Ok, I appear to be without network connectivity at home at least over the
weekend, and master.kernel.org is going down for some maintenance this
afternoon, so here's my current tree.
Architecture updates, threading improvements, shm fix (the cause of
the Oracle problems), networking, scsi, modules, you name it, it's here.
Due to my lacking network connection over the weekend, I'd suggest
discussing issues on linux-kernel, since emailing them to me won't much
help ;/
The joys of switching ISPs..
Linus
-----
Summary of changes from v2.5.48 to v2.5.49
============================================
<[email protected]>:
o [EBTABLES]: Use correct base pointer in ebt_do_table
o add necessary #ifdefs to netfilter_bridge.h, vs 2.5.48
<[email protected]>:
o [SPARC]: Add missing iounmap to display7seg driver
<[email protected]>:
o [VLAN]: Quiet some printks and free devices/groups correctly
<[email protected]>:
o [SCH_GRED]: Array overflow fixes, found by Stanford checker
<[email protected]>:
o [IPV6]: Export ipv6_chk_addr
<[email protected]>:
o vicam.c
<[email protected]>:
o dcache usage cleanups
<[email protected]>:
o ADM8513 support added;
<[email protected]>:
o fix intermezzo compile
<[email protected]>:
o [SCTP] Support for Peer address parameters socket option
o [SCTP] udp-style connect support(non-blocking)
o [SCTP] Fix for sideeffect violation in sctp_sf_heartbeat()
o [SCTP] Blocking connect() support
o [SCTP] MSG_EOR support for recvmsg()
<[email protected]>:
o [IPSEC]: Fix double unlock in esp/ah
Adrian Bunk <[email protected]>:
o dv1394 devfs missing brace
o fix compile error in usb-serial.c
Alexey Kuznetsov <[email protected]>:
o [IPSEC]: Policy timeout and pfkey acquire fixes
Andi Kleen <[email protected]>:
o Some leftover nsec stat fixes (ADFS,AFS,CIFS)
Andrew Morton <[email protected]>:
o fix the build for egcs-1.1.2
o detect uninitialised per-cpu storage
o explicitly initialise kstat per-cpu storage
o misc
o shmdt bugfix
o radix-tree reinitialisation fix
o Add SMP barrier to ipc's grow_ary()
o reduce CPU cost in loop
o Expanded bad page handling
o Make inode_ops->setxattr value parameter const
o fix endian problem in ext3 htree code
o remove a warning from __block_write_full_page()
o ext2/ext3 Orlov directory accounting fix
o bootmem crash fix
o Fix busy-wait with writeback to large queues
o Remove mapping->vm_writeback
o strengthen the `incremental min' logic in the page
o handle zones which are full of unreclaimable pages
o no-buffer-head ext2 option
Andries E. Brouwer <[email protected]>:
o inode_mknod parameters
o compilation fix (tpyo)
o *_mknod prototype
o kill i_dev
o fcntl fix
Andy Grover <[email protected]>:
o ACPI: Fix compilation error when CONFIG_SOFTWARE_SUSPEND is not set
(Shawn Starr)
o ACPI: fix debug print levels, and use down() instead of
down_interruptible(), and some whitespace.
o ACPI: Interpreter fixes Fixed memory leak in method argument
resolution Fixed Index() operator to work properly with a target
operand Fixed attempted double delete in the Index() code Code size
improvements Improved debug/error messages and levels Fixed a
problem with premature deletion of a buffer object
o ACPI: Add ec_read and ec_write external functions Other ec.c
cleanups, too
Anton Blanchard <[email protected]>:
o disable old stat on ppc64
Arnaldo Carvalho de Melo <[email protected]>:
o tr: make CONFIG_TR depend on CONFIG_LLC=y
o wdt: fix up header cleanups: add include linux/interrupt.h
o hotplug: fix up header cleanups: add include linux/interrupts.h
o sound: more fixups for header cleanup: add include
<linux/interrupt.h>
o i2c: fix up header cleanups: add include <linux/interrupt.h>
o input: fix up header cleanups: add include <linux/interrupt.h>
o smbfs: fixup header cleanups: forward declare struct sock, add
include/uio.h
o scsi: fix up header cleanups: add include <linux/interrupt.h>
o hd: fix up header cleanup: add include <linux/interrupt.h>
o i2o: fix up header cleanups: add include <linux/interrupt.h>
o tcic: fix up header cleanups: add include <linux/interrupt.h>
o sound: fix up header cleanups: add include <linux/interrupt.h>
o net/core: export sk_send_sigurg, its needed by x25 when built as a
module
o net: fix up header cleanups: remove unneeded sched.h include
o drivers/net: fix up header cleanup: remove unneeded sched.h
includes
o o drivers/net/hamradio: fix up header cleanups: remove uneeded
sched.h includes
o o drivers/net/wan/lmc: fix up header cleanups: remove uneeded
sched.h includes
Art Haas <[email protected]>:
o C99 initializer for fs/afs/inode.c
o C99 initializers for drivers/serial
o C99 initializer for drivers/zorro/proc.c
o C99 initializers for drivers/message/i2o/i2o_config.c
o C99 initializers for drivers/pci
