Linux v2.6.12-rc6

Linus Torvalds

unread,

Jun 6, 2005, 2:20:26 PM6/6/05

to

It's being uploaded right now, the git tree is already up-to-date, and by
the time this hits the mailing list the mirroring of the tar-ball will
hopefully be done too.

And since Jeff wrote me a shortlog script for git, the easist way to tell
what's new since -rc5 is to just do the shortlog and diffstat output.
Network drivers, USB and CPU-freq stand out.

And the good news is that people do seem to have taken my rumblings about
calming down for 2.6.12 seriously. Let's hope that pans out, and I can
release that one asap.. But give this a good beating first, and holler
(again, if you must) about any issues you have,

Linus

---
Adrian Bunk:
USB: remove drivers/usb/media/pwc/ChangeLog
drivers/net/hamradio/baycom_epp.c: cleanups
[IPV6]: Kill export of fl6_sock_lookup.
[IPVS]: remove net/ipv4/ipvs/ip_vs_proto_icmp.c
arch/i386/kernel/cpu/intel_cacheinfo.c: section fix
Input:
SIS900 must select MII

Alan Cox:
remove non-cleanroom pwc driver compression

Albert Lee:
libata: Fix zero sg_dma_len() on 64-bit platform

Alexander Nyberg:
acpi build fix: x86 setup.c
Note on ACPI build fix
Fixup VIA IRQ quirk
x86_64: CONFIG_BUG=n fixes

Alexey Dobriyan:
[TOKENRING]: net/802/tr.c: s/struct rif_cache_s/struct rif_cache/
[TOKENRING]: be'ify trh_hdr, trllc, rif_cache_s

Andi Kleen:
x86_64 CONFIG_ACPI=n build fix
x86_64: More fixes for compilation without CONFIG_ACPI

Andrew Morton:
Input: Fix a warning in psmouse-base.c

Andrew Vasquez:
[SCSI] qla2xxx: fix bad locking during eh_abort

Andy Currid:
PCI: amd74xx patch for new NVIDIA device IDs

Anton Blanchard:
ppc64: remove decr_overclock
ppc64: cleanup iseries runlight support
ppc64: cleanup SPR definitions
ppc64: allow timer based profiling on iseries

Bartlomiej Zolnierkiewicz:
convert IDE device drivers to driver-model

Benjamin Herrenschmidt:
ppc64: Fix result code handling in prom_init
ppc32/ppc64: cleanup /proc/device-tree
ppc32: Apple device-tree bug fix
ppc64: Fix a device-tree bug on Apple's
ppc32: Fix cpufreq vs. sleep issue
ppc32: Fix Alsa PowerMac driver on old machines
ppc32: small cpufreq update

Benjamin LaHaise:
ns83820 update

Bodo Stroesser:
s390: uml ptrace fixes

Christoph Hellwig:
[NET]: Fix locking in shaper driver.
[XFS] remove an over-zealous WARN_ON
Merge with /.../torvalds/linux-2.6.git

Colin Leroy:
therm_adt746x: show correct sensor locations
Make sure therm_adt746x only handles known hardware

Craig Shelley:
USB: CP2101 Add support for flow control

Dan Williams:
wireless/airo: WEXT and quality corrections

Daniel Ritz:
3c574_cs: disable interrupts in el3_close

Daniele Venzano:
More ethtool support for sis900 and warning fix

Dave Jones:
[CPUFREQ] Typos.
[CPUFREQ] longhaul - adjust transition latency.
[CPUFREQ] Longhaul: Magic timer frobbing.
[CPUFREQ] longhaul - disable PCI mastering around transition.
[CPUFREQ] ondemand governor default sampling downfactor as 1
[CPUFREQ] ondemand governor automatic downscaling
[CPUFREQ] ondemand,conservative governor idle_tick clean-up
[CPUFREQ] ondemand,conservative governor store the idle ticks for all cpus
[CPUFREQ] ondemand,conservative minor bug-fix and cleanup
[CPUFREQ] Allow ondemand stepping to be changed by user.
[CPUFREQ] Prevents un-necessary cpufreq changes if we are already at min/max
[CPUFREQ] Add support to cpufreq_ondemand to ignore 'nice' cpu time
[CPUFREQ] Conservative cpufreq governer
[CPUFREQ] fix up comment in cpufreq.h
[CPUFREQ] dual-core powernow-k8
[CPUFREQ] make cpufreq_gov_dbs static
[CPUFREQ] Recalibrate cpu_khz [2/2]
[CPUFREQ] Recalibrate cpu_khz [1/2]
[CPUFREQ] AMD Elan SC520 cpufreq driver.
[CPUFREQ] Add warning comment about default governors.
[CPUFREQ] speedstep-smi: it works on at least one P4M
[CPUFREQ] ondemand: trivial clean-ups
[CPUFREQ] speedstep-centrino: Pentium 4 - M (HT) support
[CPUFREQ] cpufreq-core: reduce warning messages.
[CPUFREQ] powernow-k7: don't print khz element of FSB.
Fix up pwc driver compilation.

David Brownell:
USB: resolve Zaurus problem
USB: add sl811_cs support
USB: sl811-hcd fixes

David Mosberger-Tang:
[IA64] Correct convert_to_non_syscall()
[IA64] Avoid .spillpsp directive in handcoded assembly
[IA64] fix "section mismatch" compile-time-error
[IA64] Fix stack placement when INIT hits in kernel mode.

David S. Miller:
[SPARC64]: Refine PCI strbuf ctx-based flush.
[SPARC64]: Fix streaming buffer flushing on PCI and SBUS.
Merge of /home/davem/src/GIT/tg3-2.6/
[NET]: Use %lx for netdev->features sysfs formatting.
[IPV6]: Clear up user copy warning in flowlabel code.
Merge of /home/davem/src/GIT/linux-2.6/.git/
Merge of da...@nuts.davemloft.net:/disk1/GIT/sparc-2.6/.git/
[SPARC64]: Add boot option to force UltraSPARC-III P-Cache on.

David Woodhouse:
Speedtouch resync after lost signal.

Dmitry Torokhov:
Input: synaptics - reduce verboseness of synaptics driver - there
Input: yet another model that does not play nicely when i8042 is
Input: automatically disable MUX mode on Toshiba Satellite P10
Input: gunze - fix out-of-bound array access reported by Adrian Bunk.
Input: Tone down the severity of a printk() in i386/ia64 arch code

Domen Puncer:
drivers/scsi/ahci: add #include req'd for the DMA_{64,32}BIT_MASK constants
drivers/scsi/sata_vsc: add #include req'd for DMA_32BIT_MASK constant

Don Fry:
pcnet32: fix resource leak with loopback test

Edgar E Iglesias:
[IPSEC]: Fix esp_decap_data size verification in esp4.

Edward Falk:
libata: update inline source docs

Francisco Javier:
sata_promise: add PCI ID for FastTrak TX2200 2-ports

Francois Romieu:
r8169: incoming frame length check

Frank Pavlic:
s390: qeth bug fixes
s390: qeth bug fixes
s390: fakell for high speed token ring
s390: qeth bug fixes
s390: enable iucv_send2way_xxx functions
s390: ctc code cleanup
s390: schedule_timeout cleanup in ctctty
s390: set online race in the lcs driver
s390: multicast address registration in lcs
s390: claw driver wiring

Gerald Schaefer:
s390: deadlock in appldata

Gerd Knorr:
v4l: bttv i2c oops fix

Goffredo Baroncelli:
UDF filesystem: array '__mon_yday' declared as not static

Greg Kroah-Hartman:
USB: add Vernier devices to HID blacklist
USB: fix usb-serial generic initialization

Greg Ungerer:
m68knommu: fix scheduling and race problems in idle loop

Harald Welte:
[NETFILTER]: Fix deadlock with ip_queue and tcp local input path.
[IPV4]: Primary and secondary addresses

Herbert Xu:
[IPV4]: Fix BUG() in 2.6.x, udp_poll(), fragments + CONFIG_HIGHMEM
Fw: [Bugme-new] [Bug 4482] New: natsemi: incorrect initialization of IPv6 Neighbor-discovery multicast

