Summary of changes from v2.6.4 to v2.6.5-rc1 ============================================ PPC32: Added big-endian cfg_addr access PPC32: Simplified handling of big/little endian pci indirect access [NET] add netpoll API [NET] Add netpoll support for tg3 [NET] use the netpoll API to transmit kernel printks over UDP [PATCH] netpoll for 3c59x >From the old -aa tree with minor changes. Orginally done by Andrea I think. [PATCH] netpoll for tulip Netpoll for tulip. Uses disable_irq() because tulip is unfortunately still lockless. [PATCH] netpoll for amd8111e netpoll for amd8111e [PATCH] Netpoll for pcnet32 netpoll for pcnet32 [PATCH] fix tg3 netpoll No need to use disable_irq because tg3 is properly spinlocked. Can just call the interrupt handler directly. [PATCH] netpoll for eepro100 netpoll for eepro100 This was in Ingo's old original netconsole patches. [PATCH] 3c59x netpoll typo Poll code (in net-drivers-2.5-exp) was calling undefined function. [PATCH] netpoll: fix compilation with CONFIG_NETPOLL_RX Fix compilation without CONFIG_NETPOLL_RX [PATCH] netpoll: push zap_completion_queue for lkcd Move zap_completion_queue call inside netpoll so we don't need to export it separately. [netdrvr smc-ultra] netpoll support Hi Jeff, Below is the pollcontroller patch for smc ultra net driver. This patch can be applied over 2.6.0-test9-bk25-netdrvr-exp1.patch [netdrvr tlan] netpoll support Hi Jeff, Below is the pollcontroller patch for tlan network device driver. This patch can be applied over 2.6.0-test9-bk25-netdrvr-exp1.patch [netdrvr via-rhine] add netpoll support [PATCH] netpoll abort for bad interface Fix bug freeing nonexistent dev when aborting on bad interface name. tiny-mpm/net/core/netpoll.c | 2 +- 1 files changed, 1 insertion(+), 1 deletion(-) [netdrvr] add netpoll support to several 8390-based drivers [PATCH] netconsole init return code On Sat, Jan 10, 2004 at 08:07:09PM -0500, Jeff Garzik wrote: > Matt Mackall wrote: > > tiny-mpm/drivers/net/netconsole.c | 4 ++-- > > 1 files changed, 2 insertions(+), 2 deletions(-) > > > >diff -puN drivers/net/netconsole.c~netconsole-error > >drivers/net/netconsole.c > >17:40:59.000000000 -0600 > >+++ tiny-mpm/drivers/net/netconsole.c 2004-01-10 17:42:12.000000000 -0600 > >@@ -105,11 +105,11 @@ static int init_netconsole(void) > > > > if(!configured) { > > printk("netconsole: not configured, aborting\n"); > >- return 1; > >+ return -EINVAL; > > } > > > > if(netpoll_setup(&np)) > >- return 1; > >+ return -EINVAL; > > > oops. obviously I read this after I sent a previous email. > > Can you please resend the "configured" patch, which I already deleted? tiny-mpm/drivers/net/netconsole.c | 14 ++++++++++---- 1 files changed, 10 insertions(+), 4 deletions(-) [PATCH] netconsole init return code tiny-mpm/drivers/net/netconsole.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) [PATCH] netpoll carrier handling On Sat, Jan 10, 2004 at 08:04:10PM -0500, Jeff Garzik wrote: > > if you are gonna all verbal n such with 'atleast' and 'atmost', then use > time_after() rather than !time_before() ;-) tiny-mpm/net/core/netpoll.c | 16 ++++++++++++---- 1 files changed, 12 insertions(+), 4 deletions(-) [PATCH] fix netpoll printk bug [PATCH] fix for netpoll braindamage for 64-bit Spotted by Jim Houston on amd64. [PATCH] netconsole warning fix drivers/net/netconsole.c:99: warning: initialization from incompatible pointer type [PATCH] I2C: i2c-voodoo3.c needs I2C_ADAP_CLASS_TV_ANALOG The Voodoo3 i2c bus has either a bt869 tv-out chip, or also a tv tuner, decoder and msp3400. Without I2C_ADAP_CLASS_TV_ANALOG, the i2c clients would have to do a strcmp of the adapter name to distiguish between the i2c and ddc adapters. Yes, they should be able to tell if the chip at a given address is what they are looking for, but in the case of the v3tv module, which is the v4l device, in the 2.4 kerenl I've got it set to create a dummy client, and the strcmp is the only way to distinguish the i2c from the ddc. In the 2.6 kernel, class can be defined, simplifying things for the v3tv module. [PATCH] I2C: Enable debugging in fscher I just noticed that I forgot to enable debugging in the new fscher driver. Sorry for that. Here is a patch that fixes this. [PATCH] I2C: Credit James Bolt in w83l785ts I forgot to give the credit he deserves to James Bolt for testing the latest w83l785ts driver changes. Here is a patch that fixes that. [PATCH] I2C: New chip driver ported: lm80 Here is a port for the lm80 driver, which supports LM80 chips by National Semiconductor. The port is mostly the work of Tiago Sousa, which I reviewed and polished as necessary. [PATCH] I2C: update I2C help text VIDEO_BT848 selects I2C_ALGOBIT. [PATCH] I2C: Support for IXP42x GPIO-based I2C Following is an updated patch to add support for using the GPIO lines on an Intel IXP42x SOC for I2C. Changes since last patch include: - Cleanups as per Jean Delvare's comments - Rename functions as 42x instead of 425 - Change the comments at the top of the file to better reflect the purpose of this driver. This code has been tested with the eeprom client driver and worked w/o any problems. I2C: fix oops in i2c-ali1535 driver if no hardware is present. Thanks to Dave Jones for pointing this out. PPC32: More cleanups of the IBM Spruce code. From Randy Vinson . PPC32: Fix a thinko in the gen550 code. From Randy Vinson . PPC32: Make {in,out}[bwl] be consistent on all platforms. This exposes warnings in todc_time.c for everyone, and then fixes them. PPC32: Make sure the read in in_8, in_{le,be}{16,32} happens before we return. A forward port of changes from Paul Mackerras. [PATCH] eepro100.c warning fix drivers/net/eepro100.c:545: warning: `poll_speedo' declared `static' but never defined ALSA CVS update - Takashi Iwai Intel8x0 driver - added the ac97 quirk for Intel D845WN (82801BA) ALSA CVS update - Takashi Iwai AC97 Codec Core - fixed the mic gpio switch handling on the old ALC650. - fixed the wrong register initialization on ALC655/658. ALSA CVS update - Takashi Iwai AC97 Codec Core - fixed the detection of surround/LFE VRA on ALC650. ALSA CVS update - Takashi Iwai USB generic driver - assign PCM unique (sub)streams for each USB format type. this will avoid the mix up of format I and format III on M-audio transit. ALSA CVS update - Takashi Iwai Digigram VX core - fixed the wrong lock (bug #2052) - use spin_lock_irqsave() now. ALSA CVS update - Takashi Iwai Documentation,ALSA Core,PCI drivers,MIXART driver,IGNORE - added snd-mixart driver for Digigram miXart boards. ALSA CVS update - Takashi Iwai Digigram VX core - fixed another wrong lock. ALSA CVS update - Takashi Iwai ICE1712 driver,ICE1724 driver - added the support of independent surround PCM for ice1724. ALSA CVS update - Takashi Iwai Intel8x0 driver - return with IRQ_RETVAL() for the unknown IRQ bits (often found on nForce2) with debug messages. ALSA CVS update - Takashi Iwai VIA82xx driver - changed the default DXS of GA-7VAXP to NO_VRA. ALSA CVS update - Takashi Iwai AC97 Codec Core - fix for AD1885 - set the default MISC register value. - disabled 'digital audio mode', which seems problematic on many boards. ALSA CVS update - Takashi Iwai AC97 Codec Core Clemens Ladisch : added AD1888 support. ALSA CVS update - Takashi Iwai USB generic driver Clemens Ladisch : Terratec PHASE 26 MIDI support ALSA CVS update - Takashi Iwai Digigram VX core Alain Cretet : - fixed the missing input VU meter. ALSA CVS update - Takashi Iwai Digigram VX core Alain Cretet : - fixed the 24bit mono recording. ALSA CVS update - Takashi Iwai ALSA sequencer - fixed the race conditions. ALSA CVS update - Takashi Iwai AC97 Codec Core - fixed the initial value of AD_MISC register for AD1885. ALSA CVS update - Jaroslav Kysela Intel8x0 driver Added spinlock to pointer callback - ichdev->position is not changed atomically ALSA CVS update - Takashi Iwai VIA82xx driver - added the dxs default for MSI KT6 Delta-SR. - fixed the calculation of rate bits (based on 0x100000). ALSA CVS update - Takashi Iwai MIXART driver - replaced the debug messages with snd_printdd(). ALSA CVS update - Takashi Iwai Intel8x0 driver - clean up the irq status bit debugging. - added nVidia Ck8S support. ALSA CVS update - Takashi Iwai ALSA<-OSS sequencer - fixed the behavior of SNDCTL_SEQ_IOCTL. (don't close the device). ALSA CVS update - Jaroslav Kysela ALSA<-OSS emulation Added period_frames to fix poll behavior ALSA CVS update - Jaroslav Kysela ALSA<-OSS emulation Fixed oss.period_frames setup ALSA CVS update - Jaroslav Kysela PCM Midlevel Added OSS period frames to proc interface ALSA CVS update - Jaroslav Kysela ALSA<-OSS emulation Fixed oops regarding last period_frames update ALSA CVS update - Takashi Iwai USB generic driver Clemens Ladisch : - added the quirk for Edirol UA-3FX. ALSA CVS update - Jaroslav Kysela AC97 Codec Core Fixed swap_headphone() when headpone controls do not exist ALSA CVS update - Jaroslav Kysela Trident driver Fixed s/pdif control initialization ALSA CVS update - Jaroslav Kysela RME HDSP driver Fixed wrong assert, added checks for copy_*_user functions ALSA CVS update - Jaroslav Kysela ALSA sequencer Clemens Ladisch Timestamping (if enabled on a subscription or a port) is not applied to the quoted event but to the quoting event. This patch adds a function to copy only selected fields into the event to be delivered. - fix KERNEL_QUOTE event timestamping - fix typo in port_broadcast_event ALSA CVS update - Jaroslav Kysela MPU401 UART Clemens Ladisch remove unneeded technical information from port names ALSA CVS update - Jaroslav Kysela ALSA sequencer Clemens Ladisch This patch reverses the order of the 'Rawmidi x' and rawmidi name parts of client names to enable selecting clients by a unique prefix (as snd_seq_parse_address does). ALSA CVS update - Jaroslav Kysela USB generic driver Clemens Ladisch - add device-specific port names - begin numbering ports at 1 ALSA CVS update - Jaroslav Kysela ALSA sequencer Clemens Ladisch - fix typo in port flags - add GM2 capability bit ALSA CVS update - Jaroslav Kysela MIXART driver Added missing header file inclusion ALSA CVS update - Jaroslav Kysela Documentation,ALSA Core,PCMCIA Kconfig,PCMCIA Sound Core PDAudioCF driver Added Sound Core PDAudioCF driver ALSA CVS update - Jaroslav Kysela EMU10K1/EMU10K2 driver Fixed page overflow ALSA CVS update - Jaroslav Kysela Serial BUS drivers Moved AK4117 from alsa-driver tree to satisfy dependency for PDAudioCF driver ALSA CVS update - Jaroslav Kysela AK4117 receiver Added missing ak4117.h file Fixed compilation of PDAudioCF driver [PATCH] I2C: fix space in message BTW, I found something to be fixed in i2c-core: [PATCH] I2C: fix mor rmmod oopses > Oh nevermind, that's just a dumb driver. It's doing a release_region > on memory it didn't get. Stupid, stupid, stupid... While we're at it, what about fixing two other drivers that obviously have the same problem? (BTW I didn't get an oops as I tried reproducing the problem, only a "Trying to free nonexistent resource" in dmesg.) [PATCH] I2C: Lowercase chips name This patch brings the name field of adm1021.c, it87.c and via686a.c in conformance with the defined standard ("all lowercase, as simple as the driver name itself"). [PATCH] I2C: fix Hangs with w83781d Here is a patch for the w83781d driver that prevents register bits from being arbitrary changed when we force temp2/3 to comparator mode. Keith Duthie had been reporting various problems with that driver and finally found that this arbitrary change was the cause of them. He also tested this patch, which he confirmed to work. [PATCH] I2C: fix it87 sensor type Here comes a patch by Takeru Komoriya which fixes the way the it87 driver handles temperature sensor types selection. * Use the same values as the CVS driver and sensors program. * Better comments. * Get rid of the old setting method (already gone in CVS). * Handle invalid values correctly. [PATCH] I2C: fix another oops in i2c-core with debug Some times ago, you fixed an oops in i2c-core when debugging is enabled: http://marc.theaimsgroup.com/?l=linux-kernel&m=107585749612115&w=2 Looks like you missed that second one: [PATCH] I2C: Remove asb100 support from w83781d Now that we have a separate asb100 driver, we can remove the (bad) support of it from the w83781d driver. Following patch does this. I've already cleaned this up in our CVS repository. [PATCH] I2C: update for sysfs-interface documentation [PATCH] I2C: rename sysfs files, part 1 of 2 Here it is. The associated libsensors patch is here: http://jdelvare.net1.nerim.net/sensors/libsensors-sysfs-names-1.diff (not applied yet, on purpose) [PATCH] I2C: rename sysfs files, part 2 of 2 Here is the second step of my sysfs renaming plan. This one does the following renames (as I already announced on the LKML): temp_hyst -> temp_max_hyst or temp_crit_hyst sensor -> temp_type pwm -> fan_pwm pwm_enable -> fan_pwm_enable vid -> in_ref The associated libsensors patch is here: http://jdelvare.net1.nerim.net/sensors/libsensors-sysfs-names-2.diff (not applied yet, on purpose) Note that the w83781d part is a bit more complex, not only because it is the only driver to require the 5 changes, but also because at some point the macros assume that the internal variable names match the sysfs names, so I had to change them too (better than rewriting the macros, methinks). For reference, here is the list of changes, by driver: asb100: hyst -> max_hyst (4) pwm -> fan_pwm (1) pwm_enable -> fan_pwm_enable (1) vid -> in_ref (1) fscher: pwm -> fan_pwm (3) gl518sm: hyst -> max_hyst (1) it87: sensor -> temp_type (3) lm75: hyst -> max_hyst (1) lm78: hyst -> max_hyst (1) vid -> in_ref (1) lm85: pwm -> fan_pwm (3) pwm_enable -> fan_pwm_enable (3) vid -> in_ref (1) lm90: hyst -> crit_hyst (2) via686a: hyst -> max_hyst (3) w83781d: hyst -> max_hyst (2) sensor -> temp_type (3) pwm -> fan_pwm (4) pwm_enable -> fan_pwm_enable (1) vid -> in_ref (1) There's also a documentation update. There will be one more after that (to make it more readable, no contents change), and a patch to lm83.c to bring it to compliance with the (new) standard. (If you wonder why I did not change it with the other drivers: because it was *already* not in compliance with the old standard. There's some real work to do for this one.) [PATCH] PCI: fix i2c quirk for SiS735 chipset SMBus driver I2C: fix compiler warnings in 2 drivers. PPC32: print useful flags in oops, like x86 / ppc64. [PATCH] SCSI: Recognize device type 0x0C Device type 0x0C is RAID, so show it as such [PATCH] SCSI: remove some SGI devices from the device list Some SGI RAID devices identify as being SCSI-3, so they don't need to be in the scsi_static_device_list any more. I've tested to make sure that all luns still show up in a sparse, large config. So I'm proposing removing them. Thanks to hch for the suggestion. ALSA CVS update - Jaroslav Kysela USB generic driver usb_ch9.h is already included in usb.h ALSA CVS update - Takashi Iwai ALS4000 driver - fixed the build without joystick support. ALSA CVS update - Takashi Iwai CMIPCI driver - use 0x201 as the default joystick port address. ALSA CVS update - Takashi Iwai Documentation - fixed the description for snd-cmipci module option. ALSA CVS update - Takashi Iwai AC97 Codec Core - fixed the handling of S/PDIF rates. the IEC958 status bits are updated according to the current rate. ALSA CVS update - Takashi Iwai MIXART driver - fixed the race condition in message flow. - removed obsolete debug prints. - make prepare callback non-atomic. - synchronize