o C99 initiailzers for fs/intermezzo
o C99 initializers for drivers/pcmcia
o C99 initializers for drivers/ieee1394
o C99 initializers for drivers/input
o C99 initializers for drivers/block/cciss_scsi.c
o C99 initializers for drivers/scsi
o C99 initializers for drivers/cdrom
o C99 initializers for drivers/i2c
o C99 initializers for drivers/parisc
o C99 initializers for drivers/parport
o C99 initializers for drivers/pnp
o C99 initializers for drivers/sgi
o C99 initializers for drivers/s390
o C99 initializers for drivers/tc
o C99 initializers for drivers/telephony
o fix C99 initializers fix
o C99 initializer for init/initramfs.c
o C99 initializer for drivers/parisc/ccio-dma.c
o C99 initializer for fs/ext2/dir.c
o C99 initializer for fs/sysv/super.c
Ben Fennema <[email protected]>:
o fix nanosecond stat timefields in UDF
Brian Gerst <[email protected]>:
o break up fs/devices.c
Chris Wright <[email protected]>:
o sys_capget should use current if the pid argument is 0
o remove duplicated assignment from sys_capget
Christoph Hellwig <[email protected]>:
o make scsi_ioctl.h useable without including scsi.h
o rationalize allocation and freeing of struct scsi_device
o remove dead struct/typedef from hosts.h
o remove unused includes and misleading comments from scsi_lib.c
o fix compilation for !CONFIG_SWAP
o uClinux bits for /dev/zero
Dave Jones <[email protected]>:
o A new Athlon 'bug'
o Move watchdog drivers to drivers/char/watchdog/
David S. Miller <[email protected]>:
o [SPARC64]: PCI device name changes
o [SPARC64]: Missing linux/interrupt.h includes
o arch/sparc64/kernel/sys_sunos32.c: Include net/sock.h
o [SPARC64]: More missing includes after the cleanups
o [SPARC64]: Nanosecond time changes
o [SPARC64]: More nanosecond timestamp updates
o [SCSI]: drivers/scsi/scsi.h needs linux/init.h
o [PARPORT]: drivers/parport/ieee1284.c needs sched.h and timer.h
o [QLOGICFC]: drivers/scsi/qlogicfc.c needs interrupt.h
o [SPARC64]: drivers/sbus/char/bbc_i2c.c needs interrupt.h
o [SCSI]: drivers/scsi/qlogicpti.c needs interrupt.h
o [SCSI]: drivers/scsi/qlogicisp.c needs interrupt.h
o [SCSI]: drivers/scsi/aic7xxx_old.c needs interrupt.h
o [SCSI]: drivers/scsi/esp.c needs interrupt.h plus fix abort/reset
locking
o [SCSI]: drivers/scsi/sym53c8xx_2/sym_glue.h needs interrupt.h
o [SPARC]: Module loading API updates
o [SOUND]: sound/sparc/{amd7930,cs4231}.c needs interrupt.h
o [SPARC]: Implement module_{init,core}_size
o [SPARC64]: isa_{device,bus} --> sparc_isa_{device,bus}
o [SPARC]: Add set_tid_address syscall vectors
o [IPSEC]: Make xfrm_user key manager return proper errors
o [CRYPTO]: Forgot to add crypto_null.c in previous commit
o [CRYPTO]: Kill stray CRYPTO_ALG_TYPE_COMP
o [NET]: Make sock_ioctl truly static
o [SPARC]: Add epoll syscall entries
o [TG3]: Use spin_lock_irq{save,restore} on tx_lock
o [VLAN]: remove vlan_devices[] entries properly
o [SPARC64]: Move data.cacheline_aligned right before edata
o [XFRM_USER]: Index xfrma array correctly
o [SPARC]: Remove schedule_tail ifdefs
o [SPARC]: Update for new do_fork semantics
o [SPARC]: Handle clone flag name changes
David S. Miller <[email protected]>:
o Missing unlock_kernel() in fs/block_dev.c
David Woodhouse <[email protected]>:
o JFFS2 update
o Fix nanosecond merge
Dominik Brodowski <[email protected]>:
o cpufreq: cleanups
Douglas Gilbert <[email protected]>:
o scsi_debug 1.65 for lk 2.5.48
Eric Brower <[email protected]>:
o [SPARC]: Make APC idle a boot time cmdline option
Greg Kroah-Hartman <[email protected]>:
o USB: vicam.c driver fixes
o ISDN: Convert usages of pcibios_* functions to pci_*
o PCMCIA: remove usage of pcibios_read_config_dword
o PCI: removed pcibios_read_config_* and pcibios_write_config_*
functions
o USB: usb-serial core updates
Hideaki Yoshifuji <[email protected]>:
o IPv6: Fix BUG When Received Unknown Protocol
Ingo Molnar <[email protected]>:
o threading enhancements, tid-2.5.48-C0
Ivan Kokshaysky <[email protected]>:
o alpha: initrd vs. "mem=" fix
o PCI: transparent bridge detection fix
o PCI setup: misc cleanups and fixes
o compile fixes
o PCI: rename exported pbus_* functions
James Bottomley <jejb@mulgrave.(none)>:
o fix queue run on returning I/O [[email protected]]
o fix queue plug performance problem found by akpm
James Morris <[email protected]>:
o [CRYPTO] kstack cleanup (v0.28)
o [CRYPTO]: Fix non-modular build
o [CRYPTO]: Add maintainers entry
o [CRYPTO]: Minor doc update
o [CRYPTO]: Add null algorithms and minor cleanups
Jeff Garzik <[email protected]>:
o [CRYPTO]: Kill accidental double memset
Jon Grimm <[email protected]>:
o [SCTP] sctp_params cleanup (jgrimm)
o [SCTP] More param handling, mostly handling hostname parm (jgrimm)
o [SCTP]: sctp_process_init can fail; cleanup and bail on errors.
