gmane.linux.kernel

[PATCH] Fix build warning in kernel/auditsc.c

Hannes Eder — 2008-08-29T20:33:55

Fix this: kernel/auditsc.c: In function 'audit_match_perm': kernel/auditsc.c:249: warning: ISO C90 forbids mixed declarations and code Signed-off-by: Hannes Eder hanneseder.net> --- kernel/auditsc.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/kernel/auditsc.c b/kernel/auditsc.c index 972f8e6..e72b161 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c < at >< at > -243,10 +243,12 < at >< at > static inline int open_arg(int flags, int mask) static int audit_match_perm(struct audit_context *ctx, int mask) { +unsigned n; + if (unlikely(!ctx)) return 0; -unsigned n = ctx->major; +n = ctx->major; switch (audit_classify_syscall(ctx->arch, n)) { case 0:/* native */ if ((mask & AUDIT_PERM_WRITE) &&

Re: Is SKAS still required for UML

Nix — 2008-08-30T11:46:03

On 27 Aug 2008, Jeff Dike stated: Nobody's forward-ported it to recent kernels that I can see, though :/

Re: sched_mc_power_savings broken with CGROUPS+CPUSETS

Peter Zijlstra — 2008-08-30T11:26:53

I don't think iterating the domains and setting the flag is sufficient. Look at this crap (found in arch/x86/kernel/smpboot.c): cpumask_t cpu_coregroup_map(int cpu) { struct cpuinfo_x86 *c = &cpu_data(cpu); /* * For perf, we return last level cache shared map. * And for power savings, we return cpu_core_map */ if (sched_mc_power_savings || sched_smt_power_savings) return per_cpu(cpu_core_map, cpu); else return c->llc_shared_map; } which means we'll actually end up building different domain/group configurations depending on power savings settings. Yeah - I was primarity hinting at the sched_group and find_*_group() fudge, esp find_busiest_group() is an utter nightmare. I'm still struggeling to understand _why_ we need those group things to begin with, why aren't the child domains good enough?

Fw: [PATCH 3/5] V4L/DVB: follow lspci device/vendor style

Mauro Carvalho Chehab — 2008-08-30T11:22:48

I'm forwarding this message to LKML, since the previous one got rejected there, due to an error at CC list. Cheers, Mauro. Forwarded message: Date: Fri, 29 Aug 2008 17:57:56 -0300 From: Mauro Carvalho Chehab infradead.org> To: Bjorn Helgaas hp.com> Cc: Jesse Barnes virtuousgeek.org>, Thomas Gleixner linutronix.de>, Ingo Molnar redhat.com>, H. Peter Anvin zytor.com>, David Airlie linux.ie>, Samuel Ortiz sortiz.org>, Matthew Wilcox wil.cx>, linux-kernel< at >vger.kernel.org Subject: Re: [PATCH 3/5] V4L/DVB: follow lspci device/vendor style On Mon, 25 Aug 2008 15:45:09 -0600 Bjorn Helgaas hp.com> wrote: I've applied at V4L/DVB ML and should add it at linux-next. Thanks. Cheers, Mauro Cheers, Mauro

Re: 2.6.27 mtrr fixes do not work

Helge Hafting — 2008-08-30T11:11:14

Thank you for taking interest. The files are attached. The card is a nvidia card, but I use it in VESA mode because the binary driver hangs the machine once or twice a week. Helge Hafting Linux version 2.6.27-rc4-hh1 (helgehaf< at >hh2) (gcc version 4.3.1 (Debian 4.3.1-2) ) #4 SMP Thu Aug 28 14:58:33 CEST 2008 Command line: root=/dev/sda2 ro mtrr_spare_reg_nr=3 debug KERNEL supported cpus: Intel GenuineIntel AMD AuthenticAMD Centaur CentaurHauls BIOS-provided physical RAM map: BIOS-e820: 0000000000000000 - 000000000009f000 (usable) BIOS-e820: 000000000009f000 - 00000000000a0000 (reserved) BIOS-e820: 0000000000100000 - 000000007fe5a800 (usable) BIOS-e820: 000000007fe5a800 - 0000000080000000 (reserved) BIOS-e820: 00000000f8000000 - 00000000fc000000 (reserved) BIOS-e820: 00000000fec00000 - 00000000fec10000 (reserved) BIOS-e820: 00000000fed18000 - 00000000fed1c000 (reserved) BIOS-e820: 00000000fed20000 - 00000000fed90000 (reserved) BIOS-e820: 00000000feda0000 - 00000000feda6000 (reserved) BIOS-e8

Re: [PATCH 4/6] VM: Annotate pagealloc

Pekka Enberg — 2008-08-30T10:57:55

Hi Steve, On Sat, Aug 30, 2008 at 2:16 AM, Steve VanDeBogart nerdbox.net> wrote: No SHOUTING macros please, and I think you should just pass page and order here and do the calculation elsewhere. Same here. Other than that, looks good to me.

