform are capable of performing DMA transfers to all of memory (Some arches already support this: uml, avr32, sh sh64, parisc and IA64/Altix). In order to switch ZONE_DMA off conditionally, one would have to establish a scheme by which one can assure that no drivers are enabled that are only capable of doing I/O to a part of memory, or one needs to provide an alternate means of performing an allocation from a specific range of memory (like provided by alloc_pages_range()) and insure that all drivers use that call. In that case the arches alloc_dma_coherent() may need to be modified to call alloc_pages_range() instead of relying on GFP_DMA. Signed-off-by:
-
Christoph Lameter authored
sh / sh64: Remove ZONE_DMA remains. Both arches do not need ZONE_DMA Signed-off-by:
Paul Mundt <lethal@linux-sh.org> Signed-off-by:
-