(jgrimm)
o [SCTP] sctp_addr code cleanup
o [SCTP] Handle "no route" case for output handler (jgrimm)
o [SCTP] Addr family cleanup part 2. (jgrimm)
o [SCTP]: enable v6 autobinding
o [SCTP] PF_INET6 sockets should accept v4 addresses into association
Kai Germaschewski <[email protected]>:
o kbuild: Fix KBUILD_MODNAME
o kbuild: arch/i386/oprofile/Makefile cosmetics
Linus Torvalds <[email protected]>:
o Devfs was broken by the nanosecond inode times. Fix properly
o Parts of "module.c" was needed even when no module support was
enabled, so split it up into "extable.c"
o Broken applications do not realize that a zero return from
"write()" is an error condition, and hang retrying.
o Don't trust "rq->cmd_len", most of the internal IDE-CD command
sending routines don't set it up.
o Add <asm/io.h> include for readw uses in <pcmcia/mem_op.h>
Luca Barbieri <[email protected]>:
o dup_mmap tiny optimization
o Small futex improvement
Martin Schwidefsky <[email protected]>:
o s390: config & make
o s390: system calls
o s390: mman
o s390: 64bit sector_t
o s390: module loader
o s390: missing includes
o s390: uaccess bug
o s390: gcc 3.2 fixes
o s390: ebcdic conversion bug
o s390: flushtlb bug
o s390: isclean bug
o s390: 31bit emulation
o s390: xpram driver
o s390: warnings
o s390: sclp driver part 1
o s390: sclp driver part 2
Matthew Dharm <[email protected]>:
o usb-storage: change function signatures and cleanup debug msgs
o usb-storage: fix missed changes in freecom.c and isd200.c
o usb-storage: code consolidation
Matthew Wilcox <[email protected]>:
o missing smp.h in topology.c
o remove sched.h from sctp/sm.h
o rename get_lease to break_lease
Neil Brown <[email protected]>:
o Fix *_mergeable_bvec routines for linear/raid0
o Fix r5 bug - wrong variable used
o Tidy up some handling of sb_dirty in md.c
o Remove unused variable in umem.c
o Avoid 'defined but not used' warning with i386/xor.h
o NFSv3 to extract large symlinks from paginated requests
o Fix err in size calculation for readdir response
o Fix bug in svc_udp_recvfrom
o Avoid copying unfragmented udp NFS requests
o Only set dest addr in NFS/udp reply, not NFS/tcp
o kNFSd - 1 of 2 - Change NFSv4 xdr decoding to cope with separate
pages
o kNFSd - 2 of 2 - Change NFSv4 reply encoding to cope with multiple
pages
[email protected] <[email protected]>:
o Via KT400 agp support
Patrick Mansfield <[email protected]>:
o Re: 2.5.48 /proc/scsi directory missing
o remove unused includes and misleading comments from scsi_lib.c
Patrick Mochel <[email protected]>:
o make sure DEBUG is #undef'd so it's really turned off
o make sure cpu class is registered before cpu driver
o driver model: make sure driver is added to class it belongs to
o driver model: catch one last #define DEBUG 0
o driver model: don't double up() if device registration fails
o driver model: typo in intf.c
o kobject - expose backend helpers to registration interface
o sysfs: do permission checking on open
o driver model: exploit kobject contstructs
o driver model: keep reference to device during device_add()
o sysfs: various updates
o driver model: update and clean bus and driver support
o driver model: make classes and interfaces use kobject
infrastructure
o USB: minor driver model-related updates
o partitions: use the name in disk->kobj.name, instead of
disk->disk_name
Paul Mackerras <[email protected]>:
o PPC32: adjust some includes in response to recent include cleanups
o PPC32: adjust <asm/atomic.h> so it doesn't need <asm/system.h>
o PPC32: Move BUG() definition from asm-ppc/page.h to
asm-ppc/processor.h
o PPC32: clean up the arch/ppc/boot Makefiles
Pete Zaitcev <[email protected]>:
o [SPARC]: Fix finish_arch_switch and factor PIL users
Petr Vandrovec <[email protected]>:
o [LLC]: Fix timer_init calls
o Merge lcall7 and lcall27 code paths in ia32
o Mark executable files as executable on ncpfs
o Small matroxfb fixes
Richard Henderson <[email protected]>:
o Fix carry ripple in 3 and 4 word addition and subtraction macros
o Fix a whole pile of signed/unsigned comparison warnings
o Remove duplicate sys_rx164.o in GENERIC kernel
o Avoid multi-line string literal for asm block
o Add Alpha entry in MAINTAINERS
o Misc alpha compilation fixes
o Include IEEE1394 configury
o Add new syscalls
o Remove dead module code so that builds with CONFIG_MODULES=n will
succeed.