Hideaki YOSHIFUJI:
commit 2f872f0401d4b470990864fbf99c19130f25ad4d

Ian Abbott:
USB: ftdi_sio: new PID for ELV UM100

James Harper:
fix PROMISC/bridging in TLAN driver

Jan Beulich:
[ATM]: fix ATM makefile for out-of-source-tree builds

Jan Kara:
ext3: fix list scanning in __cleanup_transaction
ext3: fix log_do_checkpoint() assertion failure

Jay Vosburgh:
[BONDING]: bonding using arp_ip_target may stay down with active path

Jeff Dike:
uml: remove unused code
uml: fix segfault on exit with CONFIG_GCOV
uml: single-space a help message
uml: remove 2_5compat.h
uml: turn off kmalloc always on a fatal signal
uml: fix a couple of warnings

Jeff Garzik:
Automatic merge of /spare/repo/netdev-2.6 branch r8169-fix
Automatic merge of /spare/repo/linux-2.6/.git branch HEAD
libata: kernel-doc warning fixes
libata: more docs updates
libata: doc updates
libata: more doc updates
libata: minor DocBook update
libata: bump version
Automatic merge of /spare/repo/netdev-2.6 branch use-after-unmap
libata: Fix use-after-iounmap
Automatic merge of rsync://rsync.kernel.org/.../torvalds/linux-2.6.git branch HEAD
Automatic merge of rsync://rsync.kernel.org/.../torvalds/linux-2.6.git branch HEAD
Automatic merge of /spare/repo/netdev-2.6 branch tlan
Automatic merge of /spare/repo/netdev-2.6 branch sis900
Automatic merge of /spare/repo/netdev-2.6 branch veth
Automatic merge of /spare/repo/netdev-2.6 branch qeth
Automatic merge of /spare/repo/netdev-2.6 branch ns83820
Automatic merge of /spare/repo/netdev-2.6 branch natsemi
Automatic merge of /spare/repo/netdev-2.6 branch forcedeth
Automatic merge of /spare/repo/netdev-2.6 branch airo
Automatic merge of /spare/repo/netdev-2.6 branch atmel
Automatic merge of /spare/repo/netdev-2.6 branch amd8111
Automatic merge of /spare/repo/netdev-2.6 branch pcnet32
Automatic merge of /spare/repo/netdev-2.6 branch ixgb
Automatic merge of /spare/repo/netdev-2.6 branch e1000
Automatic merge of /spare/repo/netdev-2.6 branch e100
Merge of /spare/repo/netdev-2.6 branch misc-fixes

Jens Axboe:
Relax idecd dma alignment check
relax ide-cd dma restrictions

Jesper Juhl:
[ATM]: [drivers] kill pointless NULL checks and casts before kfree()

Jesse Barnes:
update sn2 maintainer

Jiri Benc:
[NET]: Fix HH_DATA_OFF.

John Hawkes:
drop note_interrupt() for per-CPU for proper scaling

John W. Linville:
tulip: add return to ULI526X clause in tulip_mdio_write

Jon Mason:
[NET]: Add ethtool support for NETIF_F_HW_CSUM.

Kenji Kaneshige:
PCI Hotplug: SHPCHP driver doesn't enable PERR and SERR properly
PCI Hotplug: shpchp driver doesn't program _HPP values properly

Kumar Gala:
ppc32: i8259 PIC should not be initialized if PCI is not configured
ppc32: Add soft reset to MPC834x
ppc32: MPC834x BCSR_SIZE too small for use in a BAT.
ppc32: Simplified load string emulation error checking
ppc32: Fix building MPC8555 CDS when CONFIG_PCI is disabled
ppc32: Add VIA IDE support to MPC8555 CDS platform
ppc32: Fix some minor issues related to FSL Book-E KGDB support
ppc32: Fix uImage make target to report success correctly

Kurt Garloff:
Input: Avoid double unregistering of i8042 PnP driver. This can happen

Len Brown:
ACPI build fix
VIA IRQ quirk

Linus Torvalds:
Linux 2.6.12-rc6
Automatic merge of 'misc-fixes' branch from
Automatic merge of rsync://www.parisc-linux.org/~jejb/git/scsi-for-linus-2.6
Automatic merge of rsync://rsync.kernel.org/.../davem/net-2.6
Merge of 'docs' branch from
Merge of master.kernel.org:/.../aegl/linux-2.6
Automatic merge of rsync://rsync.kernel.org/.../sfrench/cifs-2.6
Automatic merge of rsync://rsync.kernel.org/.../davem/net-2.6
Automatic merge of rsync://rsync.kernel.org/.../davem/sparc-2.6
Automatic merge of rsync://rsync.kernel.org/.../gregkh/usb-2.6
Automatic merge of rsync://rsync.kernel.org/.../gregkh/i2c-2.6
Automatic merge of rsync://rsync.kernel.org/.../gregkh/pci-2.6
Automatic merge of rsync://rsync.kernel.org/.../aegl/linux-2.6
Merge of rsync://rsync.kernel.org/.../davem/tg3-2.6
Automatic merge of 'misc-fixes' branch from
Automatic merge of 'for-linus' branch from
Automatic merge of rsync://rsync.kernel.org/.../hch/xfs-2.6
ide-cd: revert DMA mask test change
Automatic merge of rsync://rsync.kernel.org/.../davem/net-2.6
Automatic merge of 'for-linus' branch from
Merge of 'misc-fixes' branch from
Merge of rsync://rsync.kernel.org/.../davem/sparc-2.6
Merge of 'new-ids' branch from
Merge of 'for-linus' branch from

Liu Tao:
drivers/net/amd8111e.c: fix NAPI interrupt in poll

Lonnie Mendez:
USB: hid-core: add Earthmate lt-20 productid to blacklist table

Malli Chilakala:
e100: Driver version, white space, comments, device id
e100: Performance optimizations to e100 Tx Path
e100: Fix Wake on lan related issues
e100: Synchronize interface link state with poll routine
e100: Render e100 NAPI state machine
e100: Execute tx_timeout task outside interrupt context
ixgb: Driver version, white space, comments, device id
ixgb: Fixed msec_delay in osdep to use msleep
ixgb: Code optimization
ixgb: Remove hook for suspend, no power management
ixgb: Support for ethtool -d
ixgb: Fix EEPROM functions to be endian-aware
ixgb: Reset status in the Rx
ixgb: Mask RXO interrupt
ixgb: Change RDT write bump size to 4
ixgb: Do not set the RS bit on context descriptors
ixgb: Fix multi-cast packet count in statistics
e1000:Driver version,white space,comments,device id
e1000:Adjust flow control watermarks for Jumbo Frames
e1000:Fix Packet Buffer Allocation logic for 82547_rev_2
e1000:82573 specific code & packet split code
e1000: Modified e1000_clean: exit poll
e1000:Removed redundant statement in e1000_clean_tx_irq
e1000: Implement a workaround for 82546 errata 10
e1000: e1000 stops working after resume
e1000:Fix computation of netdev stats from controller stats counters
e1000: Dump information on Tx ring
e1000: Delay clean-up of last Tx packet
e1000: Fix kernel panic with 82541 LOM
e1000: Enable polling before enabling interrupts
e1000: MSI support for PCI-e adapters
e1000: Fix msec-delay definition to use msleep
e1000: made loopback test robust

Manfred Spraul:
forcedeth: Update error handling

Manu Abraham:
dvb: Small cleanup
dvb: Fix 22k tone control
dvb: Fix LNB power switching
dvb: Remove unnecessary casts
dvb: Fix Mini DiSEqC bug

Marcello Maggioni:
timeout at boottime with NEC3500A (and possibly others) when inserted a CD in it

Martin Schwidefsky:
s390: in_interrupt vs. in_atomic
s390: ptrace peek and poke

Matthias Urlichs:
USB: add Option Card driver

Michael Chan:
[TG3]: Fix bug in tg3_load_firmware_cpu
[TG3]: Add interrupt test
[TG3]: Add loopback test
[TG3]: Add memory test
[TG3]: Add register test
[TG3]: Add parameter to tg3_halt
[TG3]: Add link test
[TG3]: Add nvram test
[TG3]: Add basic selftest infrastructure
[BNX2]: New Broadcom gigabit network driver.