with the pending messages in prepare and hw_free callbacks. ALSA CVS update - Takashi Iwai PCM Midlevel,Intel8x0 driver - added snd_pcm_limit_hw_rates() to determine the min/max rates from rates bits. ALSA CVS update - Takashi Iwai VIA82xx driver - added dxs_support default for EPIA MII. ALSA CVS update - Jaroslav Kysela EMU10K1/EMU10K2 driver James Courtier-Dutton , some additions ALSA CVS update - Jaroslav Kysela Sound Core PDAudioCF driver akpm@osdl.org Fix pdaudiocf_irq.c for gcc-3.5 ALSA CVS update - Takashi Iwai Documentation - fixed the example code of ctl info callback with enum type. ALSA CVS update - Takashi Iwai VIA82xx driver - added the dxs_support default for Uniwill laptop. SCSI: mptfusion update to 3.00.04 From: Moore, Eric Dean [PATCH] I2C: add w83627hf driver The following patch against kernel 2.6.3 adds the w83627hf driver, I have ported from the 2.4 version. Jean Delvare has asked me to send this patch to you for applying. I have tested it on w83627thf. Mark Hoffman has reviewed the code and has also tested it on w83627thf. ia64: minor 2.6 sba_iommu update This is a minor update to sba_iommu initialization code. I moved the relaxed ordering to the zx1 specific init routine, those registers don't do anything on sx1000. The iommu page size setup is more generalized on ACPI ID, so the sx1000 specific init routine goes away. And the virtual dma_mask on zx1 was tweaked a little to better reflect the memory address space. ALSA CVS update - Takashi Iwai PPC Tumbler driver fixed the resume of bass/treble volumes on snapper. ALSA CVS update - Jaroslav Kysela ALSA Core Fixed snd_info_set_text_ops() wwhen CONFIG_PROC_FS is not defined ALSA CVS update - Jaroslav Kysela Sound Core PDAudioCF driver Fixed pcm->name settings ALSA CVS update - Jaroslav Kysela AK4531 codec Aux Input Route -> Aux Capture Route renaming ALSA CVS update - Jaroslav Kysela PCI drivers Select CONFIG_VIDEO_DEV when CONFIG_SND_FM801_TEA575X is wanted ALSA CVS update - Jaroslav Kysela OPL3,OPL4,Synth Fixed sequencer dependency for opl3, opl4 and emux objects. ALSA CVS update - Jaroslav Kysela PCI drivers fix Kconfig thinko ALSA - 1.0.3 [PATCH] I2C: it87 reset option I wrote a patch which adds reset option to the it87 driver talking with Jean Delvare. * Do not reset the registers unless users want to do because resetting registers makes all fans go to full power and we can usually rely on values set by BIOS * Remove all limit initializations as they should be done from user-space * Better register mask for start of monitoring [PATCH] ia64: add zx1_defconfig Here's my attempt at a zx1_defconfig. I haven't checked it on many of the zx1 boxes (only zx2000 and rx2600) but I've set some options I know are needed on the other boxes. [PATCH] ia64: SAL cleanup This patch reorganises sal.c and adds a small amount of new functionality. - Introduce sal_revision to report what revision of the SAL spec is supported by the system. - Introduce sal_version to report what version of the vendor's SAL implementation is present. - Introduce SAL_VERSION_CODE to allow for easy comparisons. - Print the version at boot, and remove the 'oem=' and 'vendor=' strings. - Refactor ia64_sal_init() into several smaller functions. - Delete the dead variables 'max' and 'min'. - Stop printing the pal_proc and sal_proc entry addresses. - Print "None" if there are no SAL platform features. [PATCH] ia64: Add support for extended PCI config space Support for extended config space on ia64. - Add the new parameter 'type' to ia64_sal_pci_config_{read,write} - Change callers to match. - Don't check `value' for NULL -- drivers/pci/access.c guarantees it isn't. - Make pci_sal_ops static. - Add pci_sal_ext_ops. - Introduce pci_set_sal_ops() as an arch_initcall to ensure the raw_pci_ops get set before we walk the ACPI namespace to discover PCI root bridges ia64: Move irq_enter()/irq_exit() from hardirq.h to irq_ia64.c. The work done by these routines is very special and needs to be done at exactly the right time. Removing it from the header-file reduces the risk of accidental misuse. Other arch maintainers agree that this is the Right Thing to do. [PATCH] ia64: simserial module refcounting update PPC32: Kill off arch/ppc/boot/prep and rearrange some files. [PATCH] ini9100u build fix From: Christoph Hellwig - Remove dead forward declarations - Fix compilation of the interrupt handler. [PATCH] SCSI sysfs host name support OK, let's export the proc_name then. It's some name that can be used to identify the driver, works for code both compiled statically or as module, is reasonably short and very often resembles the module name. For the module_name, let's wait for a generic sysfs solution. PPC32: Update the TODC code from 2.4. PPC32: Add and make use of ppc_md.rtc_{read,write}_val. We can then fix IBM PReP machines and their RTCs. PPC32: Fix 'make znetboot' on CONFIG_PPC_MULTIPLATFORM. PPC32: Fix an old thinko in arch/ppc/boot/simple/relocate.S ia64: Rename ia64_invoke_kernel_thread_helper() to start_kernel_thread() for symmetry with start_kernel() and to make it obvious when the end of the call-chain has been reached. ia64: More SAL cleanups/fixes. Based on patch by Matthew Wilcox: make kernel work with old (broken) firmware again and fix UP build problems. [PATCH] SCSI tape sysfs name fixes - The sysfs directory names are changed from stxmy[n] to the names in devices.txt. Below is an example for the first tape drive: /sys/class/scsi_tape/ |-- nst0 |-- nst0a |-- nst0l |-- nst0m |-- st0 |-- st0a |-- st0l `-- st0m - Name generation for non-default number of modes fixed so that same minors get same names even with different number of modes. - devfs minor generation fixed to work with non-default number of modes - The alignment requirement is printed in the line logged when a tape is attached. ia64: fix missing include in include/asm-ia64/sn/router.h [PATCH] I2C: IOP3xx i2c driver update Here's a small patch update to the i2c-iop3xx.c in drivers/i2c/busses/. It fixes some functions' return value and updated the irq handler to be compatible with kernel 2.6. Thanks! [PATCH] I2C: show adapter name in i2c-dev class directory to make it easier for userspace tools. [PATCH] ia64: Don't assume iosapic interrupt controllers Some ia64 machines don't use an iosapic interrupt controller (a flagrant violation of the DIG64 spec), so don't assume it's there in mca.c. Platforms that provide a ACPI_INTERRUPT_CPEI vector are responsible for registering its interrupt controller type in platform specific code, iosapic.c:iosapic_register_platform_intr() does this for platforms with an iosapic where the CPEI vector is listed in the ACPI namespace, and on sn2, all external interrupts are assigned the irq_type_sn type, so this change of assumption should be safe. [PATCH] I2C: keep i2c-dev numbers in sync with i2c adapter numbers This makes userspace tools easier to figure out which i2c-dev device is assigned to which i2c adapter. Yes, we can overflow the i2c dev array right now, but that would take a lot of i2c adapter modprobe/rmmod cycles. That will be fixed up soon. asmlinkage acpi_enter_sleep_state_s4bios() - from Pavel Machek ALSA CVS update - Jaroslav Kysela FM801 driver tea575x can be module, too ALSA CVS update - Jaroslav Kysela Big DMA cleanup originated by Russell King * Russel - introduced 'struct device' support for 2.6 dma_alloc_coherent() * Jaroslav - removed all bus-specific allocation functions - extended snd_dma_alloc_pages/snd_dma_free_pages to handle all bus types - recoded all (or almost all) device drivers - sgbuf functions are bus independent now ALSA CVS update - Jaroslav Kysela CMI8330 driver,ES18xx driver,AD1816A driver,AD1848 driver,CS4231 driver ES1688 driver,GUS Library,Opti9xx drivers,SB16/AWE driver,SB8 driver Fixed old function name (snd_pcm_isa_flags -> snd_pcm_dma_flags) ALSA CVS update - Jaroslav Kysela ALSA Core Russell King This is part of a patch series to clean up sound/core/Makefile in Linux 2.6.4-rc1. - Add SND_TIMER, SND_PCM, SND_HWDEP and SND_RAWMIDI configuration symbols. These symbols select which modules in sound/core get built, building snd-timer, snd-pcm, snd-hwdep and snd-rawmidi respectively. - Add reverse dependencies (select) to select these symbols for core components where necessary. - Hide SND_OSSEMUL - we can select this when SND_MIXER_OSS, SND_PCM_OSS or SND_SEQUENCER_OSS are selected automatically. - Tweak Makefile to use these new symbols to build these modules. - Since we now build appropriate modules for core components according to the new configuration symbols, (eg, snd-timer if SND_SEQUENCER is selected) we can delete these duplications. ALSA CVS update - Clemens Ladisch USB generic driver Fix for Creamware Noah: search class-specific endpoint descriptor in the extra descriptors of the sync ep, too ALSA CVS update - Takashi Iwai Memalloc module,ALSA Core - fixed the lock up with SG-buffer handler. - removed non-existing export symbol. - clean up ifdefs. ALSA CVS update - Takashi Iwai ES1968 driver - fixed the handling of DMA buffer with the recent API change. ALSA CVS update - Takashi Iwai Intel8x0 driver - fixed the allocation/release of buffer descriptor table. Add SCSI lots of disk support From: Kurt Garloff Based on earlier patches from Badari Pulavarty , Matthew Wilcox [PATCH] add missing free sgtable in scsi_init_io error path I was testing different errors while running multipath, and becuase of a bug in dm-multiapth I hit the "Incorrect number of segments after building list" error. The attached patch just adds a missing scsi_release_buffers() at the bottom of the scsi_init_io where the table was allocated successfully, but you can return BLKPREP_KILL when someone messes up segment counts. ALSA CVS update - Takashi Iwai Memalloc module - added back the output of PCI dma buffers in proc file. ALSA CVS update - Jaroslav Kysela Intel8x0 driver Converted to new DMA allocation API ALSA CVS update - Jaroslav Kysela ARM,ALSA Core,Generic drivers,ISA,PARISC,PCI drivers,PCMCIA Kconfig,PPC SPARC,USB This is part of a patch series to clean up sound/core/Makefile in Linux 2.6.4-rc1. - Add 'select SND_PCM' statements to appropriate Kconfig entries for drivers whose configuration symbol is used to build snd-pcm, snd-timer, and snd-page-alloc. - Remove snd-pcm, snd-timer and snd-page-alloc from these in sound/core/Makefile. - Remove snd from these entries as well - all SND_xxx configuration symbols depend on CONFIG_SND, so we won't even consider building any of these drivers unless SND is already set to 'y' or 'm'. ALSA CVS update - Jaroslav Kysela ALSA Core,Generic drivers,ISA,PCI drivers,USB Russell King This is part of a patch series to clean up sound/core/Makefile in Linux 2.6.4-rc1. - Add SND_RAWMIDI for drivers which use the snd-rawmidi module. - Remove snd-rawmidi from these drivers entries in sound/core/Makefile - Remove any sound/core/Makefile entries which are left empty. ALSA CVS update - Jaroslav Kysela ALSA Core,ISA,PCI drivers,PCMCIA Kconfig Russell King This is part of a patch series to clean up sound/core/Makefile in Linux 2.6.4-rc1. - Add SND_HWDEP for drivers which use the snd-hwdep module. - Remove snd-hwdep from these drivers entries in sound/core/Makefile, removing any sound/core/Makefile entries which are left empty. ALSA CVS update - Jaroslav Kysela ALSA Core,Generic drivers Russell King This is part of a patch series to clean up sound/core/Makefile in Linux 2.6.4-rc1. - Add SND_TIMER for drivers which use the snd-timer module. - Remove snd-timer from these drivers entries in sound/core/Makefile, removing any sound/core/Makefile entries which are left empty. - Since the 'top level module dependency' lists are now gone, remove the comment. - Also, since we only mention objects once, remove the sorting of obj-m ALSA CVS update - Jaroslav Kysela Generic drivers,MPU401 UART,OPL3,OPL4,ISA,PCI drivers More Kconfig and Makefile cleanups following Russell's direction: - added SND_MP401_UART tristate - added SND_OPL3_LIB tristate - added SND_OPL4_LIB tristate ALSA CVS update - Jaroslav Kysela Generic drivers,Digigram VX core,PCI drivers,PCMCIA Kconfig More Kconfig and Makefile cleanups following Russell's direction: - added SND_VX_LIB tristate ALSA CVS update - Clemens Ladisch USB generic driver - fix non-working control port on Roland U-8 - more port names - new MIDI quirks for Roland MMP-2, V-SYNTH, VariOS, FP-*, GI-20, BOSS GS-10, Edirol UR-80, PCR-A, PCR-1 ALSA CVS update - Takashi Iwai Memalloc module fixed the missing inclusion. ALSA CVS update - Takashi Iwai PCI drivers,AC97 Codec Core - added CONFIG_SND_AC97_CODEC and simplify Kconfig and Makefiles. ALSA CVS update - Takashi Iwai Memalloc module - suppress allocation failure warnings with large buffers. ALSA CVS update - Clemens Ladisch USB generic driver fixes for broken SB Audigy 2 NX descriptors ALSA - fixed compilation [PATCH] SCSI Midlayer initiated START_UNIT The ipr driver that is currently out for review talks to disk array devices that require a START_UNIT prior to media ops, similar to normal scsi devices. However, any time the adapter gets reset, these devices end up needing another START_UNIT. This causes problems with the current error handling and these devices get taken offline when this occurs. Attached is a patch which will better handle these devices and issue a START_UNIT from the error handler when appropriate. Add SCSI transport attributes From: Martin Hicks Transport attributes are classes which can be attached to by a scsi driver to export (and later control) transport based properties. ia64: Fix staircase effect on Altix serial console Serial console output on Altix didn't convert NL to CRNL. [ACPI] comments [ACPI] global lock macro fixes (Paul Menage, Luming Yu) http://bugzilla.kernel.org/show_bug.cgi?id=1669 [PATCH] sort SCSI blacklist No changes whatsoever, just moves entries around to sort on vendor. I kept the multiple sections, and just sorted per section. [PATCH] USB 6-in-1 card reader blacklist addition Yes, someone really was stupid enough to make a device with such a dumb vendor name. In combination with the device name and rev number though, it shouldn't cause any false positives. Device in question is some no-name 6-in-1 usb card reader, whose slots all appear on different LUNs, requiring you fiddle with scis_mod's max_luns argument unless you have this patch. [PATCH] Add full complement of SPI transport attributes This patch does two things: 1. Add all the other ppr type transport attributes to the spi class 2. make period settable as the ppr/sdtr period, but display in ns for the user. [PATCH] gcc-3.5: acpi build fix Current gcc requires that both the declaration and the definition of functions describe the same register conventions. ALSA CVS update - Takashi Iwai ALSA Core - show the error message when the given card index is not available. ALSA CVS update - Takashi Iwai PPC Tumbler driver - fixed the mic input on snapper. ALSA CVS update - Jaroslav Kysela ICE1724 driver,ICE1712 driver Dirk Kalis Added num_total_adcs. Separated analog input / s/pdif input controls ALSA CVS update - Jaroslav Kysela ALSA<-OSS sequencer mpkelly - fixed channel settings for input events ALSA CVS update - Takashi Iwai Opti9xx drivers fixed the code with obsolete check_region(). ALSA CVS update - Takashi Iwai AC97 Codec Core,Intel8x0 driver,VIA82xx driver,CS46xx driver - fixed ALC100/P Master/PCM volume handling (h/w bug) - added ALC65x JACK quirk - disabled IC5 PERL mobo quirk - fixed Mic/Center sharing switch