o Fix scsi build wrt include files
o Cset exclude: [email protected]|ChangeSet|20021118212616|10632
o Add dummy <asm/suspend.h> for alpha
Richard Henderson <[email protected]>:
o Remove osf_swapon; fall back to sys_swapon immediately
o [ALPHA] Update clone syscall for child_tid argument
Rik van Riel <[email protected]>:
o advansys.c buffer overflow
Robert Love <[email protected]>:
o ALSA compiler warnings fixes
Russell King <[email protected]>:
o [ARM] Add cpu_flush_pmd()
o [ARM] Acorn SCSI build fixes
o [ARM] 2.5.48 Build fixes (round 1)
o [ARM] 2.5.48 module fixups (and disable module loading for ARM)
o [ARM] Undefine symbol "arm", update mach-types
o [ARM] 2.5.48 Build fixes (round 2)
o [ARM] Update ARM bitops to operate on unsigned long quantities
o [ARM] Fix ARM module support
o [ARM] Fixups for 2.5.48-bkcur
Rusty Russell <[email protected]>:
o kallsyms for new modules
o PPC32: In-kernel module linker for PPC
o module device table restoration
o Module length calculation fix and module with no init fix
Sam Ravnborg <[email protected]>:
o [SPARC64]: Use kbuild more consistently, add archhelp target
o [SPARC64]: Makefile cleanups
Stelian Pop <[email protected]>:
o sonypi driver update
o meye driver update
Steve French <[email protected]>:
o Properly emulate POSIX semantics for deleting of open files and
renaming over open files. Fix oops in readpages caused by
unitialized aops field
Tom Rini <[email protected]>:
o Add back in <asm/system.h> and <linux/linkage.h> to
<linux/interrupt.h>
Trond Myklebust <[email protected]>:
o Support for micro/nanosecond [acm]time in the NFS client
o Split buffer overflow checking out of struct nfs4_compound
William Lee Irwin III <[email protected]>:
o Numa-Q bootup failure fix
o input: fix up header cleanups: add <linux/interrupt.h>
o sched: privatizes the sibling inlines to sched.c, the sole caller
of them
Zwane Mwaikambo <[email protected]>:
o USB core/config.c == memory corruption
> Anyone got a copy, yet?
Kernel.org seems to be down an hour early.
I can put the patch on a fast U.K. based server if somebody E-Mails it
to me, preferably with the .sign as well.
John.
Ok, I've just put up 2.5.49 tarballs, .sign's, and patches of the
.gz and .bz2 varieties at:
ftp://zeniiia.linux.org.uk/pub/linux/kernel/v2.5
ftp://zeniii.linux.org.uk/pub/linux/kernel/v2.5
http://ftp.linux.org.uk/pub/linux/kernel/v2.5
ftp://ftp.linux.org.uk/pub/linux/kernel/v2.5
That's 3 machines at two ISPs in the UK, both all with 100mbit bandwidth
to the outside world.
(People behind windows routers may have problems with zeniii.linux.org.uk)
I'd like to thank Andre Hedrick for retrieving these files and thereby
making this possible.
--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html
Linus,
Russell and I have hooked togather and relayed it to the UK.
So just wait for Russell to annouce the download points.
Cheers,
Andre Hedrick
LAD Storage Consulting Group
On Fri, 22 Nov 2002, Linus Torvalds wrote:
>
> Ok, I appear to be without network connectivity at home at least over the
> weekend, and master.kernel.org is going down for some maintenance this
> afternoon, so here's my current tree.
>
> Architecture updates, threading improvements, shm fix (the cause of
> the Oracle problems), networking, scsi, modules, you name it, it's here.
>
> Due to my lacking network connection over the weekend, I'd suggest
> discussing issues on linux-kernel, since emailing them to me won't much
> help ;/
>
> The joys of switching ISPs..