Michael Ellerman:
[NET]: Add is_multicast_ether_addr() in include/linux/etherdevice.h
iseries_veth: Cleanup skbs to prevent unregister_netdevice() hanging
iseries_veth: Don't leak skbs in RX path
iseries_veth: Set dev->trans_start so watchdog timer works right
iseries_veth: Don't send packets to LPARs which aren't up

Michal Schmidt:
forcedeth: netpoll support

NAKAMURA Kenta:
sata_sil: new ID 1002:437A for ATI IXP400

Nathan Lynch:
prom_find_machine_type typo breaks pSeries lpar boot
[SCSI] fix slab corruption during ipr probe

Nathan Scott:
[XFS] Fix directory inodes ioctl compat code, minor code consistency cleanups

Neil Horman:
ipmi build fix

Nick Piggin:
h8300 sleep problem

Oliver Korpilla:
x86_64: signal.c build fix

Paolo 'Blaisorblade' Giarrusso:
uml: remove jail mode + other leftovers
uml: fixlet for arch_prctl_skas
irq code: Add coherence test for PREEMPT_ACTIVE
uml: fix PREEMPT_ACTIVE
uml: stack dump fix
uml: split CONFIG_FRAME_POINTER from DEBUG_INFO
uml: add MOD_LICENSE to random driver
uml: add modversions support

Patrick McManus:
intelfb section fix

Paul Jackson:
cpuset exit NULL dereference fix

Paul Mackerras:
ppc64: actually call prom_send_capabilities

Paul Mundt:
sh: PREEMPT_ACTIVE fix

Paulo Marques:
USB: make MODALIAS code a bit smaller devices

Pavel Machek:
fix jumpy mouse cursor on console

Pete Zaitcev:
USB: Support multiply-LUN devices in ub

Peter Chubb:
[IA64] fix compilation warning in sys32_epoll_wait()
[IA64] Cleanup compile warnings for ski config
pcdp.c build fix

Phil Dibowitz:
USB Storage: Add unusual_devs for Trumpion Voice Recorder

Ping Cheng:
USB: add new wacom device to usb hid-core list

Pravin B. Shelar:
[IPV4]: Kill MULTIPATHHOLDROUTE flag.

Qu Fuping:
mpage_end_io_write() I/O error handling fix

Roland Dreier:
IB: fix endianness of path record MTU field
IB: fix potential ib_umad leak
IB: allow NULL sa_query callbacks

Roland McGrath:
i386: fix prevent_tail_call

Roman Kagan:
USB: update urb documentation

Roman Zippel:
flush icache in correct context

Rudolf Marek:
I2C: ALI1563 SMBus driver fix

Russ Anderson:
[IS64-SGI] Set Altix error handling features
[IA64-SGI] Make Altix SAL call to POD reentrant
[IA64-SGI] cpe interrupts are not being enabled.

Russell Cattelan:
[XFS] Fix a bug in xfs_iomap for extent handling of write cases

Scott Murray:
PCI Hotplug: more CPCI updates

Shaohua Li:
swsusp: ahd_dv_0 can't be stopped

Siddha, Suresh B:
x86: fix smp_num_siblings on buggy BIOSes

Simon Kelley:
atmel wireless

Stephen Hemminger:
[BRIDGE]: receive path optimization
[BRIDGE]: prevent bad forwarding table updates
[BRIDGE]: set features based on enslaved devices
[BRIDGE]: make dev->features unsigned
[BRIDGE]: features change notification
[PKT_SCHED] netem: allow random reordering (with fix)
[PKT_SCHED] netem: use only inner qdisc -- no private skbuff queue
[PKT_SCHED]: netem: reinsert for duplication
tlan: restore deleted module parameters.

Stephen Rothwell:
ppc64 iSeries: make virtual DVD-RAMs writable again
ppc64 iSeries: fix boot time setting
ppc64: fix initialisation of gettimeofday calculations

Steve French:
Merge with rsync://rsync.kernel.org/.../torvalds/linux-2.6.git
[CIFS] fix casts of unicode strings to match function definition
[CIFS] Fix oops in cifs_unlink. Caused in some cases when renaming over existing,
[CIFS] missing break needed to handle < when mount option "mapchars" specified

Stuart Hayes:
ide-scsi: kmap scatter/gather before doing PIO

Thomas Graf:
[PKT_SCHED]: Disable dsmark debugging messages by default
[PKT_SCHED]: make dsmark try using pfifo instead of noop while grafting
[PKT_SCHED]: Fix dsmark to count ignored indices while walking

Tony Luck:
[IA64] Use "PER_CPU" form of EXPORT macro
[IA64] initialize spinlock pfm_alt_install_check
Sync with Linus - rsync://rsync.kernel.org/.../torvalds/linux-2.6.git
[IA64] alternate perfmon handler
Merge with linus
Merge with /home/aegl/GIT/linus

Venkatesh Pallipadi:
cpufreq-stats driver documentation
cpufreq-stats driver updates

Vitaly Bordug:
ppc32: Support for 82xx PQII on-chip PCI bridge

Vojtech Pavlik:
Input: Fix fast scrolling scancodes in atkbd.c
input: Fix fast scrolling scancodes in atkbd.c
Input: Fix button mapping in joydev - BTN_TRIGGER was being
Input: Workaround for Sunrex K8561 IR Keyboard/Mouse. The mouse
Input: Only write the CTR in i8042 resume function. Reading it is
Input: Remove (now) unused variable in i8042.c
Input: Add a missing KERN_INFO message designation, fix behavior

Yoichi Yuasa:
serial: update NEC VR4100 series serial support

Zhang Yanmin:
[IA64] sys_mmap doesn't follow posix.1 when parameter len=0

Pavel Machek

unread,

Jun 6, 2005, 3:40:25 PM6/6/05

to

Hi!

> It's being uploaded right now, the git tree is already up-to-date, and by
> the time this hits the mailing list the mirroring of the tar-ball will
> hopefully be done too.
>
> And since Jeff wrote me a shortlog script for git, the easist way to tell
> what's new since -rc5 is to just do the shortlog and diffstat output.
> Network drivers, USB and CPU-freq stand out.
>
> And the good news is that people do seem to have taken my rumblings about
> calming down for 2.6.12 seriously. Let's hope that pans out, and I can
> release that one asap.. But give this a good beating first, and holler
> (again, if you must) about any issues you have,

> Pavel Machek:
> fix jumpy mouse cursor on console

This one was from Dmitry, and git logs know that:

author Pavel Machek <pa...@suse.cz> Fri, 27 May 2005 12:53:03 -0700
committer Linus Torvalds <torv...@ppc970.osdl.org> Sat, 28 May 2005 11:14:01 -0700

[PATCH] fix jumpy mouse cursor on console

Do not send empty events to gpm. (Keyboards are assumed to have scroll
wheel these days, that makes them part-mouse. That means typing on
keyboard generates empty mouse events).

From: Dmitry Torokhov <dtor...@ameritech.net>
Signed-off-by: Pavel Machek <pa...@suse.cz>
Signed-off-by: Andrew Morton <ak...@osdl.org>
Signed-off-by: Linus Torvalds <torv...@osdl.org>

...perhaps shortlog script needs some updating?
Pavel

Pavel Machek

unread,

Jun 6, 2005, 4:20:15 PM6/6/05

to

Hi!

> >...perhaps shortlog script needs some updating?
>
>

> In git-whatchanged, you are listed as the author:
>
>
> >diff-tree c1e4c8d3ee3300f363a52fd4cf3d90fdf5098f5a (from
> >8bd7f125e2f217c8aa3dff0
> >Author: Pavel Machek <pa...@suse.cz>
> >Date: Fri May 27 12:53:03 2005 -0700

> >
> > [PATCH] fix jumpy mouse cursor on console
> >
> > Do not send empty events to gpm. (Keyboards are assumed to have scroll
> > wheel these days, that makes them part-mouse. That means typing on
> > keyboard generates empty mouse events).
> >
> > From: Dmitry Torokhov <dtor...@ameritech.net>
> > Signed-off-by: Pavel Machek <pa...@suse.cz>
> > Signed-off-by: Andrew Morton <ak...@osdl.org>
> > Signed-off-by: Linus Torvalds <torv...@osdl.org>
>

> '^Author: ' is what git-shortlog looks at.