Re: MTD/block regression (was Re: Slub debugging NAND error in 2.6.25.10.atmel.2)

Haavard Skinnemoen — 2008-08-30T10:49:45

On Sat, 30 Aug 2008 10:34:12 +0900 FUJITA Tomonori lab.ntt.co.jp> wrote: Ah, great. Sorry for not searching the list before posting. Yeah, I sort of suspected that MTD was doing something unusual. Unfortunately, I'm not familiar enough with the MTD and block code to help out here... Haavard

Re: [PATCH 5/6] slab: Annotate slab

Pekka Enberg — 2008-08-30T10:50:01

Hi Steve, On Sat, Aug 30, 2008 at 2:17 AM, Steve VanDeBogart nerdbox.net> wrote: OK, I'm biased (I'm one of the kmemcheck developers) but these hooks to SLAB are too ugly to live with. My preferred solution is that you reuse the kmemcheck annotations kmemcheck_slab_alloc()/kmemcheck_slab_free() we have: http://git.kernel.org/?p=linux/kernel/git/x86/linux-2.6-tip.git;a=commitdiff;h=30532cb3c49a2a9fed94127aab26003c52398a51 Maybe making valgrind code a "UML port" (plus some more) of kmemcheck. I'm not sure why you want to treat caches with constructor differently. Sure, the memory regions *are* initialized but from programmer's point of view you're not supposed to be touching the memory unless you got it from kmalloc() or kmem_cache_alloc(). Same goes for kfree() and kmem_cache_free() -- no touchy touchy after you pass a pointer to either of the functions (unless you're RCU, of course). Pekka

[PATCH] Re: Linux 2.6.27-rc5

Helge Deller — 2008-08-30T10:34:51

commit da574983de9f9283ba35662c8723627096e160de Author: Al Viro zeniv.linux.org.uk> [PATCH] fix hpux_getdents() - Missing checks for -EFAULT, broken handling of overflow. introduced a build breakage: CC arch/parisc/hpux/fs.o arch/parisc/hpux/fs.c: In function `filldir': arch/parisc/hpux/fs.c:107: error: `buffer' undeclared (first use in this function) make[1]: *** [arch/parisc/hpux/fs.o] Error 1 This patch fixes it. Signed-off-by: Helge Deller gmx.de> diff --git a/arch/parisc/hpux/fs.c b/arch/parisc/hpux/fs.c index 69ff671..12c04c5 100644 --- a/arch/parisc/hpux/fs.c +++ b/arch/parisc/hpux/fs.c < at >< at > -104,7 +104,7 < at >< at > static int filldir(void * __buf, const char * name, int namlen, loff_t offset, buf->count -= reclen; return 0; Efault: - buffer->error = -EFAULT; + buf->error = -EFAULT; return -EFAULT; }

[PATCH] Avoid double flush in do_tlb_flush_all()

Andi Kleen — 2008-08-30T10:20:12

Avoid double flush in do_tlb_flush_all() leave_mm already flushes the TLBs, so don't need to flush directly before that in do_flush_tlb_all(). Avoid the double flush. Pointed out by Rainer Weikusat Signed-off-by: Andi Kleen linux.intel.com> --- arch/x86/kernel/tlb_32.c | 5 ++--- arch/x86/kernel/tlb_64.c | 7 +++---- 2 files changed, 5 insertions(+), 7 deletions(-) Index: linux-2.6.27-rc4-misc/arch/x86/kernel/tlb_32.c =================================================================== --- linux-2.6.27-rc4-misc.orig/arch/x86/kernel/tlb_32.c +++ linux-2.6.27-rc4-misc/arch/x86/kernel/tlb_32.c < at >< at > -229,11 +229,10 < at >< at > EXPORT_SYMBOL(flush_tlb_page); static void do_flush_tlb_all(void *info) { -unsigned long cpu = smp_processor_id(); - -__flush_tlb_all(); if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_LAZY) leave_mm(cpu); +else +__flush_tlb_all(smp_processor_id()); } void flush_tlb_all(void) Index: linux-2.6.27-rc4-misc/arch/x86/kernel/tlb_64.c ======================================