>
> Linus
>
> -----
>
> Summary of changes from v2.5.48 to v2.5.49
> ============================================
>
> <[email protected]>:
> o [EBTABLES]: Use correct base pointer in ebt_do_table
> o add necessary #ifdefs to netfilter_bridge.h, vs 2.5.48
>
> <[email protected]>:
> o [SPARC]: Add missing iounmap to display7seg driver
>
> <[email protected]>:
> o [VLAN]: Quiet some printks and free devices/groups correctly
>
> <[email protected]>:
> o [SCH_GRED]: Array overflow fixes, found by Stanford checker
>
> <[email protected]>:
> o [IPV6]: Export ipv6_chk_addr
>
> <[email protected]>:
> o vicam.c
>
> <[email protected]>:
> o dcache usage cleanups
>
> <[email protected]>:
> o ADM8513 support added;
>
> <[email protected]>:
> o fix intermezzo compile
>
> <[email protected]>:
> o [SCTP] Support for Peer address parameters socket option
> o [SCTP] udp-style connect support(non-blocking)
> o [SCTP] Fix for sideeffect violation in sctp_sf_heartbeat()
> o [SCTP] Blocking connect() support
> o [SCTP] MSG_EOR support for recvmsg()
>
> <[email protected]>:
> o [IPSEC]: Fix double unlock in esp/ah
>
> Adrian Bunk <[email protected]>:
> o dv1394 devfs missing brace
> o fix compile error in usb-serial.c
>
> Alexey Kuznetsov <[email protected]>:
> o [IPSEC]: Policy timeout and pfkey acquire fixes
>
> Andi Kleen <[email protected]>:
> o Some leftover nsec stat fixes (ADFS,AFS,CIFS)
>
> Andrew Morton <[email protected]>:
> o fix the build for egcs-1.1.2
> o detect uninitialised per-cpu storage
> o explicitly initialise kstat per-cpu storage
> o misc
> o shmdt bugfix
> o radix-tree reinitialisation fix
> o Add SMP barrier to ipc's grow_ary()
> o reduce CPU cost in loop
> o Expanded bad page handling
> o Make inode_ops->setxattr value parameter const
> o fix endian problem in ext3 htree code
> o remove a warning from __block_write_full_page()
> o ext2/ext3 Orlov directory accounting fix
> o bootmem crash fix
> o Fix busy-wait with writeback to large queues
> o Remove mapping->vm_writeback
> o strengthen the `incremental min' logic in the page
> o handle zones which are full of unreclaimable pages
> o no-buffer-head ext2 option
>
> Andries E. Brouwer <[email protected]>:
> o inode_mknod parameters
> o compilation fix (tpyo)
> o *_mknod prototype
> o kill i_dev
> o fcntl fix
>
> Andy Grover <[email protected]>:
> o ACPI: Fix compilation error when CONFIG_SOFTWARE_SUSPEND is not set
> (Shawn Starr)
> o ACPI: fix debug print levels, and use down() instead of
> down_interruptible(), and some whitespace.
> o ACPI: Interpreter fixes Fixed memory leak in method argument
> resolution Fixed Index() operator to work properly with a target
> operand Fixed attempted double delete in the Index() code Code size
> improvements Improved debug/error messages and levels Fixed a
> problem with premature deletion of a buffer object
> o ACPI: Add ec_read and ec_write external functions Other ec.c
> cleanups, too
>
> Anton Blanchard <[email protected]>:
> o disable old stat on ppc64
>
> Arnaldo Carvalho de Melo <[email protected]>:
> o tr: make CONFIG_TR depend on CONFIG_LLC=y
> o wdt: fix up header cleanups: add include linux/interrupt.h
> o hotplug: fix up header cleanups: add include linux/interrupts.h
> o sound: more fixups for header cleanup: add include
> <linux/interrupt.h>
> o i2c: fix up header cleanups: add include <linux/interrupt.h>
> o input: fix up header cleanups: add include <linux/interrupt.h>
> o smbfs: fixup header cleanups: forward declare struct sock, add
> include/uio.h
> o scsi: fix up header cleanups: add include <linux/interrupt.h>
> o hd: fix up header cleanup: add include <linux/interrupt.h>
> o i2o: fix up header cleanups: add include <linux/interrupt.h>
> o tcic: fix up header cleanups: add include <linux/interrupt.h>
> o sound: fix up header cleanups: add include <linux/interrupt.h>
> o net/core: export sk_send_sigurg, its needed by x25 when built as a
> module
> o net: fix up header cleanups: remove unneeded sched.h include
> o drivers/net: fix up header cleanup: remove unneeded sched.h
> includes
> o o drivers/net/hamradio: fix up header cleanups: remove uneeded
> sched.h includes
> o o drivers/net/wan/lmc: fix up header cleanups: remove uneeded
> sched.h includes
>
> Art Haas <[email protected]>:
> o C99 initializer for fs/afs/inode.c
> o C99 initializers for drivers/serial
> o C99 initializer for drivers/zorro/proc.c
> o C99 initializers for drivers/message/i2o/i2o_config.c
> o C99 initializers for drivers/pci
> o C99 initiailzers for fs/intermezzo
> o C99 initializers for drivers/pcmcia
> o C99 initializers for drivers/ieee1394
> o C99 initializers for drivers/input
> o C99 initializers for drivers/block/cciss_scsi.c
> o C99 initializers for drivers/scsi