Linus, perhaps your scripts are doing something wrong? They should
have taken From in the description; or did I provide wrong changelog?

Linus Torvalds

unread,

Jun 6, 2005, 4:20:13 PM6/6/05

to

On Mon, 6 Jun 2005, Pavel Machek wrote:

> > Pavel Machek:
> > fix jumpy mouse cursor on console
>
> This one was from Dmitry, and git logs know that:

No it doesn't. Somebody sent me a patch in the wrong format..

> author Pavel Machek <pa...@suse.cz> Fri, 27 May 2005 12:53:03 -0700
> committer Linus Torvalds <torv...@ppc970.osdl.org> Sat, 28 May 2005 11:14:01 -0700
>
> [PATCH] fix jumpy mouse cursor on console
>
> Do not send empty events to gpm. (Keyboards are assumed to have scroll
> wheel these days, that makes them part-mouse. That means typing on
> keyboard generates empty mouse events).
>
> From: Dmitry Torokhov <dtor...@ameritech.net>
> Signed-off-by: Pavel Machek <pa...@suse.cz>
> Signed-off-by: Andrew Morton <ak...@osdl.org>
> Signed-off-by: Linus Torvalds <torv...@osdl.org>
>
> ...perhaps shortlog script needs some updating?

No, it isn't going to go through the body of the email.

The way that author attributions get done right is if the first line of
the body of the email has a "From: xyzzy <a...@xyz.com>" in it.

Linus

Pavel Machek

unread,

Jun 6, 2005, 4:30:23 PM6/6/05

to

Hi!

> > > fix jumpy mouse cursor on console
> >
> > This one was from Dmitry, and git logs know that:
>
> No it doesn't. Somebody sent me a patch in the wrong format..

Okay, that was probably me, but...

> > author Pavel Machek <pa...@suse.cz> Fri, 27 May 2005 12:53:03 -0700
> > committer Linus Torvalds <torv...@ppc970.osdl.org> Sat, 28 May 2005 11:14:01 -0700
> >
> > [PATCH] fix jumpy mouse cursor on console
> >
> > Do not send empty events to gpm. (Keyboards are assumed to have scroll
> > wheel these days, that makes them part-mouse. That means typing on
> > keyboard generates empty mouse events).
> >
> > From: Dmitry Torokhov <dtor...@ameritech.net>
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > Signed-off-by: Pavel Machek <pa...@suse.cz>
> > Signed-off-by: Andrew Morton <ak...@osdl.org>
> > Signed-off-by: Linus Torvalds <torv...@osdl.org>
> >
> > ...perhaps shortlog script needs some updating?
>
> No, it isn't going to go through the body of the email.
>
> The way that author attributions get done right is if the first line of
> the body of the email has a "From: xyzzy <a...@xyz.com>" in it.

There is "From: Dmitry..." in the changelog. Do your script move first
"From:" into author header and delete it from changelog? That would
explain it...
Pavel

Jeff Garzik

unread,

Jun 6, 2005, 4:40:13 PM6/6/05

to

In git-whatchanged, you are listed as the author:

> diff-tree c1e4c8d3ee3300f363a52fd4cf3d90fdf5098f5a (from 8bd7f125e2f217c8aa3dff0
> Author: Pavel Machek <pa...@suse.cz>

> Date: Fri May 27 12:53:03 2005 -0700

>
> [PATCH] fix jumpy mouse cursor on console
>
> Do not send empty events to gpm. (Keyboards are assumed to have scroll
> wheel these days, that makes them part-mouse. That means typing on
> keyboard generates empty mouse events).
>
> From: Dmitry Torokhov <dtor...@ameritech.net>
> Signed-off-by: Pavel Machek <pa...@suse.cz>
> Signed-off-by: Andrew Morton <ak...@osdl.org>
> Signed-off-by: Linus Torvalds <torv...@osdl.org>

'^Author: ' is what git-shortlog looks at.

Regards,

Jeff

Linus Torvalds

unread,

Jun 6, 2005, 5:20:12 PM6/6/05

to

On Mon, 6 Jun 2005, Pavel Machek wrote:
>

> Linus, perhaps your scripts are doing something wrong? They should
> have taken From in the description; or did I provide wrong changelog?

My scripts definitely do the expected thing.

In git, the author is always in the fixed header, and you never look for
it anywhere else. However, in order for the author to _get_ there in the
first place, the person who commits the thing needs to haev the author
info.

In this case it was me, and I get the author information from the email
when I commit an emailed patch. I take it from the first line of the body
if that one is a valid "From:" line, and otherwise I fall back to taking
it from the headers of the email.

So in this case you got tagged, either because the patch came through
Andrew (it has his sign-off) and _he_ sent the email but incorrectly had
you as the "From:" person, or alternatively because you sent the email and
took Andrew's sign-off but didn't put the "From:" in the right spot.

Linus

Linus Torvalds

unread,

Jun 6, 2005, 5:20:11 PM6/6/05

to

On Mon, 6 Jun 2005, Pavel Machek wrote:
>

> There is "From: Dmitry..." in the changelog. Do your script move first
> "From:" into author header and delete it from changelog? That would
> explain it...

Yes. But note how it doesn't even take the "first" From: line, it
literally takes the From: line _only_ if that line is the first line in
the email body.

See the "git-tools" archive if you want to see all the ugly details (start
from http://www.kernel.org/git)

Linus

Pavel Machek

unread,

Jun 6, 2005, 5:30:16 PM6/6/05

to

Hi!

> > There is "From: Dmitry..." in the changelog. Do your script move first
> > "From:" into author header and delete it from changelog? That would
> > explain it...
>
> Yes. But note how it doesn't even take the "first" From: line, it
> literally takes the From: line _only_ if that line is the first line in
> the email body.
>
> See the "git-tools" archive if you want to see all the ugly details (start
> from http://www.kernel.org/git)

Aha, okay, it was going to you through andrew, and it was me who
posted the changelog in form

Description

From: XXX
Signed-off-by: YYY

I thought you are taking "first From: in the body", not "From: only if
it is first line in the body". [Could you perhaps modify your scripts
to take "first From: in the body"? It seems logical to put From "near"
Signed-of-by: lines...
Pavel

Linus Torvalds

unread,

Jun 6, 2005, 5:40:13 PM6/6/05

to

On Mon, 6 Jun 2005, Pavel Machek wrote:
>

> I thought you are taking "first From: in the body", not "From: only if
> it is first line in the body". [Could you perhaps modify your scripts
> to take "first From: in the body"? It seems logical to put From "near"
> Signed-of-by: lines...

I really don't want to, for a number of reasons. Most notably because I
don't want to mix things up with the sign-off, because authorship and
sign-off are really separate things (sign-offs accumulate, authorship
stays), but also because it's not entirely unambiguous to parse these
things. With the "first line only" rule, it ends up being pretty clear
what's going on when the script suddenly ate one line..

Linus

Pavel Machek

unread,

Jun 6, 2005, 5:50:11 PM6/6/05

to

Hi!

> > I thought you are taking "first From: in the body", not "From: only if
> > it is first line in the body". [Could you perhaps modify your scripts
> > to take "first From: in the body"? It seems logical to put From "near"
> > Signed-of-by: lines...
>
> I really don't want to, for a number of reasons. Most notably because I
> don't want to mix things up with the sign-off, because authorship and
> sign-off are really separate things (sign-offs accumulate, authorship
> stays), but also because it's not entirely unambiguous to parse these
> things. With the "first line only" rule, it ends up being pretty clear
> what's going on when the script suddenly ate one line..

Okay, I see. I'm little afraid that during forwards blank line will be
inserted before "From: " and break this, but lets see how it works.