[PATCH v2 5/5] Add PERL script to conrol the PLL via the sysfs attribute

Kevin Diggs — 2008-08-30T10:07:50

This patch adds a PERL script that can be used to manipulate the sysfs attribute provided by pll_if. It can query the current state, change the state of the inactive PLL, and switch PLLs (assumming that the inactive PLL is locked to a valid frequency). My name is Kevin Diggs and I approve this patch. Signed-off-by: Kevin Diggs hypersurf.com> Index: Documentation/cpu-freq/pll.pl =================================================================== --- /dev/null2004-08-10 18:55:00.000000000 -0700 +++ Documentation/cpu-freq/pll.pl2008-08-29 13:34:36.000000000 -0700 < at >< at > -0,0 +1,773 < at >< at > +#!/usr/bin/perl -w + +=head1 NAME + +pll.pl - Provide user friendly interface to sysfs attribute for the 750GX PLL. +This uses the pll_if module. + +=head1 SYSNOPSIS + + pll.pl [ -bdhtv ] [ -f { clk | ratio }] + b:print bus frequency + d:dump current pll configuration + h:print simple usage information + f:set frequency to specified clk (MHz) or ratio (r suffix) + t:toggle selected pll. Use with -f to cause a swit

Re: [PATCH, RFC, tip/core/rcu] v3 scalable classic RCU implementation

Lai Jiangshan — 2008-08-30T09:58:26

I just had a fast review. so my comments is nothing but cleanup. Thanks, Lai. Paul E. McKenney wrote: I tried a mask like qsmaskinit before. The system came to deadlock when I did on/offline cpus. I didn't find out the whys for I bethought of these two problem: problem 1: ----race condition 1: synchronize_rcu rcu_offline_cpu -----race condition 2: rcu_online_cpu synchronize_rcu in these two condition, synchronize_rcu isblocked for ever for synchronize_rcu have to wait a cpu in rnp->qsmask, but this cpu don't run. problem 2: we need call rcu_offline_cpu() in these two cases in rcu_cpu_notify() since qsmaskinit had changed by rcu_online_cpu() case CPU_UP_CANCELED: case CPU_UP_CANCELED_FROZEN: long mask -> unsigned long mask long mask -> unsigned long mask struct rcu_state has a field: struct rcu_data *rda[NR_CPUS] so we can move these lines around __call_rcu into

[PATCH v2 3/5] Various arch/powerpc changes to support the 750GX cpufreq driver

Kevin Diggs — 2008-08-30T09:49:04

This patch includes various changes necessary to support the cpufreq driver for the PowerPC 750GX. Highlights include adding entries to recognize the 750GX to the cputable (This includes the ... unfortunate pvr that the 750GX used in the PowerLogix PowerForce 750GX => 0x0008 0203). The PLL switch in idle_6xx.S during sleep (or nap) was also removed (I tried to add ability to disable this but the result would not boot?). My name is Kevin Diggs and I approve this patch. Signed-off-by: Kevin Diggs hypersurf.com> Index: arch/powerpc/kernel/Makefile =================================================================== --- arch/powerpc/kernel/Makefile.orig2008-08-13 02:19:18.000000000 -0700 +++ arch/powerpc/kernel/Makefile2008-08-14 02:50:18.000000000 -0700 < at >< at > -17,6 +17,7 < at >< at > obj-y:= cputable.o ptrace.o syscalls init_task.o process.o systbl.o idle.o \ signal.o obj-y+= vdso32/ +obj-y+= cpu/ obj-$(CONFIG_PPC64)+= setup_64.o sys_ppc32.o \ signal_64.o ptrace32.o

[PATCH v2 2/5] Add cpufreq driver for the IBM PowerPC 750GX

Kevin Diggs — 2008-08-30T09:40:18

This adds the actual cpufreq driver for the 750GX. It supports all integer ratios that are valid for the processor model and bus frequency. It is called cf750gx. It has two modes of operation. Normal mode uses all valid frequencies. In minmaxmode, only the minimum and maximum are used. This provides the ability for very low latency frequency switches. There is also a sysfs attribute to have it switch off the unused PLL for additional power savings. This does NOT support SMP. I fixed a deadlock if the module is unloaded before _target() is ever called. My name is Kevin Diggs and I approve this patch. Signed-off-by: Kevin Diggs hypersurf.com> Index: Documentation/DocBook/cf750gx.tmpl =================================================================== --- /dev/null2004-08-10 18:55:00.000000000 -0700 +++ Documentation/DocBook/cf750gx.tmpl2008-08-27 13:59:45.000000000 -0700 < at >< at > -0,0 +1,441 < at >< at > + +

Re: [PATCH, RFC, tip/core/rcu] v3 scalable classic RCUimplementation

Peter Zijlstra — 2008-08-30T09:33:00

Already done and available in the -tip tree, curtesy of Mathieu.