on ALC65x. - fixed Mic BIAS on ALC650. - added extra delay in the resume if needed. - renamed 'External Amplifier Power Down' to 'External Amplifier' - added a workaround for the reversed EAPD of cs46xx voyetra. ALSA CVS update - Takashi Iwai Intel8x0 driver fixed the interrupt problem with NForce(2). ALSA CVS update - Jaroslav Kysela PCM Midlevel Fix in playback_silence routine - don't silence whole buffer at start if samples are filled ALSA CVS update - Jaroslav Kysela ALSA Core PCM API is 2.0.6 ALSA CVS update - Takashi Iwai Memalloc module,PCM Midlevel,ALSA Core,CMI8330 driver,ES18xx driver Sound Scape driver,AD1816A driver,AD1848 driver,CS4231 driver ES1688 driver,GUS Library,Opti9xx drivers,SB16/AWE driver,SB8 driver ALS4000 driver,AZT3328 driver,BT87x driver,CMIPCI driver,CS4281 driver ENS1370/1+ driver,ES1938 driver,ES1968 driver,FM801 driver Intel8x0 driver,Maestro3 driver,RME32 driver,RME96 driver SonicVibes driver,VIA82xx driver,ALI5451 driver,CS46xx driver EMU10K1/EMU10K2 driver,ICE1712 driver,ICE1724 driver,KORG1212 driver MIXART driver,RME HDSP driver,RME9652 driver,Trident driver YMFPCI driver,Sound Core PDAudioCF driver,USB generic driver - clean up the DMA code again. now only struct device pointer is handled for every BUS type. the pointer must be given via the corresponding macro snd_dma_xxx_data(). - added the hack for dma_alloc_coherent() to accept dev = NULL for ISA buffers. - added the missing include files. ALSA CVS update - Takashi Iwai PARISC Harmony driver,SPARC AMD7930 driver,SPARC cs4231 driver fixed for the new DMA buffer handler. ALSA - fix compilation (header files) SCSI: Make SPI transport attributes mutable This adds the final missing piece to the transport attributes: A published API by which they can be set and retrieved (SPI attributes only). The sysfs field only appears writeable if the driver supplied a set method in the attribute template, so unsettable attributes show up as read only. The consequence now is that the spi transport attribute class is no longer a simple exported structure: the driver has to attach to the spi transport class at module initialisation. SCSI: implement transport attributes for 53c700 his patch moves the internal storage of the offset and period to the transport class and adds methods to set them (there's no need for a get method since the value in the transport class is exactly what the driver believes the transport agreement to be). You can see how this type of use of the transport class is supposed to work: the driver now has init and exit routines attaching and releasing the transport class (that's the reason for the Makefile perturbation because it has to init before its users). The correct attributes are set up in slave_configure() for negotiation and we now export the period and offset setting functions. ia64: Reserve 3 syscall numbers for Andi Kleen's NUMA interface. [PATCH] ia64: fix SN2 console driver to use console_initcall() fix Kconfig select problem with SCSI_SPI_ATTRS spotted by akpm (the select was misspelled in the SCSI_SIM710 Kconfig option) CONFIG_SCSI_AIC7XXX Kconfig bug From: Randy.Dunlap ,olh@suse.de make Adaptec AIC7xyx drivers depend on SCSI tristate (without this you can compile aic7xxx into the kernel when SCSI is only modular) [libata] Split up shared IO register locations into individual components Most ATA host controllers follow a standard layout for the ATA shadow registers, where command/status, error/feature, and devctl/altstatus share a single bus I/O address, because one register of each pair is read-only, and the other is write-only. On the Vitesse/Intel chip, all registers are given distinction bus I/O addresses, which necessitates changing the libata data structures to cope with this. This simply involves storing a few more bus addresses. [libata] misc fixes, and an export * export ata_host_intr * fix debugging printk * the write of Device Control register, which occurs between SATA phy reset and the read of device signature, apparently clobbers some of the register values (namely Error register) that are needed to determine whether or not the device passed diagnostics. Make this write follow the device diag checking and classification, instead. [libata] Add new driver for Vitesse VSC-7174. Driver core: make CONFIG_DEBUG_DRIVER implementation a whole lot cleaner [PATCH] I2C: Prevent i2c-dev oops with debug Looks like i2c-dev suffers the same problem with dev_dbg as i2c-core did twice. Here is a patch that fixed a oops I and another user were experiencing when running sensors-detect. [PATCH] I2C: Fix i2c_use_client() i2c_use_client() contains a bogosity. If i2c_inc_use_client() returns success, i2c_use_client() returns an error. If i2c_inc_use_client() fails, i2c_use_client() might succeed. Fix it so that (a) we get the correct sense between these two functions, and (b) propagate the error code from i2c_inc_use_client(), rather than making our own one up. [PATCH] I2C: sysfs interface update for w83627hf This patch updates the sysfs names of the w83627hf driver to match the new standard. The patch applies on top of one recently applied to your tree [1]. I have tested it using a w83627thf & the latest lm_sensors CVS. Please apply. [1] http://archives.andrew.net.au/lm-sensors/msg06746.html [PATCH] I2C: fix i2c-prosavage.c section usage prosavage_remove() is called during init, so it shouldn't be marked as exit code. (It matters when CONFIG_HOTPLUG=n.) [PATCH] I2C: fix i2c adapters class for now Please accept this temporary fix for the class issue. I'd like users of the i2c-ali1535, i2c-sis5595 and i2c-via driver to be able to use sensors as soon as possible, even if we have not yet determined what our policy WRT classes should be. Thanks. BTW, don't you think I2C_ADAP_CLASS_SMBUS is a misnomer? It's about hardware monitoring, not the SMBus protocol. I'd say I2C_ADAP_CLASS_HWMON or I2C_ADAP_CLASS_SENSORS would have been more appropriate (although it may be a bit late for a change...) [PATCH] I2c: Kconfig for non-sensors i2c chip drivers Quoting myself: > I think that it would make sense to have a specific menu entry for > these, separate from the sensors stuff. Looks like an easy thing to > do. Here is a proposed patch that does this. Comments welcome. Make the SCSI mempool allocations variable We still retain the default MAX_PHYS_SEGMENTS (which is currently 128). However, a mechanism for raising the limit correctly is added for people who wish to benchmark with larger sglists. A decision on how we present this to the user will be taken at a later time when more information about the usefulness of a higher limit is available MPT Fusion driver 3.01.00 update From: Moore, Eric Dean [PATCH] dc395x [1/5] - formatting cleanups Formatting cleanups. - Remove lots of comments that were out-of-date, wrong, or now described things that had been simplified that much that the comments were no longer needed. - Move the "static" definition from a line before functions onto the same line as the function itself. - Remove "return;" from the end of void functions. - Remove braces from around single instructions. - Merge the declaration and assignment to variables in places where the followed each other. - Remove braces from around single statement that obviously didn't need them. - Remove unused code and variables. - Fix debug statement in disconnect to not deref potentially null ptr. - Change a lot of short comments that take up three lines down to one line. [PATCH] dc395x [2/5] - sg list handling cleanups Cleanup a lot of the sg list handling to make it easier to follow in preparation for making this work on sparc64 (big-endian & 64bit). [PATCH] dc395x [3/5] - remove old debugging stuff Remove the special versions of kmalloc and kfree and remove the trace related stuff from the code. None of this is really needed and it is not really a nice implementation. This also removes all non 7-bit clean characters to help with David Eger's work to remove and/or convert them all to UTF8. [PATCH] dc395x [4/5] - debugging cleanup Clean up debugging statements. Work on making output consistent - use the same format each time the same bit of information is output and output the same bits of information in the same order each time it's output. Update a lot of the actual messages to include useful information and to explain better what some things are. Remove some old comments and commented out code. Remove some of the debugging states. Find few instance of spaces that should have been tabs and trailing white space and remove them. [PATCH] dc395x [5/5] - version update Move version information from .h to .c so that it's more likely to get updated as changes are made (the .h file is rarely modified). Update the version from 2.04 to 2.05 [PATCH] ia64: minor cleanups for SN2 console driver [PATCH] New SCSI host_byte status code. At present there is no way to retry, if allowed, a scsi command without decrementing the scsi command's retry count. Please find a patch attached, which is submitted, for adding a new host_byte status code called "DID_IMM_RETRY". This enables the Low Level SCSI drivers to initiate retry of a scsi command without decrementing the scsi commad's retry count. [PATCH] qlogicfas: use a static string as name [PATCH] qlogic_cs: don't release region [PATCH] qlogic_cs: use a static string as name [PATCH] qlogic_cs: use scsi_host_put [PATCH] qlogicfas: force can_queue [PATCH] qlogic_cs: use own detect and release functions [PATCH] qlogic_cs: don't call qlogic_release on fail [PATCH] qlogicfas: kill QL_USE_IRQ [PATCH] qlogicfas: use qlogicfas_name instead qinfo [PATCH] qlogicfas: begin to convert qlogicfas to new driver [PATCH] qlogicfas: disable irqs on exit [PATCH] qlogicfas: support multiple cards [PATCH] qlogicfas: move common definitions to qlogicfas.h [PATCH] qlogicfas: finish to convert to new scsi driver [PATCH] qlogic_cs: use own MODULE_ macros [PATCH] SCSI: megaraid /proc dir fix I know that LSI are working on an updated driver for 2.6, but would really like to see this bug fix placed in the main tree before then, whenever then happens to be. This patch fixes the problem of the /proc entries for this driver being created in the wrong location. [PATCH] buslogic init. section fix V: linux-264-rc2 D: BusLogic_AnnounceDriver() can be called after init; diffstat:= drivers/scsi/BusLogic.c | 2 +- 1 files changed, 1 insertion(+), 1 deletion(-) [PATCH] ia64: fix preempt bug in IA32 subsystem Without this patch, IA32 programs will stall when CONFIG_PREEMPT is enabled. [PATCH] ia64: Altix affinity fix [CRYPTO]: Fix arc4 test vector. [IPVS]: Fix typo in Config.in [PATCH] I2C: Cleanup fan_div in w83781d Here is a proposed patch that cleanups the fan_div code from w83781d. The original code was obviously taken from the 2.4 driver, but the way things were done in 2.4 do not make any sense anymore (because we now have a single value per interface file). Since fan divisor bits are spread over three different regs with various bitmask manipulations, I don't think it makes much sense to have a single function as we do in most other cases. Having three different functions makes more sense, although they are of course similar. The size increment is only 581 bytes, I don't think its a problem since it also makes the code much more efficient and readable too IMHO. I agree that the original code was working - it was simply doing far too much work each time one would want to change a fan divisor. Originally, I took a look at the code because I wanted to fix the fact that changing fan divisors breaks fan mins. But it looked like a good cleanup before doing that was required. Once this patch will have been accepted, I'll work on the other one. I've tested my patch successfully on a W83781D and an AS99127F rev.1. Note that this means that one part of the code wasn't tested, because these are the two chips of the family that do not support divisors greater than 8. If anyone can test it, please do. [IPV4]: Do not leak cork.opt in ip_push_pending_frames(). [PATCH] Patch to hook up PPP to simple class sysfs support * Hanna Linder (hannal@us.ibm.com) wrote: > + ppp_class = class_simple_create(THIS_MODULE, "ppp"); > + class_simple_device_add(ppp_class, MKDEV(PPP_MAJOR, 0), NULL, "ppp"); What happens if that class_simple_create() fails? Actually, class_simple_device_add could fail too, but doesn't seem anybody is checking for that. > err = devfs_mk_cdev(MKDEV(PPP_MAJOR, 0), > S_IFCHR|S_IRUSR|S_IWUSR, "ppp"); > - if (err) > + if (err) { > unregister_chrdev(PPP_MAJOR, "ppp"); > + class_simple_device_remove(MKDEV(PPP_MAJOR,0)); > + } need to destroy the class on error path to avoid leak. > @@ -2540,6 +2547,7 @@ static void __exit ppp_cleanup(void) > if (unregister_chrdev(PPP_MAJOR, "ppp") != 0) > printk(KERN_ERR "PPP: failed to unregister PPP device\n"); > devfs_remove("ppp"); > + class_simple_device_remove(MKDEV(PPP_MAJOR, 0)); ditto. this will leak and would cause oops on reload of module. something like below. [PATCH] class_simple clean up in lp Error condition isn't caught on class_simple_create, and parport_register_driver failure doesn't do proper cleanup. [PATCH] class_simple cleanup in input Doesn't catch error on class_simple_add, and existing error return paths forget to class_simple_destroy. [PATCH] class_simple cleanup in misc Error path doesn't class_simple_destroy. [PATCH] class_simple cleanup in sg The only spot that seems to care about class_simple_device_add possibly failing, but it gets the wrong error test. [IPV6]: Kill unused warnings in addrconf.c [PATCH] Add sysfs simple class support for netlink Patch adds sysfs simple class support for netlink character device (Major 36). Feedback appreciated. Thanks, [libata] make set_{pio,udma}mode hooks optional Only one driver actually uses them. [PATCH] G5 temperature control update This makes the temperature control code more robust, putting less pressure on i2c, and work around occasional misconfiguration of the ADC chips leading to incorrect temperature readings. Delete (void)func() casts considered cruft in Linux style. GCC's inability to warn when return values are ignored has conditioned Linux programmers into thinking that this is actually normal. delete some #define's -- suggested by Matt Wilcox [ACPI] fix printk and build warning from previous csets [PATCH] x86-64 merge for 2.6.4 The biggest new feature is fixed 32bit vsyscall (SYSCALL+SYSENTER) support, mostly from Jakub Jelinek. This increases 32bit syscall performance greatly (latency halved and better). The SYSENTER for Intel support required some infrastructure changes, but seems to work now too. The 64bit vsyscall vtime() just references xtime.tv_sec now. This should make it a lot faster too. A fix for some Intel IA32e systems. Also a few long standing bugs in NMI like exception handlers were fixed. And a lot of other bug fixes. Full changeLog: - Clean up 32bit address room limit handling, fix 3gb personality - Move memcpy_{from,to}io export to ksyms.c file. This seems to work around a toolchain bug (Andreas Gruenbacher) - Update defconfig - ACPI merges from i386 (SBF should work now, acpi=strict) - Implement mmconfig support based on i386 code (untested) - Fix i386/x86-64 pci source file sharing - Implement ptrace access for 32bit vsyscall page - Always initialize all 32bit SYSENTER/SYSCALL MSRs. - Export run time cache line size to generic kernel - Remove explicit CPUID in ia32 syscall code - Fill in most of boot_cpu_data early - Remove unused PER_LINUX32 setup - Fix syscall