Pavel

Bill Davidsen

unread,

Jun 6, 2005, 6:00:22 PM6/6/05

to

Linus Torvalds wrote:
>
> On Mon, 6 Jun 2005, Pavel Machek wrote:
>
>>Linus, perhaps your scripts are doing something wrong? They should
>>have taken From in the description; or did I provide wrong changelog?
>
>
> My scripts definitely do the expected thing.
>
> In git, the author is always in the fixed header, and you never look for
> it anywhere else. However, in order for the author to _get_ there in the
> first place, the person who commits the thing needs to haev the author
> info.
>
> In this case it was me, and I get the author information from the email
> when I commit an emailed patch. I take it from the first line of the body
> if that one is a valid "From:" line, and otherwise I fall back to taking
> it from the headers of the email.
>
> So in this case you got tagged, either because the patch came through
> Andrew (it has his sign-off) and _he_ sent the email but incorrectly had
> you as the "From:" person, or alternatively because you sent the email and
> took Andrew's sign-off but didn't put the "From:" in the right spot.

Any process which only works when multiple people do everything
correctly is not going to be robust. Perhaps you want to use the first
"signed-off-by" line or some such, rather than relying on mail headers?

Linus Torvalds

unread,

Jun 6, 2005, 6:00:17 PM6/6/05

to

On Mon, 6 Jun 2005, Pavel Machek wrote:
>

> Okay, I see. I'm little afraid that during forwards blank line will be
> inserted before "From: " and break this, but lets see how it works.

Oh, I skip blank lines (and that means any line that is "whitespace only",
ie tabs/spaces etc won't confuse the scripts), so at least it's not _that_
subtle.

Linus

Jesper Juhl

unread,

Jun 6, 2005, 6:10:19 PM6/6/05

to

On 6/6/05, Linus Torvalds <torv...@osdl.org> wrote:
>
>
> On Mon, 6 Jun 2005, Pavel Machek wrote:
> >
> > There is "From: Dmitry..." in the changelog. Do your script move first
> > "From:" into author header and delete it from changelog? That would
> > explain it...
>
> Yes. But note how it doesn't even take the "first" From: line, it
> literally takes the From: line _only_ if that line is the first line in
> the email body.
>

A lot of times I see mails getting forwarded to you/Andrew/other
maintainer by someone without adding a From: or other indication of
who was the original author, but in almost all cases the original
author is the one listed as the first Signed-off-by: since authors are
the first to sign off on a patch, so, wouldn't it make more sense to
pick the author like this ;

1) If there's a "From:" at the start of the email, use that (note: a
lot of times this actually breaks since From: is often set by
maintainers to the person who forwarded the mail, not the actual
author, but that's a matter of educating maintainers).

2) if there's no "From:" in the mail body, pick first "signed-off-by:"

3) Fall back on email headers.

--
Jesper Juhl <jespe...@gmail.com>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html

Russell King

unread,

Jun 6, 2005, 6:20:10 PM6/6/05

to

On Mon, Jun 06, 2005 at 11:57:00PM +0200, Jesper Juhl wrote:
> On 6/6/05, Linus Torvalds <torv...@osdl.org> wrote:
> >
> >
> > On Mon, 6 Jun 2005, Pavel Machek wrote:
> > >
> > > There is "From: Dmitry..." in the changelog. Do your script move first
> > > "From:" into author header and delete it from changelog? That would
> > > explain it...
> >
> > Yes. But note how it doesn't even take the "first" From: line, it
> > literally takes the From: line _only_ if that line is the first line in
> > the email body.
> >
>
> A lot of times I see mails getting forwarded to you/Andrew/other
> maintainer by someone without adding a From: or other indication of
> who was the original author, but in almost all cases the original
> author is the one listed as the first Signed-off-by: since authors are
> the first to sign off on a patch, so, wouldn't it make more sense to
> pick the author like this ;

Not necessarily. Re-read what Signed-off-by: is all about and who
may provide that line. You should find that the first Signed-off-by:
line may not be the author themselves, but someone else who is able
to satisfy our requirements.

I think people will just have to accept that there's no way to _always_
_automatically_ get the proper author for every patch. (and that calling
it author in git was probably the first mistake - we never had these
issues with BK which didn't specifically indentify anything as being
the "author" as such.)

--
Russell King
Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/
maintainer of: 2.6 Serial core

Voluspa

unread,

Jun 7, 2005, 2:20:14 AM6/7/05

to

CC arch/x86_64/kernel/irq.o
CC arch/x86_64/kernel/ptrace.o
arch/x86_64/kernel/ptrace.c: In function `putreg':
arch/x86_64/kernel/ptrace.c:285: error: duplicate case value
arch/x86_64/kernel/ptrace.c:280: error: previously used here
make[1]: *** [arch/x86_64/kernel/ptrace.o] Error 1
make: *** [arch/x86_64/kernel] Error 2

real 0m5.524s
user 0m4.227s
sys 0m0.724s

Mvh
Mats Johannesson
--

Ingo Molnar

unread,

Jun 7, 2005, 2:30:14 AM6/7/05

to

* Voluspa <lis...@telia.com> wrote:

> CC arch/x86_64/kernel/irq.o
> CC arch/x86_64/kernel/ptrace.o
> arch/x86_64/kernel/ptrace.c: In function `putreg':
> arch/x86_64/kernel/ptrace.c:285: error: duplicate case value
> arch/x86_64/kernel/ptrace.c:280: error: previously used here
> make[1]: *** [arch/x86_64/kernel/ptrace.o] Error 1
> make: *** [arch/x86_64/kernel] Error 2

builds fine here - and i cannot see how those case values could be
duplicate. Are you sure your build is completely clean?

Ingo

Voluspa

unread,

Jun 7, 2005, 2:40:11 AM6/7/05

to

On Tue, 7 Jun 2005 08:18:31 +0200 Ingo Molnar wrote:
>
> * Voluspa <lis...@telia.com> wrote:
>
> > CC arch/x86_64/kernel/irq.o
> > CC arch/x86_64/kernel/ptrace.o
> > arch/x86_64/kernel/ptrace.c: In function `putreg':
> > arch/x86_64/kernel/ptrace.c:285: error: duplicate case value
> > arch/x86_64/kernel/ptrace.c:280: error: previously used here
> > make[1]: *** [arch/x86_64/kernel/ptrace.o] Error 1
> > make: *** [arch/x86_64/kernel] Error 2
>
> builds fine here - and i cannot see how those case values could be
> duplicate. Are you sure your build is completely clean?

Ah, sorry about the noise... I've been away from kernel testing too
long. I patched a 2.6.11.11 tree without noticing all the rejects (this
new machine is fast). But from what I remember, it was decided to do
the -rc patches against the latest stable codebase, in this case .11
Shrug.

Mvh
Mats Johannesson
--

Grant Coady

unread,

Jun 7, 2005, 3:10:22 AM6/7/05

to

On Tue, 7 Jun 2005 08:37:31 +0200, Voluspa <lis...@telia.com> wrote:
>Ah, sorry about the noise... I've been away from kernel testing too
>long. I patched a 2.6.11.11 tree without noticing all the rejects (this
>new machine is fast). But from what I remember, it was decided to do
>the -rc patches against the latest stable codebase, in this case .11
>Shrug.

I dunno what the change was, patch didn't apply cleanly to 2.6.11,
(no idea if bad .bz2, finger trouble), so I download whole thing
instead, now running on three x86 boxen.
--Grant.

Matthias Andree

unread,

Jun 7, 2005, 4:30:21 AM6/7/05

to

On Tue, 07 Jun 2005, Voluspa wrote:

> Ah, sorry about the noise... I've been away from kernel testing too
> long. I patched a 2.6.11.11 tree without noticing all the rejects (this
> new machine is fast). But from what I remember, it was decided to do
> the -rc patches against the latest stable codebase, in this case .11
> Shrug.

Try adding "-s" to the patch command then.

--
Matthias Andree

Voluspa

unread,

Jun 7, 2005, 4:50:11 AM6/7/05

to

On Tue, 7 Jun 2005 17:44:00 +1000 Dave Airlie wrote:

> > On Tue, 7 Jun 2005 08:37:31 +0200, Voluspa wrote:
> > >Ah, sorry about the noise... I've been away from kernel testing too
> > >long. I patched a 2.6.11.11 tree without noticing all the rejects
> > >(this new machine is fast). But from what I remember, it was
> > >decided to do the -rc patches against the latest stable codebase,
> > >in this case .11 Shrug.