[PATCH v2 1/5] Add low level PLL config register interface module

Kevin Diggs — 2008-08-30T09:28:23

This adds a small module to handle the low level details of dealing with the PLL config register (HID1) found in the IBM 750GX. It provides 2 possible interfaces, both selectable via kernel config options. One is a sysfs attribute and the other is a normal function API. It is called pll_if. The determination of the bus frequency is what worked on a PowerMac 8600. Any suggestions on a more general solution are welcome. After receiving comments, I added an argument for the data item to the notifier register functions exported for the cpufreq API. I also fixed a deadlock if the module is unloaded before _modify_PLL() is ever called. My name is Kevin Diggs and I approve this patch. Signed-off-by: Kevin Diggs hypersurf.com> Index: arch/powerpc/kernel/cpu/pll_if.c =================================================================== --- /dev/null2004-08-10 18:55:00.000000000 -0700 +++ arch/powerpc/kernel/cpu/pll_if.c2008-08-29 13:42:35.000000000 -0700 < at >< at > -0,0 +1,807 < at >< at > +/* + * pll_if.c - low level i

2.6.27-rc5 nfs stall - more info

Grant Coady — 2008-08-30T09:28:11

Hi there, Had another stall, server running 2.6.24.7, clients running 2.6.27-rc5. On the server the netstat -t gave during stall: grant< at >deltree:~$ netstat -t Active Internet connections (w/o servers) Proto Recv-Q Send-Q Local Address Foreign Address State tcp 0 0 localhost:50087 localhost:4743 ESTABLISHED tcp 0 52 deltree.mire.mine.n:ssh sillywin.mire.mine:3052 ESTABLISHED tcp 0 0 deltree.mire.mine.n:ssh sillywin.mire.mine:3051 ESTABLISHED tcp 0 0 deltree.mire.mine.n:ssh sillywin.mire.mine:1056 ESTABLISHED tcp 0 0 localhost:4743 localhost:50087 ESTABLISHED after stall, shows the delay seems to be in making the NFS connection: grant< at >deltree:~$ netstat -t Active Internet connections (w/o servers) Proto Recv-Q Send-Q Local Address Foreign Address State tcp 0 0 localhost:50087 localhost:4743 ESTABLISHED tcp 0 0 deltree.mire.mine.n:ssh silly

[PATCH v2 0/5] Add a cpufreq driver for the IBM PowerPC 750GX

Kevin Diggs — 2008-08-30T09:18:36

Hi, This patch set adds a cpufreq driver for the IBM PowerPC 750GX processor. It "should" also work for the 750FX. The patches are: 1) Add low level PLL config register interface module 2) Add cpufreq driver for the 750GX 3) Other PowerPC kernel changes necessary to support the above 4) Add kernel doc for the completion feature, fix split-man.pl in kernel-doc-nano-HOWTO.txt 5) Add pll script to interface with pll_if sysfs attribute These changes are against 2.6.26. Thanks for all who took the time to review v1! Documentation/DocBook/Makefile | 2 Documentation/DocBook/cf750gx.tmpl | 441 ++++++++++++++++ Documentation/cpu-freq/pll.pl | 773 ++++++++++++++++++++++++++++ Documentation/kernel-doc-nano-HOWTO.txt | 4 arch/powerpc/kernel/Makefile | 1 arch/powerpc/kernel/cpu/Makefile | 6 arch/powerpc/kernel/cpu/cpufreq/Kconfig | 33 + arch/powerpc/kernel/cpu/cpufreq/Makefile | 1 arch/powerpc/kernel/cpu/cpufreq/cf750gx.c | 741 +++

[Regression] sshd hangs on close

Matthias Urlichs — 2008-08-30T09:18:47

"top" sees multiple instances of: 6008 root 20 3184 D 0 0:00.06 flush_cpu sshd: root< at >pts/13 This is 2.6.27rc5, preemption and tickless turned on, on a 4-CPU machine. (OK, two hyperthreaded ones :-P) strace() shows that this happens on the last close of its /dev/ptmx file descriptor, which hangs. I'll do a bisect early next week unless somebody else identifies a likely suspect.

Re: SATA Cold Boot problems on >2.6.25 with NV

Tejun Heo — 2008-08-30T09:14:39

Hmm... Can you post full dmesg output? We used to see things like above when ATAPI CHECK SENSE handling somehow failed to tell EH that it was an exception not worth whining about. Maybe EH action mask is not being cleared properly? I somehow feel pretty optimistic about that part. :-)

Re: [Patch 0/6] MIPS: Hardware watch register support for gdb (version 3).

David Daney — 2008-08-30T08:37:05

I think there will have to be at least one more pass at this. The current design assumes that all debug registers support an identical set of the I, R, and W bits and Mask. However sections 8.23 and 8.24 of my MIPS32® Architecture For Programmers Volume III: The MIPS32® Privileged Resource Architecture indicate that they do not have to uniform. I will have to augment the ptrace structures to report the values for each register instead of a single global value. David Daney