trace in fast 32bit calls (Suresh B. Siddha) - Tighten first line of the oops again. - Set up ptrace registers correctly for debug,ss,double fault exceptions - Fix 64bit bug in sys_time64 - Optimize time syscall/vsyscall to only read xtime - Fix csum_partial_copy_nocheck - Remove last traces of FPU emulation - Check properly for rescheduling in exceptions with own stack - Harden exception stack entries (#SS,#NMI,#MC,#DF,#DB) against bogus GS - Use an exception stack for machine checks - Handle TIF_SINGLESTEP properly in kernel exit - Add exception stack for debug handler - Disable X86_HT for Opteron optimized builds because it pulls in ACPI_BOOT - Fix CONFIG_ACPI_BOOT compilation without CONFIG_ACPI - Fix eflags handling in SYSENTER path (Jakub Jelinek) - Use atomic counter for enable/disable_hlt - Support 32bit SYSENTER vsyscall too (Jakub Jelinek) - Don't redefine Dprintk - Change some cpu/apic id arrays to char - Support arbitary cpu<->apicid in hard_smp_processor_id (Surresh B Sidda) - Move K8 erratum #100 workaround into slow path of page fault handler. - Fix 32bit cdrom direct access ioctls (Jens Axboe) - Enable 32bit vsyscalls by default - Fix 32bit vsyscalls (Jakub Jelinek) [PATCH] Fix a 64bit bug in kobject module request From Takashi Iwai kobj_lookup had a 64bit bug, which caused the request of a unknown character device to burn CPU instead of failing quickly. ia64: don't unmask iosapic interrupts by default In ia64 kernel, IOSAPIC's RTEs for PCI interrupts are unmasked at the boot time before installing device drivers. I think it is very dangerous. If some PCI devices without device driver generate interrupts, interrupts are generated repeatedly because these interrupt requests are never cleared. I think RTEs for PCI interrupts should be unmasked by device driver. This patch fixes the problem. [PATCH] Fix class_register() always returns 0 I noticed that the class_register() function in drivers/base/class.c always returns 0 and thus will never fail. Patch below inserts simple error checking to return any errors if they occur. Feedback welcome. Thanks, PPC32: Update the Motorola PowerPlus family support. From Randy Vinson Kobject: add decl_subsys_name() macro for users who want to set the subsystem name PCI Hotplug: use the new decl_subsys_name() macro instead of rolling our own. [PATCH] add sysfs simple class support for DRI char device Patch adds sysfs simple class support for DRI character device (Major 226). Also, adds some error checking. [PATCH] I2C: sensor chip driver refactoring This patch is a refactoring of some common code among all sensors chip drivers (except asb100, which was written this way to begin with.) It saves a handful of lines and ~100-300 bytes per module. It compiles ok. I've only tested it against one of the drivers, but the changes are similar across the board and quite mechanical. Please apply. [PATCH] ia64: Convert to use the generic drivers/Kconfig mechanism. [PATCH] ia64: fix misc. sn2 warnings This patch fixes a few warnings that have cropped up in the sn2 code: - hwgfs function prototype mismatch - pconn uninitialized in pciio.c - printk formatting fixes in pcibr_dvr.c - kill volatile qualifier in pcibr_intr.c [PATCH] fix PPC64 iSeries virtual console devices While playing with udev, I discovered that the virtual console devices on iSeries had there minor numbers off by one i.e. /dev/tty1 was minor 2! This fixes it. [PATCH] user data -> request mapping This patch allows you to map a request with user data for io, similarly to what you can do with bio_map_user() already to a bio. However, this goes one step further and populates the request so the user only has to fill in the cdb (almost) and put it on the queue for execution. Patch converts sg_io() to use it, next patch I'll send adapts cdrom layer to use it for zero copy cdda dma extraction. [PATCH] CDROMREADAUDIO dma support This small patch builds on top of the blk_rq_map_user() patch just sent, and enables us to easily support DMA for CDROMREADAUDIO cdda extraction. It's quite amazing how much cool stuff you can with the new block layer :-) Patch has intelligent fall back from multi frame dma to single frame dma, and further to old-style pio ripping in case of hardware problems. [PATCH] fix ppc64 in kernel syscalls Thanks to some great debugging work by Olaf Hering and Marcus Meissner it has been noticed that the current ppc64 syscall code is corrupting 4 bytes past errno. Why we even bothered to set errno beats me, its unusable in the kernel. Since we had to reinstate the inline syscall code we can go back to using it for those few syscalls that we call. Especially now with Randy's syscall prototype cleanup we should be calling them directly but we can do that sometime later. [PATCH] ppc32: Fix G5 config space access lockup Fix a typo in the code that prevents lockup on config space access to sleeping devices on ppc32/G5. Please apply. [PATCH] print kernel version in oops messages From: Arjan van de Ven Unfortunatly a large portion of the oops reports lack the basic information about what kernel version the oops is for; it's trivial to just print this in the oops as well to improve the usefulness of bugreports... [PATCH] ppc64: fix initialisation of NUMA arrays From: Anton Blanchard We were hitting problems on machines with cpu_possible != cpu_online when NUMA was enabled. The debug checks would trip during scheduler init because we iterate through all possible cpus whereas we only set up NUMA information for online cpus. Longer term we should have a cpu_up hook which sets up its NUMA information but for now we initalise all possible cpus and memory to node 0. [PATCH] Clean up sys_ioperm stubs From: Brian Gerst Remove stubs for sys_ioperm for non-x86 arches, using sys_ni_syscall instead where applicable. Support for sys_ioperm is unconditionally no for non-x86 arches. [PATCH] readdir() cleanups From: cramfs and freevxfs explicitly mark themselves readonly (as other r/o fs do). afs marked noatime (ACKed by maintainer) filesystems that do not do update_atime() in their ->readdir() had been explicitly marked nodiratime. NOTE: cifs, coda and ncpfs almost certainly need full noatime as we currently have in nfs and afs. update_atime() call shifted to callers of ->readdir() and out of ->readdir() instances. Bugs caught: dcache_readdir() updated atime only if it reached EOF. bfs_readdir() - ditto. qnx4_readdir() - ditto. [PATCH] adaptive lazy readahead From: Suparna Bhattacharya From: Ram Pai Pipelined readahead behaviour is suitable for sequential reads, but not for large random reads (typical of database workloads), where lazy readahead provides a big performance boost. One option (suggested by Andrew Morton) would be to have the application pass hints to turn off readahead by setting the readahead window to zero using posix_fadvise64(POSIX_FADV_RANDOM), and to special-case that in do_generic_mapping_read to completely bypass the readahead logic and instead read in all the pages needed directly. This was the idea I started with. But then I thought, we can do a still better job ? How about adapting the readahead algorithm to lazy-read or non-lazy-read based on the past i/o patterns ? The overall idea is to keep track of average number of contiguous pages accessed in a file. If the average at any given time is above ra->pages the pattern is sequential. If not the pattern is random. If pattern is sequential do non-lazy-readahead( read as soon as the first page in the active window is touched) else do lazy-readahead. I have studied the behaviour of this patch using my user-level simulator. It adapts pretty well. Note from Suparna: This appears to bring streaming AIO read performance for large (64KB) random AIO reads back to sane values (since the lazy readahead backout in the mainline). [PATCH] read-only support for UFS2 From: Niraj Kumar This patch adds read-only support for ufs2 (used in FreeBSD 5.x) variant of ufs filesystem. For filesystem specific tools, see http://ufs-linux.sourceforge.com . [PATCH] fb_console_init fix From: James Simmons This patch fixes fb_console_init from being called twice. I still need to fix set_con2fb but this helps but this is still important to get in. [PATCH] time interpolator fix From: john stultz In developing the ia64-cyclone patch, which implements a cyclone based time interpolator, I found the following bug which could cause time inconsistencies. In update_wall_time_one_tick(), which is called each timer interrupt, we call time_interpolator_update(delta_nsec) where delta_nsec is approximately NSEC_PER_SEC/HZ. This directly correlates with the changes to xtime which occurs in update_wall_time_one_tick(). However in update_wall_time(), on a second overflow, we again call time_interpolator_update(NSEC_PER_SEC). However while the components of xtime are being changed, the overall value of xtime does not (nsec is decremented NSEC_PER_SEC and sec is incremented). Thus this call to time_interpolator_update is incorrect. This patch removes the incorrect call to time_interpolator_update and was found to resolve the time inconsistencies I had seen while developing the ia64-cyclone patch. [PATCH] teach /proc/kmsg about O_NONBLOCK If there's nothing available and the file is O_NONBLOCK, return -EAGAIN. This is a bit grubby - really we should push the file* down into do_syslog() and handle it inside the spinlock. [PATCH] remove __io_virt_debug From: Brian Gerst Drivers should all be converted to use ioremap() or isa_*() by now. [PATCH] genrtc: cleanups From: "Randy.Dunlap" From: Luiz Fernando Capitulino remove ifdef/endif in rtc_generic_init(). use returned error code; [PATCH] i386 very early memory detection cleanup patch From: "H. Peter Anvin" This patch cleans up the very early memory setup on the i386 platform. In particular, it removes the hard-coded 8 MB limit completely by dynamically creating the early-boot pagetables rather than having them hard coded. While I was at it, I changed head.S so that it always sets up a local GDT; this means among other things that SMP and VISWS are no longer special cases, and is conceptually cleaner to boot. The VISWS people have confirmed it works on VISWS. It also uses a separate entrypoint for non-boot processors since this is completely kernel-internal anyway. This eliminates the need to set %bx on boot. (If you think this is a bad idea I can eliminate this change; it just seemed cleaner to me to do it this way.) Additionally, zero bss with rep;stosl rather that rep;stosb. [PATCH] Allow X86_MCE_NONFATAL to be a module From: Herbert Xu By allowing X86_MCE_NONFATAL to be a module, it can be included in distribution kernels without upsetting those with strange hardware. [PATCH] dm: endio method From: Joe Thornber Add an endio method to targets. This method is allowed to request another shot at failed ios (think multipath). Context can be passed between the map method and the endio method. [PATCH] dm: list_for_each_entry audit From: Joe Thornber Audit for list_for_each_*entry* [PATCH] dm: default queue limits From: Joe Thornber Fill in missing queue limitations when table is complete instead of enforcing the "default" limits on every dm device. Problem noticed by Mike Christie. [Christophe Saout] [PATCH] dm: list targets cmd From: Joe Thornber List targets ioctl. [Patrick Caulfield] [PATCH] dm: stripe width fix dm-stripe.c: The stripe width must be at least the page size. [PATCH] selinux: clean up binary mount data From: James Morris selinux is currently inspecting the filesystem name ("nfs" vs "coda" vs watever) to work out whether it needs to hanbdle binary mount data. Eliminate all that by adding a flag to file_system_type.fs_flags. [PATCH] UDF filesystem update From: Ben Fennema - added udf 2.5 #defines - fixed prealloc discard race - fixed several bugs in inode_getblk - added S_IFSOCK support - fix unicode encoding bug - change partition allocation from kmalloc to vmalloc for large allocations [PATCH] kbuild: Remove CFLAGS assignment in i386/mach-*/Makefile From: Sam Ravnborg The EXTRA_CFLAGS assignments in the following files are a left-over from the early 2.5 days where the source was not compiled from the root of the source tree. Removing these wrong assignments fixes http://bugme.osdl.org/show_bug.cgi?id=2210 A script named 'kernel' in the .. directory no longer halt compilation. [PATCH] NUMA-aware zonelist builder From: The attached patch is NUMA-aware zonelist builder patch, which sorts zonelist in the order that near-node first, far-node last. In lse-tech and linux-ia64, where most of NUMA people resides, no objections are raised so far. The patch adds NUMA-specific version of build_zonelists which calls find_next_best_node to select the next-nearest node to add to zonelist. The patch has no effect on flat NUMA platform. [PATCH] Redundant unplug_timer deletion From: "Chen, Kenneth W" The only path to get to del_timer call in __generic_unplug_device() is when blk_remove_plug() returns 1, and in that case it already removed the unplug_timer. Patch to remove this redundant call. [PATCH] compiler.h scoping fixes From: Ville Nuorvala There are a few kernel-only things in compiler.h which should have been placed inside __KERNEL__. [PATCH] Fix elf mapping of the zero page From: William Lee Irwin III Using PAGE_SIZE rather than 4096 so that mmap() granularity is honored by whatever non-i386 architectures use MMAP_PAGE_ZERO. [PATCH] kbuild: Cause `make clean' to remove more files From: Sam Ravnborg Make the difference between 'make clean' and 'make distclean/mrproper' more explicit. make clean now removes all generated files except .config* and .version. The result is much easier to understand now. make clean deletes all generated files (except .config* and .version). make mrproper deletes configuration and all temporary files left by patch, editors and the like. Example output: > make mrproper CLEAN init CLEAN usr CLEAN scripts/kconfig CLEAN scripts CLEAN .tmp_versions include/config CLEAN include/asm-i386/asm_offsets.h include/linux/autoconf.h include/linux/version.h include/asm .tmp_versions CLEAN .version .config Form the list of files/directories deleted during make clean, removed all references that is no longer relevant for the current kernel. [PATCH] LOOP_CHANGE_FD ioctl From: Arjan van de Ven The patch below (written by Al Viro) solves a nasty chicken-and-egg issue for operating system installers (well at least anaconda but the problem domain is not exclusive to that) The basic problem is this: - The small first stage installer locates the image file of the second stage installer (which has X and all the graphical stuff); this image can be on the same CD, but it can come via NFS, http or ftp or ... as well. - The first stage installer loop-back mounts this image and gives control to the second stage installer by calling some binary there. - The graphical installer then asks the user all those questions and starts installing packages. Again the packages can come from the CD but also from NFS or http or ... Now in case of a CD install, once all requested packages from the first CD are installed, the installer wants to unmount and eject the CD and prompt the user to put CD 2 in....... EXCEPT that the unmount can't work since the installer is actually running from a loopback mount of this cd. The solution is a "LOOP_CHANGE_FD" ioctl, where basically the installer copies the image to the harddisk (which can only be done late since only late the target harddisk is mkfs'd) and then magically switches