[...]
> It was explicitly stated by Linus way back 2.6.8.1 time that
> subsequent patches are against the base of the previous release, so
> -rcs are against 2.6.x not 2.6.x.y.. which all makes great sense if
> development is happening in parallel... I'm not sure I've ever heard
> anything else stated to oppose this, but apparently some people have..

Yes, I remember that episode well. But times and policy change, don't
they? When the "stable" dot-releases were introduced, there was another
discussion, albeit brief, of where the base should lie. And my memory
must have really left for an early vacation since the .x.y decision came
so easily and quickly to my mind.

Anyway, the -rc6 applied cleanly to a 2.6.11 kernel and my uptime is now
27 minutes.

Issue: The new "conservative" qpufreq governor doesn't work. Stuck on
full blast. Loading the "ondemand" and trying to echo that choice into
the scaling_governor file doesn't work ("cannot overwrite existing
file"). Must reboot since the "conservative" don't want to be unloaded
(is in use). Guess I could have unloaded the whole ACPI module system in
correct order, but a reboot was quicker --> 20 seconds...

Ask for more info about my setup if someone would like to work on this.

Meta-issue: Anyone not using hotplug/udev/whatever will trip on the
kbd-now-being-another-mouse. Should result in plenty of user bugreports
around the lists unless the fact is posted somewhere prominently. I took
a brief tour with the -rc5 (downloaded a whole kernel) and just shook my
head when the mice in X became totally unruly. Didn't know if I
should laugh or cry, that's how funny the cursor behaved. I've now
bumped up both the internal touchpad and the usb mouse one notch higher
in the /dev/input/mouseX list and all is well.

Mvh
Mats Johannesson
--

Dave Airlie

unread,

Jun 7, 2005, 3:50:17 AM6/7/05

to

> On Tue, 7 Jun 2005 08:37:31 +0200, Voluspa <lis...@telia.com> wrote:
> >Ah, sorry about the noise... I've been away from kernel testing too
> >long. I patched a 2.6.11.11 tree without noticing all the rejects (this
> >new machine is fast). But from what I remember, it was decided to do
> >the -rc patches against the latest stable codebase, in this case .11
> >Shrug.
> I dunno what the change was, patch didn't apply cleanly to 2.6.11,
> (no idea if bad .bz2, finger trouble), so I download whole thing
> instead, now running on three x86 boxen.

It was explicitly stated by Linus way back 2.6.8.1 time that

subsequent patches are against the base of the previous release, so
-rcs are against 2.6.x not 2.6.x.y.. which all makes great sense if
development is happening in parallel... I'm not sure I've ever heard
anything else stated to oppose this, but apparently some people have..

Dave.

Johannes Stezenbach

unread,

Jun 7, 2005, 5:20:13 AM6/7/05

to

Hype-threading stopped working for me (probably due to
me not enabling ACPI). dmesg output and .config attached.
-rc5 worked fine. The board is an Asus P4P800-Deluxe.

dmesg: WARNING: 1 siblings found for CPU0, should be 2

$ cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 2
model name : Intel(R) Pentium(R) 4 CPU 2.60GHz
stepping : 9
cpu MHz : 2606.438
cache size : 512 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid xtpr
bogomips : 5144.57

Johannes

dmesg

.config

Erik Mouw

unread,

Jun 7, 2005, 9:10:10 AM6/7/05

to

On Mon, Jun 06, 2005 at 11:08:25AM -0700, Linus Torvalds wrote:

Over here the script can get the correct information from git
branches:

> Jeff Garzik:
> Automatic merge of /spare/repo/netdev-2.6 branch r8169-fix

But for your own changes it seems to fail:

> Linus Torvalds:
> Linux 2.6.12-rc6
> Automatic merge of 'misc-fixes' branch from

... from what?

> Automatic merge of rsync://www.parisc-linux.org/~jejb/git/scsi-for-linus-2.6
> Automatic merge of rsync://rsync.kernel.org/.../davem/net-2.6

And this again works.

> Merge of 'docs' branch from
> Merge of master.kernel.org:/.../aegl/linux-2.6
> Automatic merge of rsync://rsync.kernel.org/.../sfrench/cifs-2.6
> Automatic merge of rsync://rsync.kernel.org/.../davem/net-2.6
> Automatic merge of rsync://rsync.kernel.org/.../davem/sparc-2.6
> Automatic merge of rsync://rsync.kernel.org/.../gregkh/usb-2.6
> Automatic merge of rsync://rsync.kernel.org/.../gregkh/i2c-2.6
> Automatic merge of rsync://rsync.kernel.org/.../gregkh/pci-2.6
> Automatic merge of rsync://rsync.kernel.org/.../aegl/linux-2.6
> Merge of rsync://rsync.kernel.org/.../davem/tg3-2.6
> Automatic merge of 'misc-fixes' branch from
> Automatic merge of 'for-linus' branch from
> Automatic merge of rsync://rsync.kernel.org/.../hch/xfs-2.6
> ide-cd: revert DMA mask test change
> Automatic merge of rsync://rsync.kernel.org/.../davem/net-2.6
> Automatic merge of 'for-linus' branch from
> Merge of 'misc-fixes' branch from
> Merge of rsync://rsync.kernel.org/.../davem/sparc-2.6
> Merge of 'new-ids' branch from
> Merge of 'for-linus' branch from

I don't see what's going wrong, maybe you can figure out.

Erik

--
+-- Erik Mouw -- www.harddisk-recovery.com -- +31 70 370 12 90 --
| Lab address: Delftechpark 26, 2628 XH, Delft, The Netherlands

Linus Torvalds

unread,

Jun 7, 2005, 11:50:15 AM6/7/05

to

On Tue, 7 Jun 2005, Erik Mouw wrote:
>
> Over here the script can get the correct information from git
> branches:
>
> > Jeff Garzik:
> > Automatic merge of /spare/repo/netdev-2.6 branch r8169-fix
>
> But for your own changes it seems to fail:
>
> > Linus Torvalds:
> > Linux 2.6.12-rc6
> > Automatic merge of 'misc-fixes' branch from
>
> ... from what?

Yeah, I guess I need to redo my merge messages. Or alternatively, I should
just remove merges from the shortlog.

The merge message that goes along with that shortlog entry is

Author: Linus Torvalds <torv...@ppc970.osdl.org>
Date: Sat Jun 4 08:18:39 2005 -0700

Automatic merge of 'misc-fixes' branch from

rsync://rsync.kernel.org/pub/scm/linux/kernel/git/jgarzik/netdev-2.6

which is pretty readable in the long format, but causes the shortlog to
pick up just the partial (largely uninteresing) first line.

Removing merges from the shortlog is actually likely the _right_ thing to
do, since we already miss a lot of merges: any truly trivial merge (ie no
parallellism) is invisible anyway, except that the committer changed.
Besides, it's what the old BK changelogs did.

So I guess I'll leave the merge message as is (unless somebody can suggest
a more readable format), and just update my release scripts to not include
merge messages.

(Using Jeff's syntax for merge messages isn't very good, since with remote
repositories the names of the repos get so long that the message gets
unwieldly..)

> > Automatic merge of rsync://www.parisc-linux.org/~jejb/git/scsi-for-linus-2.6
> > Automatic merge of rsync://rsync.kernel.org/.../davem/net-2.6
>
> And this again works.

That's because they don't have branches - when I take the HEAD of the
repostitory, the changelog entry ends up being just a one-liner again.

Do a "git log" if you have the git tree (or just look at the full
ChangeLog, which is just the output of that), you'll see what's up.

(Btw, Jeff, I think the git-shortlog script is slightly buggered, it
doesn't do the nice word-wrap, and it _only_ takes the first line, even
from a multi-line header. I think it should stop at the first empty line
in the commit, not just take the first one).