> o C99 initializers for drivers/cdrom
> o C99 initializers for drivers/i2c
> o C99 initializers for drivers/parisc
> o C99 initializers for drivers/parport
> o C99 initializers for drivers/pnp
> o C99 initializers for drivers/sgi
> o C99 initializers for drivers/s390
> o C99 initializers for drivers/tc
> o C99 initializers for drivers/telephony
> o fix C99 initializers fix
> o C99 initializer for init/initramfs.c
> o C99 initializer for drivers/parisc/ccio-dma.c
> o C99 initializer for fs/ext2/dir.c
> o C99 initializer for fs/sysv/super.c
>
> Ben Fennema <[email protected]>:
> o fix nanosecond stat timefields in UDF
>
> Brian Gerst <[email protected]>:
> o break up fs/devices.c
>
> Chris Wright <[email protected]>:
> o sys_capget should use current if the pid argument is 0
> o remove duplicated assignment from sys_capget
>
> Christoph Hellwig <[email protected]>:
> o make scsi_ioctl.h useable without including scsi.h
> o rationalize allocation and freeing of struct scsi_device
> o remove dead struct/typedef from hosts.h
> o remove unused includes and misleading comments from scsi_lib.c
> o fix compilation for !CONFIG_SWAP
> o uClinux bits for /dev/zero
>
> Dave Jones <[email protected]>:
> o A new Athlon 'bug'
> o Move watchdog drivers to drivers/char/watchdog/
>
> David S. Miller <[email protected]>:
> o [SPARC64]: PCI device name changes
> o [SPARC64]: Missing linux/interrupt.h includes
> o arch/sparc64/kernel/sys_sunos32.c: Include net/sock.h
> o [SPARC64]: More missing includes after the cleanups
> o [SPARC64]: Nanosecond time changes
> o [SPARC64]: More nanosecond timestamp updates
> o [SCSI]: drivers/scsi/scsi.h needs linux/init.h
> o [PARPORT]: drivers/parport/ieee1284.c needs sched.h and timer.h
> o [QLOGICFC]: drivers/scsi/qlogicfc.c needs interrupt.h
> o [SPARC64]: drivers/sbus/char/bbc_i2c.c needs interrupt.h
> o [SCSI]: drivers/scsi/qlogicpti.c needs interrupt.h
> o [SCSI]: drivers/scsi/qlogicisp.c needs interrupt.h
> o [SCSI]: drivers/scsi/aic7xxx_old.c needs interrupt.h
> o [SCSI]: drivers/scsi/esp.c needs interrupt.h plus fix abort/reset
> locking
> o [SCSI]: drivers/scsi/sym53c8xx_2/sym_glue.h needs interrupt.h
> o [SPARC]: Module loading API updates
> o [SOUND]: sound/sparc/{amd7930,cs4231}.c needs interrupt.h
> o [SPARC]: Implement module_{init,core}_size
> o [SPARC64]: isa_{device,bus} --> sparc_isa_{device,bus}
> o [SPARC]: Add set_tid_address syscall vectors
> o [IPSEC]: Make xfrm_user key manager return proper errors
> o [CRYPTO]: Forgot to add crypto_null.c in previous commit
> o [CRYPTO]: Kill stray CRYPTO_ALG_TYPE_COMP
> o [NET]: Make sock_ioctl truly static
> o [SPARC]: Add epoll syscall entries
> o [TG3]: Use spin_lock_irq{save,restore} on tx_lock
> o [VLAN]: remove vlan_devices[] entries properly
> o [SPARC64]: Move data.cacheline_aligned right before edata
> o [XFRM_USER]: Index xfrma array correctly
> o [SPARC]: Remove schedule_tail ifdefs
> o [SPARC]: Update for new do_fork semantics
> o [SPARC]: Handle clone flag name changes
>
> David S. Miller <[email protected]>:
> o Missing unlock_kernel() in fs/block_dev.c
>
> David Woodhouse <[email protected]>:
> o JFFS2 update
> o Fix nanosecond merge
>
> Dominik Brodowski <[email protected]>:
> o cpufreq: cleanups
>
> Douglas Gilbert <[email protected]>:
> o scsi_debug 1.65 for lk 2.5.48
>
> Eric Brower <[email protected]>:
> o [SPARC]: Make APC idle a boot time cmdline option
>
> Greg Kroah-Hartman <[email protected]>:
> o USB: vicam.c driver fixes
> o ISDN: Convert usages of pcibios_* functions to pci_*
> o PCMCIA: remove usage of pcibios_read_config_dword
> o PCI: removed pcibios_read_config_* and pcibios_write_config_*
> functions
> o USB: usb-serial core updates
>
> Hideaki Yoshifuji <[email protected]>:
> o IPv6: Fix BUG When Received Unknown Protocol
>
> Ingo Molnar <[email protected]>:
> o threading enhancements, tid-2.5.48-C0
>
> Ivan Kokshaysky <[email protected]>:
> o alpha: initrd vs. "mem=" fix
> o PCI: transparent bridge detection fix
> o PCI setup: misc cleanups and fixes
> o compile fixes
> o PCI: rename exported pbus_* functions
>
> James Bottomley <jejb@mulgrave.(none)>:
> o fix queue run on returning I/O [[email protected]]
> o fix queue plug performance problem found by akpm
>
> James Morris <[email protected]>:
> o [CRYPTO] kstack cleanup (v0.28)
> o [CRYPTO]: Fix non-modular build
> o [CRYPTO]: Add maintainers entry
> o [CRYPTO]: Minor doc update
> o [CRYPTO]: Add null algorithms and minor cleanups
>
> Jeff Garzik <[email protected]>:
> o [CRYPTO]: Kill accidental double memset
>
> Jon Grimm <[email protected]>:
> o [SCTP] sctp_params cleanup (jgrimm)
> o [SCTP] More param handling, mostly handling hostname parm (jgrimm)
> o [SCTP]: sctp_process_init can fail; cleanup and bail on errors.