the backing store FD from underneath the loop device to the one on the target harddisk (and thus unbusying the CD mount). This is obviously only allowed if the size of the new image is identical and if the loop image is read-only in the first place. It's the responsibility of root to make sure the contents is the same (but that's of the give-root-enough-rope kind) [PATCH] loop setup race fix From: Chris Mason There's a race in loopback setup, it's easiest to trigger with one or more procs doing loopback mounts at the same time. The problem is that fs/block_dev.c:do_open() only calls bdev_set_size on the first open. Picture two procs: proc1: mount -o loop file1 mnt1 proc2: mount -o loop file2 mnt2 proc1 proc2 open /dev/loop0 # bd_openers now 1 do_open bd_set_size(bdev, 0) # loop unbound, so bdev size is 0 open /dev/loop0 # bd_openers now 2 loop_set_fd # disk capacity now correct, but # bdev not updated mount /dev/loop0 /mnt do_open Because bd_openers != 0 for the last do_open, bd_set_size is not called again and a size of 0 is used. This eventually leads to an oops when the loop device is unmounted, because fsync_bdev calls block_write_full_page who decides every page on the block device is outside i_size and unmaps them. When ext2 or reiserfs try to sync a metadata buffer, we get an oops on because the buffers are no longer mapped. The patch below changes loop_set_fd and loop_clr_fd to also manipulate the size of the block device, which fixes things for me. [PATCH] kbuild: fix usage with directories containing '.o' From: Sam Ravnborg From: Daniel Mack , me modpost unconditionally searched for ".o" assuming this is always the suffix of the module. This fails in two cases: a) when building external modules where any directory include ".o" in the name. One example is a directory named: .../cvs.alsa.org/... b) when someone names a kernel directory so it contains ".o". One example is drivers/scsi/aic.ok/... case b) was triggered by renaming the directory for aic7xxx, and modifying Makefile and Kconfig. This caused make modules to fail. [PATCH] Remove unneeded unlock in ipc/sem.c From: Manfred Spraul sem_revalidate checks that a semaphore array didn't disappear while the code was running without the semaphore array spinlock. If the array disappeared, then it will return without holding a lock. find_undo calls sem_revalidate and then sem_unlock, even if sem_revalidate failed. The sem_unlock call must be removed. Mingming Cao reported a spinlock deadlock with sysv semaphores. A superflous unlock doesn't explain the deadlock, but it's obviously a bug. [PATCH] /proc data corruption check From: Arjan van de Ven If someone removes a /proc directory which still has subdirectories it will lead to very nasty things (dentries remaining on hash chains etc etc etc). The BUG_ON in the patch below will catch this nasty situation. [PATCH] Enable i810 fb on x86-64 From: Andi Kleen i810fb most likely is needed on x86-64 too because there are Intel chipsets for it now. So far it only linked on i386, fix this. [PATCH] Remove arbitrary #acl entries limits on ext[23] when reading From: Andreas Gruenbacher Remove the arbitrary limit of 32 ACL entries on ext[23] when reading from disk. This change is backward compatible; we need to have this change in to be able to also allow writing big ACLs. The second patch that removes the ACL entry limit for writes is not included. I don't want to push that patch now, because large ACLs would cause 2.4 and current 2.6 kernels to fail. My plan is to remove the second limit later, in a half-year or year or so. [PATCH] watchdog: moduleparam-patches From: Wim Van Sebroeck Convert last set of watchdog drivers to new moduleparam system. [PATCH] AMD ELAN Kconfig fix From: Adrian Bunk - remove an MELAN entry that was forgotten in the i386 processor selection menu - s/CONFIG_MELAN/CONFIG_X86_ELAN/ was missing in module.h [PATCH] fadvise(POSIX_FADV_DONTNEED) fixups From: WU Fengguang - In sys_fadvise64_64(): if the start and/or end offsets do not fall on page boundaries, preserve the partial pages. The thinking here is that it is better to preserve needed memory than to not shoot down unneeded memory. - In invalidate_mapping_pages(): we were invalidating an entire pagevec's worth of pages each time around, even if that went beyond the part of the file which the caller asked to be invalidated. Fix that up. [PATCH] Fix and harden validate_mm From: Andi Kleen I was debugging some code that corrupted the vma rb lists and for that I fixed validate_mm to not be recursive and do some more checks. It's slower now, but that shouldn't be a problem. Also make it non static to allow easier checks elsewhere. [PATCH] current_is_keventd() speedup From: Srivatsa Vaddagiri current_is_keventd() doesn't need to search across all the CPUs to identify itself. [PATCH] Fix rootfs on ramdisk From: vda Add a missing test for the "root=/dev/ram" kernel boot option. It's just an alias for /dev/ram0, but it worked in 2.4... [PATCH] Fix reading the last block on a bdev From: Chris Mason This patch fixes a problem we're hitting on ia64 with page sizes > 4k. When the page size is greater than the block size, and parts of the page fall past the end of the device, readpage will fail because blkdev_get_block returns -EIO for blocks past i_size. The attached patch changes blkdev_get_block to return holes when reading past the end of the device, which allows us to read that last valid 4k block and then fill the rest of the page with zeros. Writes will still fail with -EIO. [PATCH] wavfront.c needs syscalls.h sound/oss/wavfront.c: In function `wavefront_download_firmware': sound/oss/wavfront.c:2524: warning: implicit declaration of function `sys_open' sound/oss/wavfront.c:2533: warning: implicit declaration of function `sys_read' sound/oss/wavfront.c:2582: warning: implicit declaration of function `sys_close [PATCH] EDD: Get Legacy Parameters From: Matt Domsch Patch below from Patrick J. LoPresti and myself. Patrick describes: Why this patch? The problem is that the legacy BIOS interface (INT13/AH=3D08) for querying the disk geometry returns different values than the extended INT13 interface which the EDD code currently uses. This is because the legacy interface only provides a 10-bit cylinder field, so modern BIOSes "lie" about the head/sector counts in order to make more of the disk visible within the first 1024 cylinders. Many non-Linux applications, including the stock Windows boot loader, DOS fdisk, etc., rely upon the legacy interface and geometry. So it is useful to be able to obtain the legacy values from a running Linux kernel. What this patch does is to add new entries under /sys/firmware/edd/int13_devXX named "legacy_cylinders", "legacy_heads", and "legacy_sectors". These provide the geometry given by the legacy INT13/AH=3D08 BIOS interface, just like the current "default_cylinders" etc. provide the the geometry given by the INT13/AH=3D48 interface. Without this patch, I cannot use Linux to partition a drive and install Windows, which happens to be my application. - Pat http://unattended.sourceforge.net/ In addition, this adds two buggy BIOS workarounds in the EDD int13 calls as suggested by Ralf Brown's interrupt list. I'm also interested in moving this code out of arch/i386/kernel/edd.c and include/asm-i386/edd.h, as I believe it is applicable on x86-64 as well. However, there's no good place under drivers/ to put edd.c when it's not tied to a bus, but to several CPU architectures and their firmwares... Maybe a new directory drivers/firmware? [PATCH] cciss: init section fix From: "Randy.Dunlap" cciss_scsi_detect() can be called after init (for TAPE support). [PATCH] add nowarn to a few pte chain allocators From: Arjan van de Ven Several of the pte_chain_alloc() allocators that use GFP_ATOMIC have a fallback for failure that sleeps; they thus need to not warn on failure.. Seen during a big fork on a busy system. [PATCH] Disable Macintosh device drivers for all but PPC || MAC From: Marc-Christian Petersen The attached patch is needed to stop showing us "Macintosh device drivers" for all architectures via menuconfig || xconfig || gconfig. It's only necessary for PPC and/or MAC. ACKed by benh. [PATCH] Applicom warning From: Geert Uytterhoeven Add missing include (needed for struct inode) [PATCH] Fix CONFIG_NVRAM dependencies From: Geert Uytterhoeven Make CONFIG_NVRAM depend on the prerequisites that are explicitly checked for in drivers/char/nvram.c, or on CONFIG_GENERIC_NVRAM (for PPC). [PATCH] fix raid0 readahead size From: Arjan van de Ven Readahead of raid0 was suboptimal; it read only 1 stride ahead. The problem with this is that while it will keep all spindles busy, it will not actually manage to make larger IO's, eg each disk would just do the chunk size IO. Doing at least 2 chunks is more than appropriate so that each spindle will get a chance to merge IO's. (Neil fixed raid6 and raid6 too) [PATCH] Fix NULL pointer dereference in blkmtd.c From: Michel Marti The blkmtd driver oopses in add_device(). The following trivial patch fixes this. [PATCH] fbdev: monitor detection fixes From: James Simmons , Kronos Various fixes and enhancements to the monitor hardware detection code. The only driver that uses it is the radeon driver. Old EDID parsing code was very verbose, half of the patch address this (ie. print lots of stuff iff DEBUG). The other big change is the FB_MODE_IS_* stuff: we really need a way to know the origin of a video mode. In this way we can select video mode that comes from EDID instead of VESA or GTF. Drivers other than radeonfb won't be affected because they cannot (yet) get EDID from the monitor and don't use EDID related code. [PATCH] m68k: __test_and_set_bit() From: Geert Uytterhoeven Add missing implementation for non-atomic __test_and_set_bit() [PATCH] m68k: Amiga Framemaster II fb sysfsification From: Geert Uytterhoeven Amiga Framemaster II fb: Add sysfs support (from James Simmons) [PATCH] Apollo fb sysfsification From: Geert Uytterhoeven Apollo fb: Add sysfs support (from James Simmons) [PATCH] m68k: Macintosh IDE fixes From: Geert Uytterhoeven Mac IDE: Make sure the core IDE driver doesn't try to request the MMIO ports a second time, since this will fail. [PATCH] m68k: interrupt management cleanups From: Geert Uytterhoeven M68k interrupt management: rename routines to not confuse them with syscalls - sys_{request,free}_irq() -> cpu_{request,free}_irq() - q40_sys_default_handler[] -> q40_default_handler - sys_default_handler() -> default_handler() [PATCH] Add barriers to avoid race in mempool_alloc/free From: Chris Mason mempool_alloc() and mempool_free() check pool->curr_nr without any locks held. This can lead to skipping a wakeup when there are people waiting, and sleeping when there are free elements in the pool. I can't trigger this reliably, but sooner or later someone on ppc is probably going to hit it. [PATCH] synclinkmp.c update From: Paul Fulghum Patch for synclinkmp.c * Track driver API changes * Remove cast (kernel janitor) * Replace page_free call with kfree (to match kmalloc allocation) [PATCH] synclink_cs.c update From: Paul Fulghum * Track driver API changes * Remove cast (kernel janitor) [PATCH] synclink.c update From: Paul Fulghum * track driver API changes * remove cast (kernel janitor) [PATCH] vm: per-zone vmscan instrumentation To check on zone balancing, split the /proc/vmstat:pgsteal, pgreclaim pgalloc and pgscan stats into per-zone counters. Additionally, split the pgscan stats into pgscan_direct and pgscan_kswapd to see who's doing how much scanning. And add a metric for the number of slab objects which were scanned. [PATCH] return remaining jiffies from blk_congestion_wait() Teach blk_congestion_wait() to return the number of jiffies remaining. This is for debug, but it is also nicely consistent. [PATCH] Narrow blk_congestion_wait races From: Nick Piggin The addition of the smp_mb and the other change is to try to close the window for races a bit. Obviously they can still happen, it's a racy interface and it doesn't matter much. [PATCH] mm/vmscan.c: remove unused priority argument. From: Nikita Danilov Now that decision to reclaim mapped memory is taken on the basis of zone->prev_priority, priority argument is no longer needed. [PATCH] kswapd throttling fixes The logic in balance_pgdat() is all bollixed up. - the incoming arg `nr_pages' should be used to determine if we're being asked to free a specific number of pages, not `to_free'. - local variable `to_free' is not appropriate for the determination of whether we failed to bring all zones to appropriate free pages levels. Fix this by correctly calculating `all_zones_ok' and then use all_zones_ok to determine whether we need to throttle kswapd. So the logic now is: for (increasing priority) { all_zones_ok = 1; for (all zones) { to_reclaim = number of pages to try to reclaim from this zone; max_scan = number of pages to scan in this pass (gets larger as `priority' decreases) /* * set `reclaimed' to the number of pages which were * actually freed up */ reclaimed = scan(max_scan pages); reclaimed += shrink_slab(); to_free -= reclaimed; /* for the `nr_pages>0' case */ /* * If this scan failed to reclaim `to_reclaim' or more * pages, we're getting into trouble. Need to scan * some more, and throttle kswapd. Note that this * zone may now have sufficient free pages due to * freeing activity by some other process. That's * OK - we'll pick that info up on the next pass * through the loop. */ if (reclaimed < to_reclaim) all_zones_ok = 0; } if (to_free > 0) continue; /* swsusp: need to do more work */ if (all_zones_ok) break; /* kswapd is done */ /* * OK, kswapd is getting into trouble. Take a nap, then take * another pass across the zones. */ blk_congestion_wait(); } [PATCH] vmscan: preserve page referenced info in refill_inactive() From: Nick Piggin If refill_inactive_zone() is running in its dont-reclaim-mapped-memory mode we are tossing away the referenced infomation on active mapped pages. So put that info back if we're not going to deactivate the page. [PATCH] shrink_slab: math precision fix From: Nick Piggin In shrink_slab(), do the multiply before the divide to avoid losing precision. [PATCH] vm: shrink slab evenly in try_to_free_pages() From: Nick Piggin In try_to_free_pages(), put even pressure on the slab even if we have reclaimed enough pages from the LRU. [PATCH] vmscan: fix calculation of number of pages scanned From: Nick Piggin The logic which calculates the numberof pages which were scanned is mucked up. Fix. [PATCH] vm: scan slab in response to highmem scanning The patch which went in six months or so back which said "only reclaim slab if we're scanning lowmem pagecache" was wrong. I must have been asleep at the time. We do need to scan slab in response to highmem page reclaim as well. Because all the math is based around the total amount of memory in the machine, and we know that if we're performing highmem page reclaim then the lower zones have no free memory. [PATCH] vmscan: zone balancing fix We currently have a problem with the balancing of reclaim between zones: much more reclaim happens against highmem than against lowmem. This patch partially fixes this by changing the direct reclaim path so it does not bale out of the zone walk after having reclaimed sufficient pages from highmem: go on to reclaim from lowmem regardless of how many pages we reclaimed from lowmem. [PATCH] vmscan: drive everything via nr_to_scan Page reclaim is currently a bit schitzo: sometimes