Linus

Erik Mouw

unread,

Jun 7, 2005, 1:20:27 PM6/7/05

to

On Tue, Jun 07, 2005 at 08:28:45AM -0700, Linus Torvalds wrote:
> On Tue, 7 Jun 2005, Erik Mouw wrote:
> > > Linus Torvalds:
> > > Linux 2.6.12-rc6
> > > Automatic merge of 'misc-fixes' branch from
> >
> > ... from what?
>
> Yeah, I guess I need to redo my merge messages. Or alternatively, I should
> just remove merges from the shortlog.
>
> The merge message that goes along with that shortlog entry is
>
> Author: Linus Torvalds <torv...@ppc970.osdl.org>
> Date: Sat Jun 4 08:18:39 2005 -0700
>
> Automatic merge of 'misc-fixes' branch from
>
> rsync://rsync.kernel.org/pub/scm/linux/kernel/git/jgarzik/netdev-2.6
>
> which is pretty readable in the long format, but causes the shortlog to
> pick up just the partial (largely uninteresing) first line.

Fair enough.

> Removing merges from the shortlog is actually likely the _right_ thing to
> do, since we already miss a lot of merges: any truly trivial merge (ie no
> parallellism) is invisible anyway, except that the committer changed.
> Besides, it's what the old BK changelogs did.
>
> So I guess I'll leave the merge message as is (unless somebody can suggest
> a more readable format), and just update my release scripts to not include
> merge messages.

Well, it's nice to know which repos you pulled from for a particular
release, that makes it easier to figure out what's wrong if/when
something breaks.

> > > Automatic merge of rsync://www.parisc-linux.org/~jejb/git/scsi-for-linus-2.6
> > > Automatic merge of rsync://rsync.kernel.org/.../davem/net-2.6
> >
> > And this again works.
>
> That's because they don't have branches - when I take the HEAD of the
> repostitory, the changelog entry ends up being just a one-liner again.

Isn't the easy fix to put the long changelog information on a single
line?

Erik
[blissfully ignorant about git internals]

--
+-- Erik Mouw -- www.harddisk-recovery.nl -- 0800 220 20 20 --
| Eigen lab: Delftechpark 26, 2628 XH, Delft, Nederland
| Files foetsie, bestanden kwijt, alle data weg?!
| Blijf kalm en neem contact op met Harddisk-recovery.nl!

Linus Torvalds

unread,

Jun 7, 2005, 1:30:15 PM6/7/05

to

On Tue, 7 Jun 2005, Erik Mouw wrote:
>
> Well, it's nice to know which repos you pulled from for a particular
> release, that makes it easier to figure out what's wrong if/when
> something breaks.

Well, the thing is, you don't actually get a full list - you at best get a
_partial_ list of the merges I do. In fact, it's almost always a proper
subset, with (wild wag) 10% of the merges never showing up.

If a repo has already merged with me on the far end, my merge will be just
a "fast-forward to what they did", which won't show up as a separate
commit.

Similarly, if somebody is on a different time-zone from me and just
happens to do his work while I'm sleeping, and I pull the tree first thing
in the morning, there won't be any overlapping development and again it's
just a fast-forward. So no merges at _all_ may be showing on either side
in that case..

This very much does happen, exactly because we have a "merge often"

> Isn't the easy fix to put the long changelog information on a single
> line?

Sure. It's just a shell script that generates the message:

merge_name=$(echo "$1" | sed 's:\.git/*$::')
if [ "$2" ]
then
merge_name="'$2' branch from

$merge_name"
merge_head="refs/heads/$2"
fi

but the reason I do it on multiple lines is that that looks better in the
long format, I think.

But maybe the short format is the one I should prioritize, since that's
the one that also shows up in the summary line for the graphical tools
(webgit and gitk).

Linus

Jean Delvare

unread,

Jun 7, 2005, 2:50:14 PM6/7/05

to

Hi all,

> Alan Cox:
> remove non-cleanroom pwc driver compression

This one triggers a compilation warning. Proposed fix:

Signed-off-by: Jean Delvare <kh...@linux-fr.org>

drivers/usb/media/pwc/pwc-uncompress.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)

--- linux-2.6.12-rc6.orig/drivers/usb/media/pwc/pwc-uncompress.c 2005-06-07 07:14:50.000000000 +0200
+++ linux-2.6.12-rc6/drivers/usb/media/pwc/pwc-uncompress.c 2005-06-07 20:35:12.000000000 +0200
@@ -120,7 +120,6 @@

switch (pdev->type)
{
-#if 0
case 675:
case 680:
case 690:
@@ -128,15 +127,16 @@
case 730:
case 740:
case 750:
+#if 0
pwc_dec23_decompress(&pdev->image, &pdev->view, &pdev->offset,
yuv, image,
flags,
pdev->decompress_data, pdev->vbandlength);
break;
+#endif
case 645:
case 646:
/* TODO & FIXME */
-#endif
return -ENXIO; /* No such device or address: missing decompressor */
break;
}

--
Jean Delvare

Paul Mackerras

unread,

Jun 7, 2005, 10:10:08 PM6/7/05

to

Linus Torvalds writes:

> The merge message that goes along with that shortlog entry is
>
> Author: Linus Torvalds <torv...@ppc970.osdl.org>
> Date: Sat Jun 4 08:18:39 2005 -0700
>
> Automatic merge of 'misc-fixes' branch from
>
> rsync://rsync.kernel.org/pub/scm/linux/kernel/git/jgarzik/netdev-2.6
>
> which is pretty readable in the long format, but causes the shortlog to
> pick up just the partial (largely uninteresing) first line.

This also affects gitk, which takes the first line of the commit
message as the headline. I could make gitk take the first paragraph
(i.e. until the first blank line) as the headline but even that
wouldn't help since you put a blank line between the "Automatic merge"
line and the actual URL. Could you leave out that blank line in
future, or do you have a better suggestion?

Paul.

Johannes Stezenbach

unread,

Jun 8, 2005, 7:20:25 AM6/8/05

to

Johannes Stezenbach wrote:
> Hype-threading stopped working for me (probably due to
> me not enabling ACPI). dmesg output and .config attached.
> -rc5 worked fine. The board is an Asus P4P800-Deluxe.
>
> dmesg: WARNING: 1 siblings found for CPU0, should be 2

Indeed SMT works fine if I enable ACPI.
Is SMT without ACPI not supported?

Johannes

Brice Goglin

unread,

Jun 8, 2005, 7:30:56 AM6/8/05

to

Johannes Stezenbach a écrit :

> Johannes Stezenbach wrote:
>
>>Hype-threading stopped working for me (probably due to
>>me not enabling ACPI). dmesg output and .config attached.
>>-rc5 worked fine. The board is an Asus P4P800-Deluxe.
>>
>>dmesg: WARNING: 1 siblings found for CPU0, should be 2
>
>
> Indeed SMT works fine if I enable ACPI.
> Is SMT without ACPI not supported?
>
> Johannes

You can pass acpi=ht into the kernel command line to disable
ACPI except the minimum required to get HT support.

From Documentation/kernel-parameters.txt:
acpi= [HW,ACPI] Advanced Configuration and Power Interface
...
ht -- run only enough ACPI to enable Hyper Threading

Brice

Johannes Stezenbach

unread,

Jun 8, 2005, 7:50:12 AM6/8/05

to

Brice Goglin wrote:
> Johannes Stezenbach a écrit :
> > Johannes Stezenbach wrote:
> >
> >>Hype-threading stopped working for me (probably due to
> >>me not enabling ACPI). dmesg output and .config attached.
> >>-rc5 worked fine. The board is an Asus P4P800-Deluxe.
> >>
> >>dmesg: WARNING: 1 siblings found for CPU0, should be 2
> >
> >
> > Indeed SMT works fine if I enable ACPI.
> > Is SMT without ACPI not supported?
>

> You can pass acpi=ht into the kernel command line to disable
> ACPI except the minimum required to get HT support.

That's nice, but I was thinking along the lines of:

diff -ur linux-2.6.12-rc6.orig/arch/i386/Kconfig linux-2.6.12-rc6/arch/i386/Kconfig
--- linux-2.6.12-rc6.orig/arch/i386/Kconfig 2005-06-06 23:16:27.000000000 +0200
+++ linux-2.6.12-rc6/arch/i386/Kconfig 2005-06-08 13:35:08.000000000 +0200
@@ -503,7 +503,7 @@

config SCHED_SMT
bool "SMT (Hyperthreading) scheduler support"
- depends on SMP
+ depends on SMP && ACPI
default off
help
SMT scheduler support improves the CPU scheduler's decision making

Comments? Is this intended?