> (jgrimm)
> o [SCTP] sctp_addr code cleanup
> o [SCTP] Handle "no route" case for output handler (jgrimm)
> o [SCTP] Addr family cleanup part 2. (jgrimm)
> o [SCTP]: enable v6 autobinding
> o [SCTP] PF_INET6 sockets should accept v4 addresses into association
>
> Kai Germaschewski <[email protected]>:
> o kbuild: Fix KBUILD_MODNAME
> o kbuild: arch/i386/oprofile/Makefile cosmetics
>
> Linus Torvalds <[email protected]>:
> o Devfs was broken by the nanosecond inode times. Fix properly
> o Parts of "module.c" was needed even when no module support was
> enabled, so split it up into "extable.c"
> o Broken applications do not realize that a zero return from
> "write()" is an error condition, and hang retrying.
> o Don't trust "rq->cmd_len", most of the internal IDE-CD command
> sending routines don't set it up.
> o Add <asm/io.h> include for readw uses in <pcmcia/mem_op.h>
>
> Luca Barbieri <[email protected]>:
> o dup_mmap tiny optimization
> o Small futex improvement
>
> Martin Schwidefsky <[email protected]>:
> o s390: config & make
> o s390: system calls
> o s390: mman
> o s390: 64bit sector_t
> o s390: module loader
> o s390: missing includes
> o s390: uaccess bug
> o s390: gcc 3.2 fixes
> o s390: ebcdic conversion bug
> o s390: flushtlb bug
> o s390: isclean bug
> o s390: 31bit emulation
> o s390: xpram driver
> o s390: warnings
> o s390: sclp driver part 1
> o s390: sclp driver part 2
>
> Matthew Dharm <[email protected]>:
> o usb-storage: change function signatures and cleanup debug msgs
> o usb-storage: fix missed changes in freecom.c and isd200.c
> o usb-storage: code consolidation
>
> Matthew Wilcox <[email protected]>:
> o missing smp.h in topology.c
> o remove sched.h from sctp/sm.h
> o rename get_lease to break_lease
>
> Neil Brown <[email protected]>:
> o Fix *_mergeable_bvec routines for linear/raid0
> o Fix r5 bug - wrong variable used
> o Tidy up some handling of sb_dirty in md.c
> o Remove unused variable in umem.c
> o Avoid 'defined but not used' warning with i386/xor.h
> o NFSv3 to extract large symlinks from paginated requests
> o Fix err in size calculation for readdir response
> o Fix bug in svc_udp_recvfrom
> o Avoid copying unfragmented udp NFS requests
> o Only set dest addr in NFS/udp reply, not NFS/tcp
> o kNFSd - 1 of 2 - Change NFSv4 xdr decoding to cope with separate
> pages
> o kNFSd - 2 of 2 - Change NFSv4 reply encoding to cope with multiple
> pages
>
> [email protected] <[email protected]>:
> o Via KT400 agp support
>
> Patrick Mansfield <[email protected]>:
> o Re: 2.5.48 /proc/scsi directory missing
> o remove unused includes and misleading comments from scsi_lib.c
>
> Patrick Mochel <[email protected]>:
> o make sure DEBUG is #undef'd so it's really turned off
> o make sure cpu class is registered before cpu driver
> o driver model: make sure driver is added to class it belongs to
> o driver model: catch one last #define DEBUG 0
> o driver model: don't double up() if device registration fails
> o driver model: typo in intf.c
> o kobject - expose backend helpers to registration interface
> o sysfs: do permission checking on open
> o driver model: exploit kobject contstructs
> o driver model: keep reference to device during device_add()
> o sysfs: various updates
> o driver model: update and clean bus and driver support
> o driver model: make classes and interfaces use kobject
> infrastructure
> o USB: minor driver model-related updates
> o partitions: use the name in disk->kobj.name, instead of
> disk->disk_name
>
> Paul Mackerras <[email protected]>:
> o PPC32: adjust some includes in response to recent include cleanups
> o PPC32: adjust <asm/atomic.h> so it doesn't need <asm/system.h>
> o PPC32: Move BUG() definition from asm-ppc/page.h to
> asm-ppc/processor.h
> o PPC32: clean up the arch/ppc/boot Makefiles
>
> Pete Zaitcev <[email protected]>:
> o [SPARC]: Fix finish_arch_switch and factor PIL users
>
> Petr Vandrovec <[email protected]>:
> o [LLC]: Fix timer_init calls
> o Merge lcall7 and lcall27 code paths in ia32
> o Mark executable files as executable on ncpfs
> o Small matroxfb fixes
>
> Richard Henderson <[email protected]>:
> o Fix carry ripple in 3 and 4 word addition and subtraction macros
> o Fix a whole pile of signed/unsigned comparison warnings
> o Remove duplicate sys_rx164.o in GENERIC kernel
> o Avoid multi-line string literal for asm block
> o Add Alpha entry in MAINTAINERS
> o Misc alpha compilation fixes
> o Include IEEE1394 configury
> o Add new syscalls
> o Remove dead module code so that builds with CONFIG_MODULES=n will
> succeed.