we say "go and scan this many pages and tell me how many pages were freed" and at other times we say "go and scan this many pages, but stop if you freed this many". It makes the logic harder to control and to understand. This patch coverts everything into the "go and scan this many pages and tell me how many pages were freed" model. It doesn't seem to affect performance much either way. [PATCH] Balance inter-zone scan rates When page reclaim is working out how many pages to san in a zone (max-scan) it presently rounds that number up if it looks too small - for work batching. Problem is, this can result in excessive scanning against small zones which have few inactive pages. So remove it. Not that it is possible for max_scan to be zero. That's OK - it'll become non-zero as the priority increases. [PATCH] vmscan: avoid bogus throttling - If max_scan evaluates to zero due to a very small inactive list and high `priority' numbers, we don't want to thrlttle yet. - In balance_pgdat(), we may end up not scanning any pages because all zones happened to be above pages_high. Avoid throttling in this case too. [PATCH] kswapd: avoid unnecessary reclaiming from higher zones Currently kswapd walks across all zones in dma->normal->highmem order, performing proportional scanning until all zones are OK. This means that pressure against ZONE_NORMAL causes unnecessary reclaim of ZONE_HIGHMEM. To fix that up we change kswapd so that it walks the zones in the high->normal->dma direction, skipping zones which are OK. Once it encounters a zone which needs some reclaim kswapd will perform proportional scanning against that zone as well as all the succeeding lower zones. We scan the lower zones even if they have sufficient free pages. This is because a) the lower zone may be above pages_high, but because of the incremental min, the lower zone may still not be eligible for allocations. That's bad because cache in that lower zone will then not be scanned at the correct rate. b) pages in this lower zone are usable for allocations against the higher zone. So we do want to san all the relevant zones at an equal rate. [PATCH] kswapd: fix lumpy page reclaim As kswapd is now scanning zones in the highmem->normal->dma direction it can get into competition with the page allocator: kswapd keep on trying to free pages from highmem, then kswapd moves onto lowmem. By the time kswapd has done proportional scanning in lowmem, someone has come in and allocated a few pages from highmem. So kswapd goes back and frees some highmem, then some lowmem again. But nobody has allocated any lowmem yet. So we keep on and on scanning lowmem in response to highmem page allocations. With a simple `dd' on a 1G box we get: r b swpd free buff cache si so bi bo in cs us sy wa id 0 3 0 59340 4628 922348 0 0 4 28188 1072 808 0 10 46 44 0 3 0 29932 4660 951760 0 0 0 30752 1078 441 1 6 30 64 0 3 0 57568 4556 924052 0 0 0 30748 1075 478 0 8 43 49 0 3 0 29664 4584 952176 0 0 0 30752 1075 472 0 6 34 60 0 3 0 5304 4620 976280 0 0 4 40484 1073 456 1 7 52 41 0 3 0 104856 4508 877112 0 0 0 18452 1074 97 0 7 67 26 0 3 0 70768 4540 911488 0 0 0 35876 1078 746 0 7 34 59 1 2 0 42544 4568 939680 0 0 0 21524 1073 556 0 5 43 51 0 3 0 5520 4608 976428 0 0 4 37924 1076 836 0 7 41 51 0 2 0 4848 4632 976812 0 0 32 12308 1092 94 0 1 33 66 Simple fix: go back to scanning the zones in the dma->normal->highmem direction so we meet the page allocator in the middle somewhere. r b swpd free buff cache si so bi bo in cs us sy wa id 1 3 0 5152 3468 976548 0 0 4 37924 1071 650 0 8 64 28 1 2 0 4888 3496 976588 0 0 0 23576 1075 726 0 6 66 27 0 3 0 5336 3532 976348 0 0 0 31264 1072 708 0 8 60 32 0 3 0 6168 3560 975504 0 0 0 40992 1072 683 0 6 63 31 0 3 0 4560 3580 976844 0 0 0 18448 1073 233 0 4 59 37 0 3 0 5840 3624 975712 0 0 4 26660 1072 800 1 8 46 45 0 3 0 4816 3648 976640 0 0 0 40992 1073 526 0 6 47 47 0 3 0 5456 3672 976072 0 0 0 19984 1070 320 0 5 60 35 [PATCH] fix the kswapd zone scanning algorithm This removes a vestige of the old algorithm. We don't want to skip zones if all_zones_ok is true: we've already precalculated which zones need scanning and this just stops us from ever performing kswapd reclaim from the DMA zone. [PATCH] vmscan: less throttling of page allocators and kswapd This is just a random unsubstantiated tuning tweak: don't immediately throttle page allocators and kwapd when the going is getting heavier: scan a bit more of the LRU before throttling. [PATCH] vmscan: batch up inactive list scanning work From: Nick Piggin Use a "refill_counter" for inactive list scanning, similar to the one used for active list scanning. This batches up scanning now that we precisely balance ratios, and don't round up the amount to be done. No observed benefits, but I imagine it would lower the acquisition frequency of the lru locks in some cases, and make codepaths more efficient in general due to cache niceness. [PATCH] fix vm-batch-inactive-scanning.patch - prevent nr_scan_inactive from going negative - compare `count' with SWAP_CLUSTER_MAX, not `max_scan' - Use ">= SWAP_CLUSTER_MAX", not "> SWAP_CLUSTER_MAX". [PATCH] vm: balance inactive zone refill rates The current refill logic in refill_inactive_zone() takes an arbitrarily large number of pages and chops it down to SWAP_CLUSTER_MAX*4, regardless of the size of the zone. This has the effect of reducing the amount of refilling of large zones proportionately much more than of small zones. We made this change in may 2003 and I'm damned if I remember why. let's put it back so we don't truncate the refill count and see what happens. [PATCH] vmscan: add lru_to_page() helper From: Nick Piggin Add a little helper macro for a common list extraction operation in vmscan.c [PATCH] slab: avoid higher-order allocations From: Manfred Spraul At present slab is using 2-order allocations for the size-2048 cache. Of course, this can affect networking quite seriously. The patch ensures that slab will never use more than a 1-order allocation for objects which have a size of less than 2*PAGE_SIZE. Revert attribute_used changes in module.h. They were wrong. Cset exclude: akpm@osdl.org|ChangeSet|20040312161945|47751 [wireless] Add new Prism54 wireless driver. [wireless prism54] remove WIRELESS_EXT ifdefs [PATCH] drivers_net_wireless_airo.c '< 0' comparison make sense From: The sense of the comparison was signed, but the code was testing an unsigned variable for less-than-zero. [PATCH] forcedeth update Andrew de Quincey added wol support to forcedeth. The patch also renames additional function, to help analyzing backtraces. [PATCH] eepro init section usage eepro_print_info() can be __init. [PATCH] smctr: fix init section usage smctr_chk_mca() can be __init. [PATCH] pcnet32 correct names for changes This patch corrects the names of contributors of changes to the pcnet32 driver. [PATCH] netdevice.h add netif_msg_init helper This patch adds a helper function to initialize the debug bit mask for use with netif_msg_*. When the debug_value is out of range it returns the default_msg_enable_bits. Tested IA32. [PATCH] xirc2ps ethtool fix patch below adds bus_info for xirc2ps_cs; anaconda depends on this info to be there. [PATCH] use netdev_priv() in appletalk & fc [PATCH] use netdev_priv() in /hamradio/ [PATCH] use netdev_priv() in 3com net drivers [PATCH] use netdev_priv() in net/ lance drivers [PATCH] use netdev_priv() in net/arm drivers [PATCH] use netdev_priv() in net/ intel drivers [PATCH] use netdev_priv() in net/pcmcia/ drivers [PATCH] use netdev_priv() in net/tulip drivers [PATCH] use netdev_priv() in net/tokenring/ drivers [PATCH] use netdev_priv() in net/wireless/ drivers [PATCH] use netdev_priv() in tap/tun/plip/loop/skel [PATCH] use netdev_priv() in fusion/mptlan [PATCH] use netdev_priv() in net/wan drivers [PATCH] use netdev_priv() in drivers/net/ (others) [PATCH] 8139too assertions Add Promise SX8 (carmel) block driver. PPC32: Fix include/asm-ppc/dma-mapping.h for the !CONFIG_PCI case. PPC32: consistent_free only takes one arguement. MPT Fusion driver 3.01.01 update From: Moore, Eric Dean This is an update for the MPT Fusion drivers 2.6 kernel. Version 3.01.01. This is a fix for poor performance in RAID Volumes. The dvStatus was being cleared for hidden physical disks when mptscsih_slave_destroy is called. Also, I have fixed the warning comming from mptscsih_reset_timeouts. [PATCH] aacraid driver patch I submitted a patch last month for the aacraid driver's reset handler. I left out setting function pointers in the adapter_ops structure for the adapter_check_health element. [PATCH] sym2 2.1.18i - Correct a typo "mvram" -> "nvram". - Re-do the PQS/PDS support which I'd #if 0 out. Should even work on multiple-domain boxes now ;-) - Move all the nvram definitions to sym_nvram.h (from Gerard's 2.1.19-pre3) - hcb_p -> struct sym_hcb * - sdev_p -> struct sym_device * - Delete a lot of unused macros from sym_misc.h - Move READ_BARRIER and WRITE_BARRIER definitions to sym_glue.h - SYM_CONF_NVRAM_WRITE_SUPPORT (from Gerard's 2.1.19-pre3). Not enabled yet. - Fix some -W warnings (some courtesy of Adrian Bunk). [PATCH] add adapter support to aacraid driver (update) Mark S. said that there was another adapter added, and that they changed the names of some boards. Here is the updated version. [PATCH] add device quiescing to the SCSI API This patch adds the ability to quiesce a SCSI device. The idea is that user issued commands (including filesystem ones) would get blocked, while mid-layer and device issued ones would be allowed to proceed. This is for things like Domain Validation which like to operate on an otherwise quiet device. There is one big change: to get all of this to happen correctly, scsi_do_req() has to queue on the *head* of the request queue, not the tail as it was doing previously. The reason is that deferred requests block the queue, so anything needing executing after a deferred request has to go in front of it. I don't think there are any untoward consequences of this. [PATCH] more SPI transport attribute updates This does three things - Fix the signedness of the bit attributes (otherwise they show up as -1 when on, not 1) - Make the period adjust to the closest value rather than ignoring values it doesn't understand. - Add a visibility field to attributes, so drivers can get rid of attributes they're never going to care about [PATCH] update the 53c700 use of transport attributes This patch just brings it up to date with the previous transport attribute patch, moving it to the model where it sets the min/max of the attribute if asked for something outside its range. It also only makes period and offset visible (it doesn't care about any of the others). [PATCH] cdev 1/2: Eliminate /sys/cdev This is the first of two patches designed to make life easier for authors of device driver books - and, with luck, driver authors too. /sys/cdev is, according to Al, a mistake which was never really meant to exist. I believe nothing uses it currently - there isn't a whole lot there to use. Its existence takes up system resources, and also requires drivers to deal with the cdev's embedded kobject in their failure paths. The following patch (against 2.6.4) makes it all go away. OK, almost all. We have to keep (but not register) cdev_subsys because it's rwsem is needed to control access to cdev_map. [PATCH] cdev 2/2: hide cdev->kobj The existing cdev interface requires users to deal with the embedded kobject in two places: - The kobject name field must be set before adding the cdev, and - Should cdev_add() fail, a call to kobject_put() is required. IMO, this exposure of the embedded kobject makes the interface more brittle and harder to understand. It's also unnecessary. With the removal of /sys/cdev, a call to cdev_del() will nicely replace kobject_put(), and the name setting is easily wrapped. This is against 2.6.4, but depends on the /sys/cdev removal patch. remove cdev_set_name completely as it is not needed. [PATCH] update for sata_vsc Hch reviewed my sata_vsc driver and had a few comments. If you could please apply these changes, I'd appreciate it. The copyright is probably the most important for me (apparently I goofed with the OSL. thanks jeremy [libata] clean up module_init() hook of sata_{promise,sil,svw} drivers The more verbose form existed for decreased differences with the 2.4 backport. But differences in this area existed anyway, so why not clean up the 2.6 versions even more. [netdrvr e100] fix stray skb pointer * Not setting cb->skb = NULL after releasing skb to OS or during initialization of cbs. Reported by Deepak Saxena [dsaxena@plexity.net]. [ACPI] SMP poweroff (David Shaohua Li) http://bugzilla.kernel.org/show_bug.cgi?id=1141 [ACPI] ACPICA 20040311 from Bob Moore Fixed a problem where errors occurring during the parse phase of control method execution did not abort cleanly. For example, objects created and installed in the namespace were not deleted. This caused all subsequent invocations of the method to return the AE_ALREADY_EXISTS exception. Implemented a mechanism to force a control method to "Serialized" execution if the method attempts to create namespace objects. (The root of the AE_ALREADY_EXISTS problem.) Implemented support for the predefined _OSI "internal" control method. Initial supported strings are "Linux", "Windows 2000", "Windows 2001", and "Windows 2001.1", and can be easily upgraded for new strings as necessary. This feature allows Linux to execute the fully tested, "Windows" code path through the ASL code Global Lock Support: Now allows multiple acquires and releases with any internal thread. Removed concept of "owning thread" for this special mutex. Fixed two functions that were inappropriately declaring large objects on the CPU stack: ps_parse_loop() and ns_evaluate_relative(). Reduces the stack usage during method execution considerably. Fixed a problem in the ACPI 2.0 FACS descriptor (actbl2.h) where the S4Bios_f field was incorrectly defined as UINT32 instead of UINT32_BIT. Fixed a problem where acpi_ev_gpe_detect() would fault if there were no GPEs defined on the machine. Implemented two runtime options: One to force all control method execution to "Serialized" to mimic Windows behavior, another to disable _OSI support if it causes problems on a given machine. [blk carmel] fix bug, minor cleanups * the scan-channels message seemed to always give invalid output. Look at the constant, and discover we are sending another message entirely. Fix the constant (CARM_MSG_IOCTL). * s/MISC_SYNC_TIME/MISC_SET_TIME/ * list some additional messages * bump version number [ACPI] add boot parameters "acpi_osi=" and "acpi_serialize" acpi_osi= will disable the _OSI method -- which by default tells the BIOS to behave as if Windows is the OS. acpi_serialize is for debugging AE_ALREADY_EXISTS failures [PATCH] Add Domain Validation to the SPI transport class Domain Validation is a fairly essential element to the SCSI Parallel Interface (although if you look very few drivers actually do it). The premise is that the Parallel Bus, being a transmission line, might not be correctly tuned to the transfers you want do perform. DV probes the parameters of the transport until it finds a setting that works (for the interested, see http://www.t10.org/ftp/t10/drafts/sdv/sdv-r08b.pdf) The current code employs rather simplistic DV heuristics, although those can be improved over time. The change in scsi_scan.c is so that DV may be done easily from the slave_configure routine, which is the most natural place