Johannes

Jakob Oestergaard

unread,

Jun 8, 2005, 9:50:08 AM6/8/05

to

On Mon, Jun 06, 2005 at 11:08:25AM -0700, Linus Torvalds wrote:
>

> It's being uploaded right now, the git tree is already up-to-date, and by
> the time this hits the mailing list the mirroring of the tar-ball will
> hopefully be done too.
...

Guys, here's a (yes I know, fairly vague) error report - I'm just
beginning to narrow this down. Any suggestions would be greatly
appreciated.

Dual opteron (Sun Fire v20z), 2G RAM, untainted x86_64 kernels (no
preempt or other fancy stuff, just plain and simple K8 NUMA ACPI
configs);

Problem with 2.6.11.11: Machine hangs during backup (amanda) - right
after the estimates are done the box will freeze hard, nothing on
console.

Tried 2.6.12-rc6 in the hope that it would at least print out a panic or
oops for me: Haven't tried the backup yet, but I got the following in
the logs which I've never seen before:

gcc[18818]: segfault at 00000000332c8e54 rip 000000005555a0a8 rsp 00000000ffffba4c error 6
klogd[3116]: segfault at 000000003ef1b373 rip 000000000804a3b3 rsp 00000000ffffbbe0 error 4
gcc[18857]: segfault at 00000000332c8e54 rip 000000005555a0a8 rsp 00000000ffffbedc error 6
gcc[19049]: segfault at 00000000332c8e54 rip 000000005555a0a8 rsp 00000000ffffaefc error 6
gcc[19097]: segfault at 00000000332c8e54 rip 000000005555a0a8 rsp 00000000ffffb1fc error 6
gcc[19251]: segfault at 00000000332c8e54 rip 000000005555a0a8 rsp 00000000ffffc9dc error 6
gcc[19272]: segfault at 00000000332c8e54 rip 000000005555a0a8 rsp 00000000ffffb1bc error 6
gcc[1569]: segfault at 00000000332c8e54 rip 000000005555a0a8 rsp 00000000ffffb56c error 6
gcc[16458]: segfault at 00000000332c8e54 rip 000000005555a0a8 rsp 00000000ffffc9ac error 6

So, with 2.6.12-rc6 all of the sudden klogd and gcc starts
segfaulting...

No ECC errors or other anormalities reported via the service processor.

Anything you want me to try? Need more information?

--

/ jakob

Nicolas Pitre

unread,

Jun 8, 2005, 12:40:12 PM6/8/05

to

On Tue, 7 Jun 2005, Linus Torvalds wrote:

> (Btw, Jeff, I think the git-shortlog script is slightly buggered, it
> doesn't do the nice word-wrap, and it _only_ takes the first line, even
> from a multi-line header. I think it should stop at the first empty line
> in the commit, not just take the first one).

Also if the lists of commits could be shown in chronological order (like
the bk shortlog script did) that would be nice. Right now the list is
reversed.

Nicolas

Linus Torvalds

unread,

Jun 8, 2005, 1:30:25 PM6/8/05

to

On Wed, 8 Jun 2005, Paul Mackerras wrote:
>
> This also affects gitk, which takes the first line of the commit
> message as the headline. I could make gitk take the first paragraph
> (i.e. until the first blank line) as the headline but even that
> wouldn't help since you put a blank line between the "Automatic merge"
> line and the actual URL. Could you leave out that blank line in
> future, or do you have a better suggestion?

I'll just make future messages much denser. They should now be just

Merge repo

or

Merge 'name' branch of <repo>

(and I may decide to replace "branch of" with "from" to make it even
denser.)

Linus

Greg KH

unread,

Jun 9, 2005, 4:20:10 AM6/9/05

to

On Tue, Jun 07, 2005 at 08:47:33PM +0200, Jean Delvare wrote:
> Hi all,
>
> > Alan Cox:
> > remove non-cleanroom pwc driver compression
>
> This one triggers a compilation warning. Proposed fix:

What warning? I don't see anything here...

thanks,

greg k-h

Jean Delvare

unread,

Jun 9, 2005, 12:00:14 PM6/9/05

to

Hi Greg, all,

> > This one triggers a compilation warning. Proposed fix:
>
> What warning? I don't see anything here...

drivers/usb/media/pwc/pwc-uncompress.c: In function `pwc_decompress':
drivers/usb/media/pwc/pwc-uncompress.c:140: warning: unreachable code at beginning of switch statement

This is gcc 3.3.4. Strange that you don't have it.

Thanks,
--
Jean Delvare

Greg KH

unread,

Jun 9, 2005, 12:40:15 PM6/9/05

to

On Thu, Jun 09, 2005 at 05:52:36PM +0200, Jean Delvare wrote:
> Hi Greg, all,
>
> > > This one triggers a compilation warning. Proposed fix:
> >
> > What warning? I don't see anything here...
>
> drivers/usb/media/pwc/pwc-uncompress.c: In function `pwc_decompress':
> drivers/usb/media/pwc/pwc-uncompress.c:140: warning: unreachable code at beginning of switch statement
>
> This is gcc 3.3.4. Strange that you don't have it.

gcc 3.4.4 here does not show that warning. Odd...

thanks,

greg k-h

Bill Davidsen

unread,

Jun 17, 2005, 1:50:09 PM6/17/05

to

I would think that you can't do HT without ACPI, so there's no point in
building in HT scheduling unless you can have HT.

Is that what you were asking? I was hoping someone else would comment.

Scheduling is getting harder and harder to get right... I have this
thought of a Beowolf cluster of NUMA machines, with each node being HT
multicore SMP. By "right" I meant "optimal," I'm sure that setup would
do something reasonable with current scheduling.

--
-bill davidsen (davi...@tmr.com)
"The secret to procrastination is to put things off until the
last possible moment - but no longer" -me

Johannes Stezenbach

unread,

Jun 18, 2005, 10:00:11 AM6/18/05

to

Bill Davidsen wrote:
> Johannes Stezenbach wrote:
> >Brice Goglin wrote:
> >>Johannes Stezenbach a écrit :

> >>>Indeed SMT works fine if I enable ACPI.
> >>>Is SMT without ACPI not supported?
> >>
> >>You can pass acpi=ht into the kernel command line to disable
> >>ACPI except the minimum required to get HT support.
> >
> >
> >That's nice, but I was thinking along the lines of:
> >
> >diff -ur linux-2.6.12-rc6.orig/arch/i386/Kconfig
> >linux-2.6.12-rc6/arch/i386/Kconfig
> >--- linux-2.6.12-rc6.orig/arch/i386/Kconfig 2005-06-06
> >23:16:27.000000000 +0200
> >+++ linux-2.6.12-rc6/arch/i386/Kconfig 2005-06-08
> >13:35:08.000000000 +0200
> >@@ -503,7 +503,7 @@
> >
> > config SCHED_SMT
> > bool "SMT (Hyperthreading) scheduler support"
> >- depends on SMP
> >+ depends on SMP && ACPI
> > default off
> > help
> > SMT scheduler support improves the CPU scheduler's decision making
> >
> >Comments? Is this intended?
>
> I would think that you can't do HT without ACPI, so there's no point in
> building in HT scheduling unless you can have HT.
>
> Is that what you were asking? I was hoping someone else would comment.

In 2.6.12-rc5 SMT worked without CONFIG_ACPI. (IIRC the kernel used some
minimal ACPI stuff anyway for CPU initialisation).

I don't use power management or other features of ACPI so I
had it disabled, and my build broke with 2.6.12-rc6.

Johannes

Johannes Stezenbach

unread,

Jun 18, 2005, 1:20:05 PM6/18/05

to

Johannes Stezenbach wrote:
> In 2.6.12-rc5 SMT worked without CONFIG_ACPI. (IIRC the kernel used some
> minimal ACPI stuff anyway for CPU initialisation).
>
> I don't use power management or other features of ACPI so I
> had it disabled, and my build broke with 2.6.12-rc6.

Correction: It built fine, but didn't recognize the second "CPU".