> o Fix scsi build wrt include files
> o Cset exclude: [email protected]|ChangeSet|20021118212616|10632
> o Add dummy <asm/suspend.h> for alpha
>
> Richard Henderson <[email protected]>:
> o Remove osf_swapon; fall back to sys_swapon immediately
> o [ALPHA] Update clone syscall for child_tid argument
>
> Rik van Riel <[email protected]>:
> o advansys.c buffer overflow
>
> Robert Love <[email protected]>:
> o ALSA compiler warnings fixes
>
> Russell King <[email protected]>:
> o [ARM] Add cpu_flush_pmd()
> o [ARM] Acorn SCSI build fixes
> o [ARM] 2.5.48 Build fixes (round 1)
> o [ARM] 2.5.48 module fixups (and disable module loading for ARM)
> o [ARM] Undefine symbol "arm", update mach-types
> o [ARM] 2.5.48 Build fixes (round 2)
> o [ARM] Update ARM bitops to operate on unsigned long quantities
> o [ARM] Fix ARM module support
> o [ARM] Fixups for 2.5.48-bkcur
>
> Rusty Russell <[email protected]>:
> o kallsyms for new modules
> o PPC32: In-kernel module linker for PPC
> o module device table restoration
> o Module length calculation fix and module with no init fix
>
> Sam Ravnborg <[email protected]>:
> o [SPARC64]: Use kbuild more consistently, add archhelp target
> o [SPARC64]: Makefile cleanups
>
> Stelian Pop <[email protected]>:
> o sonypi driver update
> o meye driver update
>
> Steve French <[email protected]>:
> o Properly emulate POSIX semantics for deleting of open files and
> renaming over open files. Fix oops in readpages caused by
> unitialized aops field
>
> Tom Rini <[email protected]>:
> o Add back in <asm/system.h> and <linux/linkage.h> to
> <linux/interrupt.h>
>
> Trond Myklebust <[email protected]>:
> o Support for micro/nanosecond [acm]time in the NFS client
> o Split buffer overflow checking out of struct nfs4_compound
>
> William Lee Irwin III <[email protected]>:
> o Numa-Q bootup failure fix
> o input: fix up header cleanups: add <linux/interrupt.h>
> o sched: privatizes the sibling inlines to sched.c, the sole caller
> of them
>
> Zwane Mwaikambo <[email protected]>:
> o USB core/config.c == memory corruption
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
> I have the patch, but i forgot to dl the sign, and already unziped it,
> so the sign would work.
>
> If u want it i'll send it to you
Yes, please - I'll put it on-line for the others then.
Cheers,
John.
> Ok, I've just put up 2.5.49 tarballs, .sign's, and patches of the
> .gz and .bz2 varieties at:
>
> ftp://zeniiia.linux.org.uk/pub/linux/kernel/v2.5
> ftp://zeniii.linux.org.uk/pub/linux/kernel/v2.5
> http://ftp.linux.org.uk/pub/linux/kernel/v2.5
> ftp://ftp.linux.org.uk/pub/linux/kernel/v2.5
>
> That's 3 machines at two ISPs in the UK, both all with 100mbit bandwidth
> to the outside world.
Thanks, that's brilliant. I promised to put them on-line myself, and
although it's a bit redundant now, I've done so:
http://grabjohn.com/kernel
in Redbus, U.K.
John.
> I'd like to thank Andre Hedrick for retrieving these files and thereby
> making this possible.
I'd like to thank Russel King for thanking Andre Hedrick for
retrieving these files and thereby making this possible. :-)
John.
On Fri, 22 Nov 2002 13:55:08 -0800 (PST) Linus Torvalds (LT) wrote:
LT> Summary of changes from v2.5.48 to v2.5.49
LT> ============================================
[...]
Hello,
Someone already reported the problem for 2.5.48, and it's still present in
2.5.49:
SCSI subsystem driver Revision: 1.00
ERROR: SCSI host `ide-scsi' has no error handling
ERROR: This is not a safe way to run your SCSI host
ERROR: The error handling must be added to this driver
Call Trace:
[<c026e193>] scsi_register+0x2e3/0x2f0
[<c0210aff>] bus_add_driver+0xaf/0xd0
[<c0275882>] idescsi_detect+0x22/0x80
[<c026e1d3>] scsi_register_host+0x33/0xd0
[<c010507a>] init+0x3a/0x160
[<c0105040>] init+0x0/0x160
[<c010713d>] kernel_thread_helper+0x5/0x18
Regards,
-Udo.
On Fri, Nov 22, 2002 at 01:55:08PM -0800, Linus Torvalds wrote:
>
> Ok, I appear to be without network connectivity at home at least over the
> weekend, and master.kernel.org is going down for some maintenance this
> afternoon, so here's my current tree.
This (and -mm1) does not boot for me. Or rather, the kernel boots
but the systems hangs hard when kdm starts. No numlock, no sysrq.
Only hard reboot helps.
I can boot single-user fine. And I can boot X fine from a
single-user root login. KDE then hangs, but X definitely starts.
I can terminate X at this point with crtl-alt-backspace.
My system is a Debian testing, X is 4.2.1. The kernel has no DRI
support configured, no SMP, has preempt. Machine is P3-450, 128MB
RAM.
2.5.47mm1 works fine. My 2.5.49 .config is at
http://www.jaquet.dk/kernel/config-2.5.49
Regards,
Rasmus
Hi!
> Architecture updates, threading improvements, shm fix (the cause of
> the Oracle problems), networking, scsi, modules, you name it, it's here.
It might be nice to include .sign files in release
announcement... Fake anouncement should be
discovered very quickly.
--
Pavel
Written on sharp zaurus, because my Velo1 broke. If you have Velo you don't need...