to begin. ALSA CVS update - Clemens Ladisch ALSA sequencer remove superfluous call to snd_seq_event_port_detach ALSA CVS update - Clemens Ladisch ALSA sequencer,ALSA<-OSS sequencer use wrapper function for DELETE_PORT ioctl calls ALSA CVS update - Clemens Ladisch USB generic driver use MIN_PACKS_URB as lower bound for nrpacks parameter ALSA CVS update - Clemens Ladisch USB generic driver show one decimal place of momentary frequency in proc file ALSA CVS update - Clemens Ladisch USB generic driver prevent twenty-seconds wait when unplugging USB MIDI device with a port subscription ALSA CVS update - Takashi Iwai VIA82xx driver restrict the PCM sample rates to 32, 44.1 and 48kHz when the SPDIF switch is on. ALSA CVS update - Jaroslav Kysela DT019x driver Fixed warnings ALSA CVS update - Takashi Iwai VIA82xx driver patch was applied wrongly. fixed the rate restriction of spdif output again. ALSA CVS update - Takashi Iwai Documentation,PCI drivers,au88x0 driver added the au88x0 drivers for Aureal soundcards by Manuel Jander ALSA CVS update - Takashi Iwai PPC Tumbler driver added input source switch to select mic/line-in. ALSA CVS update - Takashi Iwai Documentation changed the description of the buffer allocation routines for the new designed functions. ALSA CVS update - Takashi Iwai Documentation fixed the files to include. ALSA CVS update - Takashi Iwai USB generic driver added fix and workaround for the mixer problem on SB Extigy. ALSA CVS update - Takashi Iwai PPC Tumbler driver fixed the info callback of mixer input source (for enum type). ALSA CVS update - Takashi Iwai au88x0 driver removed EXPORT_NO_SYMBOLS. ALSA CVS update - Takashi Iwai EMU10K1/EMU10K2 driver disabled Dell OEM Emu10k1x from the pci id list. the board isn't compatible with the normal emu10k1. ALSA CVS update - Takashi Iwai MIXART driver fixed the compile warning. ALSA CVS update - Takashi Iwai Documentation,PCI drivers,ATIIXP driver added snd-atiixp driver for the ATI IXP150/200/250 AC97 controllers. ALSA CVS update - Takashi Iwai Documentation,PCI drivers,Intel8x0-modem driver added Intel-compatible onboard MC97 modem driver by Sasha Khapyorsky ALSA CVS update - Takashi Iwai ALSA Core added the new magic numbers for atiixp and au88x0 drivers. ALSA CVS update - Takashi Iwai ALSA Core fixed the wrong release of id proc file. [PATCH] Fix removable USB drive oops The actual problem reported was because there wasn't a corresponding check on transport_classdev.class in the unregister. However, on closer inspection I also turned up a nasty thinko in the reference counting. For reasons best known to the class code authors, class devices have to obtain their own references to the devices they're attached to which they release again in their .release routines, so you have to remember to do a get_device() in the correct place after the class_device_add(). I put comments in the code so that, hopefully, we can avoid the problem in future. [WATCHDOG] v2.6.4 pcwd_pci-v1.00_20040313-patch Two small fixes: * Make cards_found a global variable so that if we remove the pci device we can count down. * If we can't find a correct I/O address for the card, then we should disable the card again. WATCHDOG] v2.6.4 wdt977-v0.03-patch Version 0.03 of wdt977.c - Changes that were made are: * Extract the stop code in a seperate function (wdt977_stop) * Extract the start code in a seperate function (wdt977_start) * Rename kick_wdog to wdt977_keepalive for consistency * Extract the watchdog's status code to a seperate function (wdt977_get_status) * Change the way we deal with the watchdog timeout: Up till now we used timeoutM (in minutes) as the correct value and then calculated timeout as being timeoutM*60 or *timeoutM*120 (depending on wether or not we have the netwinder hardware bug). From now on timeout is the correct value and we calculate timeoutM out of it. Because of this we start with checking wether or not we have a correct timeout value (if not we reset it to the default value) and we automatically calculate timeoutM. Each time we change timeout with a correct timeout value, we recalculate timeoutM. * Extended ioctl code with WDIOC_SETOPTIONS and updated the watchdog_info structure * Added notifier support Code has been tested by Woody [WATCHDOG] v2.6.4 notifier_block-patches Remove unnecessary initialization in notifier_block NFSv2/v3/v4: New attribute revalidation code that no longer relies on ctime for correctness in avoiding update races. VFS: allow filesystems to disable inode_update_time() on a per-inode basis. NFSv2/v3/v4: New file writeout strategy. Defer writes until a flush is requested by the application (or memory pressure). Configuration: simplify configuration options. Automatically select RPCSEC_GSS if NFSv4 is selected. Remove need for user to select SUNRPC_GSS, and the crypto options. Make NFSv3 a recommended option. NFSv2/v3: Ensure that we only use GETATTR+STATFS (NFSv2) and FSINFO (NFSv3) when mounting. This should allow us to use AUTH_SYS credentials when mounting, (even when the user requests RPCSEC_GSS authentication) due to the hack described in RFC2623. Remove the broken NFS_INO_FAKE_ROOT hack. NFSv2/v3/v4: Ensure that fsync() flushes all writebacks to disk rather than just the ones labelled as belonging to our file. This fixes a bug in which msync(MS_SYNC) will fail to flush the pages to disk. NFSv2/v3/v4: A patch by Greg Banks that fixes the "VFS: Busy inodes after unmount." problem. RPC: Make XIDs unique on a per-transport basis rather than globally unique. Gets rid of an unnecessary global spinlock. RPC: Sync rpc_set_timeo() up to the 2.4.x version. In particular, this will ensure that the timeout shift is clamped to a maximum value of 8. RPC: Fix by Olaf Kirch to the rpc scheduler to ensure sync tasks respect the "intr" mount flag. RPC: Ensure that we have the correct capabilities when binding a socket to a reserved port. Fixes a privilege bug when CONFIG_SECURITY is set. RPC: When trying to reconnect to a TCP port, try to bind() to the last used port number in order to ensure that the servers NFS replay cache recognizes this as being the same mount as before. RPC,NFSv2/v3/v4: Ensure that xprt_create_proto() and rpc_create_client() return full error codes. Should allow the "mount" program to print more useful error diagnostics. NFSv2/v3/v4: Parenthesize #defines in nfs?xdr.c. Fix an off-by-one error on the value of compound_decode_hdr_maxsz. NFSv4: fix a printk() typo (spotted by Linda Dunaphant). NFSv4: Ensure that nfs4_open_reclaim() copies the value of the new stateid back into the shared nfsv4 state structure. NFSv4: Don't leak NFS4ERR_WRONGSEC errors back into nfs_lookup(). RPC,NFS,Lockd: Mark the debugging code as "unlikely" so that gcc moves it out of the mainline code paths. NFSv2/v3 locking: Patch by Patrice Dumas to implement nlmsvc_proc_granted_res. When a server receives that callback it should deallocate the corresponding blocked lock using the nlmsvc_grant_reply function. NFSv2/v3 locking: Patch by Patrice Dumas that adds a check to ensure we really were requesting a blocking lock when we get a reply from the server asking us to block. NFSv2/v3 locking: Patch by Patrice Dumas to ensure that the server index blocks uniquely by using the client address in addition to the value of the NLM cookie field. NFSv2/v3 locking: A patch to ensure that blocks which are not going to time out are placed last on the ordered list nlm_block (problem reported by Olaf Kirch). RPC,NFSv3: remove the redundant "memset()" in call_encode(). Fix up the only places where this causes a padding error: xdr_encode_fhandle() and unx_marshal() RPC: patch by Chuck Lever to make the number of RPC slots a tunable parameter. This is wanted in order to allow the NFS client to send more requests before is has to block and wait for replies. This is mainly useful if you have a WAN and want to ensure that the bandwidth is being used efficiently. NFSv2: Fix up NFSv2 reads so that they report when the server returned a short read due to EOF. NFSv4: Fix a list corruption in the NFSv4 state engine. NFS: From the suse kenrel RPM: handle ENOMEM from nfs_fhget(). From: Forward-port from 2.4: The following patch pulls an NFS server IP address off root_server_path (handed out via the DHCP root-path option), if it is present. For example, you can do this sort of thing in dhcpd.conf: root-path = 192.168.1.33:/tftpboot/yip.zImage This lets you mount your root filesystem off a different machine than you booted from, without needing to use kernel command-line parameters. The patch appears to be backwards compatible. RFC2132 says this about the root-path option: This option specifies the path-name that contains the client's root disk. The path is formatted as a character string consisting of characters from the NVT ASCII character set. This is sufficiently vague to allow the path-name to include an IP-address. Also, I found some documentation for FreeBSD saying it does this too, so it must be right, because those FreeBSD guys are really smart... :-) The only downside of the patch is that the summary that ipconfig prints can be a little odd when the kernel command line overrides whatever ipconfig gets from (say) DHCP. The address from the kernel command line seems to get stripped off early, so ipconfig reports it, but it doesn't report the kernel command line NFS path, since that's handled a bit later... This small cosmetic problem looks difficult to "fix" without rewriting quite a bit of stuff... akpm@odsl.org: For complex reasons it is not possible to hold i_sem in nfs_update_inode(). Hence the i_size_write() in there is deadlocky. Go back to the old way. [PATCH] ppc64: fix NUMA compile with large cpumasks From: Anton Blanchard The recent NUMA changes fail to compile with large cpumasks, we need to use a temporary to get around the type checking. [PATCH] Use 64-bit counters for scheduler stats From: Kingsley Cheung A number of scheduler counters wrap around after 47 days. The context-switch counter can wrap around after considerably less time. Convert them to 64-bit values. [PATCH] Manfred's patch to distribute boot allocations across nodes From: Anton Blanchard Distribute boot time memory allocations across all nodes, from Manfred Spraul. We want to spread memory across nodes to avoid all allocations ending up on node 0. Spreading boot time allocations around also helps us to avoid node 0 becoming the hot node. I took it for a spin: buddyinfo before: Node 7, 0 2 1 1 0 2 1 2 1 2 1 2 741 Node 6, 0 0 0 2 0 2 1 1 2 2 2 2 1002 Node 5, 0 0 0 2 0 2 1 2 1 2 2 2 2006 Node 4, 0 0 0 2 0 2 1 2 1 2 2 2 2006 Node 3, 0 0 0 2 0 2 1 2 1 2 2 2 2006 Node 2, 0 0 0 2 0 2 1 2 1 2 2 2 2006 Node 1, 0 0 0 2 0 2 1 1 2 2 2 2 1002 Node 0, 0 0 38 7 0 1 1 1 0 0 0 0 1998 buddyinfo after: Node 7, 0 1 0 1 1 1 1 0 0 0 1 2 738 Node 6, 0 1 0 1 1 1 0 1 0 0 2 2 1002 Node 5, 0 0 0 1 1 1 1 0 0 0 2 2 2006 Node 4, 0 1 0 1 0 1 1 0 0 0 2 2 2006 Node 3, 0 0 0 1 0 1 1 0 0 0 2 2 2005 Node 2, 0 1 0 0 0 0 0 1 0 0 2 2 2006 Node 1, 0 2 1 1 0 1 1 1 0 0 2 2 1002 Node 0, 0 20 45 8 3 0 1 1 1 1 0 1 2004 Change in free memory due to patch: Node 7 -54.08 MB Node 6 -6.33 MB Node 5 -6.09 MB Node 4 -6.14 MB Node 3 -22.15 MB Node 2 -6.05 MB Node 1 -6.12 MB Node 0 107.35 MB As you can see we gained over 100MB on node 0. [PATCH] further __KERNEL_SYSCALLS__ removal From: Arnd Bergmann Dave Jones already removed some of the useless __KERNEL_SYSCALLS__ defines in various files, this gets rid of almost all the others. Replacing execve() is nontrivial, so I left those in for now. For all the other system calls that are currently used from inside the kernel, calling the sys_* function directly should always have an identical effect. [PATCH] use wait_task_inactive() in kthread_bind() From: Rusty Russell Make it clear that the reason we do wait_task_inactive is because kthread_bind frobs with k->thread_info->cpu, which is only legal because the task is definitely not running. We can't use the normal migration thread code here, because it doesn't let us bind to cpus which are offline yet, and also because we use this to start the migration threads themselves. [PATCH] md: use "shedule_timeout()" instead of yield() From: NeilBrown Use "shedule_timeout()" instead of yield() as it seems to wait for less time. [PATCH] md: allow assembling of partitioned arrays at boot time. From: NeilBrown kernel parameters: raid=partitionable will make all auto-detected md arrays partitionable md=d.... will assemble an array as a partitionable array. [PATCH] Work around an AMD768MPX erratum From: Andi Kleen This patch has been in the SuSE 2.4 kernel forever, but for some reason never made it mainline. It works around the infamous "only works stable when a mouse is plugged in" problem some AMD 768MPX Dual Athlon chipsets have. The problem happens because the chipset can hang when PCI prefetch strides from a RAM page into the VGA text buffer. When a PS2 mouse is plugged in the BIOS reserves a page before the VGA text buffer, which stops the prefetch early. This patch always reserves this page when the chipset could be AMD768MPX. This can be only done early in bootmem setup. Because it's difficult to scan the PCI bus that early it just always reserves this page when the CPU is an Athlon. Normally it should not make a difference because the BIOS will have reserved that page anyways when a PS/2 mouse is plugged in. [PATCH] Add Domain Validation to 53c700 driver This simply throws out the 53c700 driver's optimistic setting of the best possible transport parameters and replaces it with DV determination. It also adds a missing report_bus_reset() callback and finally does a delayed DV on device errors. [PATCH] DMA: Fill gaping hole in DMA API interfaces. From: "David S. Miller" Currently, for an existing DMA mapping, there is a way to transfer buffer ownership back to the cpu, yet there is no way to give it back to the device again explicitly. The latter really is needed on platforms where the PCI subsystem does not snoop the cpu caches, MIPS is one example. Many drivers were expecting the existing DMA sync interface to handle both directions, which was wrong. Now, with this change, we have explicit interfaces for DMA syncing to/from the device and the cpu. [PATCH] module unload deadlock fix From: Rusty Russell From: Andrea Arcangeli We should drop module sem before calling mod->exit, for practical reasons: too many module exit functions oops or hang, resulting in a permenantly held module sem, which blocks all module ops including lsmod. [PATCH] gcc-3.5 libata build fix drivers/scsi/sata_vsc.c: In function `vsc_sata_interrupt': include/linux/libata.h:414: sorry, unimplemented: inlining failed in call to 'ata_host_intr': function body not available drivers/scsi/sata_vsc.c:187: sorry, unimplemented: called from here [Bluetooth] Make use of the MODULE_VERSION macro Each module contains a version number and this patch uses MODULE_VERSION to make it visible through modinfo. [PATCH] PA-RISC update Updates for 2.6.4 for PARISC: - defconfigs (Randolph Chung) - copyright updates (Grant Grundler) - Fix DOS hole in sys_clone (James Bottomley) - missing hardware ID (Andy Walker) - disable interrupts during cache-flushes (LaMont Jones) - Fix crash on machines with <256MB ram (Randolph Chung) - Make SuckyIO IDE work better (Randolph Chung) - Align data_start so the extable is writable (Randolph Chung) - Extensive rewrite of virtual merging code (James Bottomley) - Fix EISA, non-PCI module builds (Matthew Wilcox) - Fix Elroy PCI config space byte & word writes (Grant Grundler) - Eliminate a warning in parport_gsc (Helge Deller) - Fix endian problem with ide mmio macros (Randolph Chung) - Delete asm/keyboard.h (Matthew Wilcox) - Delete asm/md.h (Grant Grundler) - Eliminate a warning in ALSA harmony (Matthew Wilcox) [PCMCIA] Add, fix, update PCMCIA debugging. The PCMCIA core debugging assumes that "pc_debug" is a global. If pc_debug were to be a global, it would pollute the kernel's global name space, potentially clashing with PCMCIA drivers and even ds.c. We also get rid of PCMCIA using the "DEBUG" macro, which is the standard macro to enable debugging in the kernel. Instead, we introduce cs_dbg() for core PCMCIA debugging, ds_dbg() for driver services debugging, and similar macros in the socket drivers. [PCMCIA] Clean up socket state handling around shutdown There is no need for shutdown_socket() to touch s->state directly - this can be all done in skt_shutdown(). [PCMCIA] Rename driver services constants. Driver services constants clash with similar constants defined by Card services. Prefix them all with DS_. This patch started out as Dominik Brodowski's 2.6.0-test9 ds_includes_cs_internal patch, which only changed one constant and attempted to fix PCMCIA debugging. [PCMCIA] move_pcmcia_bind_device Patch from Dominik Brodowski, updated by Russell King. pcmcia_bind_device was only used by ds.c, and it is only used for 16-bit PCMCIA devices. So, move it to ds.c. [PCMCIA] move_pcmcia_bind_mtd Patch from Dominik Brodowski, updated by Russell King. pcmcia_bind_mtd was only used by ds.c, and it is only used for 16-bit PCMCIA devices. So, move it to ds.c. [PCMCIA] move pcmcia_report_error and cs_error Patch from Dominik Brodowski, updated by Russell King. cs_error and pcmcia_report_error are only used by 16-bit PCMCIA devices, so move them to ds.c. [PCMCIA] socket user operations should take pcmcia_socket Patch from Dominik Brodowski, updated by Russell King. The pcmcia_{suspend,resume,eject,insert}_card calls are only issued by the userspace cardctl tool via ds.c. There's no need to verify the caller (ds) has a valid "handle", as long as it knows the proper struct pcmcia_socket. [PATCH] move consistent_dma_mask to the generic device From: James Bottomley pci_dev.consistent_dma_mask was introduced to get around problems in the IA64 Altix machine. Now, we have a use for it in x86: the aacraid needs coherent memory in a 31 bit address range (2GB). Unfortunately, x86 is converted to the dma model, so it can't see the pci_dev by the time coherent memory is allocated. The solution to all of this is to move pci_dev.consistent_dma_mask to dev.coherent_dma_mask and make x86 use it in the dma_alloc_coherent() calls. This should allow me to make the aacraid set the coherent mask instead of using it's current dma_mask juggling. [PATCH] s390: update for altered page_state structure From: Olaf Hering This patch is needed on s390. [PATCH] __kill_pg_info() return value fix Fix a bug which was spotted by Alex Lyashkov The fairly unobvious coding in __kill_pg_info() will cause a zero value to be incorrectly returned if the second or succeeding call to group_send_sig_info() returns an error. [PATCH] Fix voyager to boot again The very early memory detection patch broke voyager. This fixes it again. [Bluetooth] Fix compile errors with enabled debugging This patch fixes some stupid compile errors when debugging is enabled. Noticed by Christian Hesse [SPARC32]: Rework the CPU enumeration and probing code The low 2 bits of the MID are now used exclusively as the software CPU ID on 4m. This eliminates entirely the need for mid_xlate. We also eliminate linux_num_cpus and linux_cpus; since the hardware-based id is now used everywhere we don't need indexed CPU tables. The device probe is almost identical to sparc64's version. The one exception is that 4m only uses the bottom 2 bits of the MID, so we also have a separate function to get the true hardware MID for the rare times it's actually needed. Per-cpu data is now used the same way as in sparc64. As a side benefit we now display the same format Bogo and Clock values in /proc/cpuinfo. [SPARC32]: Also remove num_cpus_possible as was done on sparc64 [libata sata_sil] remove incorrect limit on drive quirk The original submittor of the quirk code noted that the errata of a lockup creating when the condition "sector % 15 == 1" is true only on chip revisions <= 1. This is limitation on chip revision is apparently not correct, therefore we apply to all affected drives in the quirk list. This is a sledgehammer fix, limiting max-sectors to 15. A much better fix would split the hardware request into two taskfiles, issuing an additional command, thereby avoiding "sectors % 15 == 1" Given the errata is limited to "sil 311x + ", it is hoped that some janitor cares enough to do a real fix. [SPARC32]: Fix build; we don't need KERNEL_SYSCALLS but asm/unistd.h is necessary [SBUS]: Fix sound build if CONFIG_SBUS is set [SPARC32]: Add per-cpu data header [SOUND]: Fix typo in SBUS memalloc changes. [NETDEV]: pcnet32, eepro100, and 8139too need asm/irq.h ALSA CVS update - Jaroslav Kysela Intel8x0 driver Added slot definitions for s/pdif pcm - ICH4 ALSA CVS update - Jaroslav Kysela AC97 Codec Core Kevin Mack Here's a quick and dirty patch that's given me basic sound from my Gateway M675 notebook (Sigmatel 9758 AC97 codec). ALSA CVS update - Clemens Ladisch Documentation,MPU401 UART integrate MPU-401 ACPI PnP from alsa-driver ALSA CVS update - Clemens Ladisch au88x0 driver fix compilation on gcc 2.95.x ALSA CVS update - Clemens Ladisch Intel8x0 driver add Intel ICH6 and ESB ALSA CVS update - Jaroslav Kysela au88x0 driver Make mchannels and rampchs static ALSA CVS update - Jaroslav Kysela au88x0 driver Fixed gameport dependency and solid kernel build [ARM PATCH] 1772/1: ARM README changes Patch from Marc Singer Updates to the ARM architecture README file. Mostly, this is done for clarity. [ARM] Update ARM README Provide more detailed information on machine types and machine classes. [ARM PATCH] 1753/1: Devfs support for the 21285 serial driver - try 2 Patch from Jonas Larsson This patch adds devfs_name member to struct uart_driver. No devfs entries are created for the serial port without this patch. [PATCH] update for pdc202xx_new driver Many thanks to Promise for their support. Expect more updates/fixes for Promise PATA support soon. - fix PIO (auto-)tuning - use pdcnew_new_tune_chipset() and always tune PIO mode even if (U)DMA is used - cleanup cable verification code a bit - remove leftovers from driver split-up - remove duplicate DISPLAY_PDC202XX_TIMINGS define [PATCH] ide-disk.c: cleanup get_command() [PATCH] remove ide_cmd_type_parser() logic Set ide_task_t fields (command_type, handler and prehandler) directly. Remove unused ide_task_t->posthandler and all ide_cmd_type_parser() logic. ide_cmd_type_parser() was meant to be used for ioctls but ended up checking validity of kernel generated requests (doh!). Rationale for removal: - it can't be used for existing ioctls (changes the way they work) - kernel shouldn't check validity of (root only) user-space requests (it can and should be done in user-space) - it wastes CPU cycles on going through parsers - it makes code harder to understand/follow (now info about request is localized) [PATCH] remove IDE_*_OFFSET_HOB and IDE_*_REG_HOB defines They are identical to non _HOB versions (except IDE_CONTROL_OFFSET_HOB). [PATCH] remove ide_init_drive_taskfile() From: Frederik Deweerdt [PATCH] piix_ide_init() can be __init From: "Randy.Dunlap" From: Luiz Fernando Capitulino The missing __init was found by reference_init26.pl. [PATCH] H8/300: Interrupt handling cleanup - merge common routine - runtime vector setup update [PATCH] H8/300: fix build error - add asm/dma-mapping.h - add CONFIG_PCI [PATCH] H8/300: fix waring - fix gcc-3.4.0 warnings [PATCH] H8/300: makefile cleanup [libata] disable clustering by default, whitespace cleanups Disable SCSI's clustering of S/G entries, typically reducing overall S/G list size. Although ATA S/G tables are getting smarter (64-bit address, 32-bit length, no DMA boundary besides 4G itself), working harder to generate large byte sizes for each S/G size is not necessarily in our best interest: * SATA consists of frames, much like ethernet, called "FIS". A Data FIS consists of up to 2048 dwords (8K) of data. We don't want generate large byte sizes in S/G entries, just to have the host controller break them up again into smaller chunks. * Reduces the possibility that ata_fill_sg (libata-core.c) must split an S/G into two pieces, because it straddles a 64K boundary. * Reduces the possibility that FIS-related errata workarounds must be enabled. One such exists in the Silicon Image driver (sata_sil), which is currently worked around by limiting the max-sectors-per-request to 15. [libata] kill the warning everybody grumbles about [ARM] Update mach-types file. [ARM] Provide userspace method for controlling LEDs in ARM machines. [ARM] Move consistent_xxx exports to arch/arm/mm/consistent.c [ARM] Remove export of kd_mksound. This symbol is no longer used in modules. [Bluetooth] Declare more functions static This patch declares all possible functions static. [PATCH] I2C: fix up CONFIG_I2C_DEBUG_BUS logic to be simpler on the .c files. [PATCH] I2C: fix up CONFIG_I2C_DEBUG_CORE logic to be simpler on the .c files. [PATCH] I2C: add CONFIG_I2C_DEBUG_ALGO to be consistant. [PATCH] I2C: fix up CONFIG_I2C_DEBUG_CHIP logic to be simpler on the .c files. [ARM] Allow run-time selection of user debugging messages. [ARM] Add asm/irq.h include - required for NR_IRQS [SPARC]: Include linux/linkage.h in asm/unistd.h [TUN]: Name fix. This fixes both the devfs name and the /sys name issues. It ends up as /dev/net/tun and /sys/class/misc/tun looks correct. [TUN]: Do not obscure error return from misc_register in tun_init. [TUN]: Fix user buffer verification. When Tun initialization fails it shouldn't obscure the error return, There are several reasons it could fail, and when diagnosing user problems the error code might provide more info. [ATM]: [suni] dev_data should really be phy_data. [TCP]: Kill westwood bw_sample, set but never used. [PATCH] I2C: fix forced i2c chip drivers have no name I just noticed that I am doing something wrong in the i2c chip drivers I ported to Linux 2.6. If these drivers are forced to a specific chip type ("kind" as we call it internally), then the device doesn't have its name set (and defaults to an empty string). Affected drivers: gl518sm, lm83, lm90, w83l785ts. I could verify the problem on my ADM1032 chip (lm90 driver). I also verified that the proposed patch fixes the issue. You may notice that I fix the problem differently for gl518sm and w83l785ts on the one hand, and lm83 and lm90 on the other hand. This is because the first two drivers are not expected to support more a single chip in the future, while lm90 already does and lm83 could someday (for example, support for the LM82 could be added on request). [PATCH] I2C: Don't handle kind errors that cannot happen A number of chip drivers in 2.6.4-mm1 try to handle an error case that cannot happen when setting the chip name. The following patch changes that. Affected drivers: adm1021, it87, lm75, lm78, lm85, w83627hf, w83781d. Note that in any case, the worst that could happen (but then again, it cannot happen) is that the chip name would be set to an empty string, which doesn't hurt much. The patch also cleans up a few things in it87, w83627hf and w83781d, which are tightly related to the rest of the changes and necessary for them to be safe. it87: There is only really one "kind" in this driver, so I removed all references to other kinds. w83627hf: The driver did not handle unknown chips. w83781d: The user shouldn't be allowed to force a kind that doesn't match the chip's bus type (I2C or ISA). The code was not meant to handle that case, although no check was done so far. Tested on my AS99127F, works as intended. [PATCH] I2C: Setting w83781d fan_div preserves fan_min This patch makes the w83781d driver preserve fan_min settings when the user changes fan_divs. This isn't done "by default" because the actual fan min value (in RPMs) depends on both the fan_min register and the fan_div register. Only two drivers handle it properly at the moment as far as I know (lm78 and asb100). Several other drivers would need to be fixed the same way, but well, once at a time ;) Tested on my AS99127F rev.1. Credits go to Philip Pokorny, since I think I remember he is the one who introduced the method in the lm78 driver in the first place. This tends to increase the size of the three set_store_regs_fan_div functions, and I am considering refactoring them at some point. Later though. [PATCH] I2C: New chip driver: ds1621 The following patch against kernel 2.6.4-mm1 adds the ds1621 driver (an I2C sensor). I have ported it from the 2.4 version. It has been reviewed by Jean Delvare, partly on IRC, and it is "compliant" with Mark Hoffman's refactoring. [NET]: Make netdevice.h more non-kernel friendly. [VLAN]: Use array for static const char 'fmt' [NET_SCHED]: Use time_after, fixes htb on 64-bit arch. [PATCH] I2C: delete the i2c-elv.c driver as it is obsoleted by the i2c-parport.c driver. [PATCH] I2C: delete the i2c_philips-par.c and i2c-veleman.c drivers They are obsolted by the i2c-parport.c driver. [SERIAL] Add alias for TTY_MAJOR character device. Patch from Steve Youngs [PATCH] cdev: warning fix Against Jon's cdev stuff kref: add kref structure to kernel tree. Based on the kobject structure, but much smaller and simpler to use. [SERIAL] fix PCI interrupt setting for ia64 Patch from Bjorn Helgaas ACPI and HCDP tell us what IRQ the serial port uses, so there's no need to have the driver probe for the IRQ. [SERIAL] serialP.h: remove a kernel 2.2 #ifdef Patch from Adrian Bunk [TCP]: Clean up some westwood comments. [PATCH] generic 32 bit emulation for System-V IPC From: Arnd Bergmann Adds a generic implementation of 32 bit emulation for IPC system calls. The code is based on the existing implementations for sparc64, ia64, mips, s390, ppc and x86_64, which can subsequently be converted to use this. [PATCH] Remove unneeded cast. [PATCH] Whitespace fixes. No code changes, but lots of trivial whitespace cleaning, and removal of a bogus set of 'defaults' at the bottom of the file. [PATCH] Fix sysfs leak. If the driver fails to load, we leave a 3c509 eisa directory in sysfs. [PATCH] Set module owner in megaraid driver This bug was discussed on linux-scsi a few weeks back, but it appears to have slipped through the cracks. This fix was originally proposed by Christoph Hellwig, but I am reposting it, since I have the hardware :-) Anyway, the problem that this fixes is that megaraid doesn't set the moduler owner for the host_template, this means that the module can be removed, even when it is in use. [PATCH] sys_swapon bad arg causing slab corruption There's an error in sys_swapon() that can cause slab corruption if you pass in a bad specialfile pointer. getname() then returns ERR_PTR(-EFAULT), but sys_swapon() doesn't clear name before calling putname() on it (thus freeing 0xfffffff2, corrupting slab). An ltp test case repeatedly crashed in later tests due to thus, irk. [PATCH] Fix CONFIG_DEBUG build on x86-64 & small cleanup This fixes the CONFIG_DEBUG_INFO build on x86-64 (there were missing .cfi_endprocs) Also some minor cleanup in the exception stack handling. Linux 2.6.5-rc1