Hmm.. Nothing really fundamental here - various updates all over
(architecture updates, networking, usb, acpi, bluetooth, the usual
suspects).
The task structure reference counting seems to have broken alpha, Richard
is still chasing that one down.
Linus
Summary of changes from v2.5.62 to v2.5.63
============================================
<[email protected]>:
o [SPARC64]: Fix ocndition code handling in do_rt_sigreturn
<[email protected]>:
o [SCTP/IPV6]: Move sockaddr storage and in6addr_{any,loopback} to
generic places
<[email protected]>:
o [ATM]: Add MAINTAINERS entry
<[email protected]>:
o [ATM]: use sock timestamp
o [ATM]: Remove cli from lec.c
o [ATM]: empty tx queue in lec when flush complete the lane spec says
the following about path switching:
o [ATM]: prevent compiler warning when compiling w/o bridging
<[email protected]>:
o typo in 53c700.c
<[email protected]>:
o /proc/ide/triflex returns incomplete data
<[email protected]>:
o [OPROF] Support EV67 ProfileMe
<[email protected]>:
o aic79xx build and lun detect problem fix
<[email protected]>:
o [SPARC64]: Fix typo in sparc64_get_context (G7 register is saved
wrongly)
<[email protected]>:
o PCI: list code cleanup
<[email protected]>:
o [IPSEC]: Add missing credit and include to xfrm_user ipv6 changes
<[email protected]>:
o [XFS] fix one more set of transaction callback ordering issues,
this was always there, but exposed by the last change in this area
and made much more likely.
<[email protected]>:
o Fusion Driver 2.05.00.03 against 2.5.62bk3
<[email protected]>:
o Re: Griffin Powermate: Aluminum
<[email protected]>:
o need archmrproper
o 64bit dynamic app fix (still needs binfmt_elf.c patch to work)
o zImage now holds vmlinux, System.map and config in sections
o handle RI for GQ processors
o ignore files for new zImage build
o Cset exclude: [email protected]|ChangeSet|20030207200510|30566
Adam Belay <[email protected]>:
o Preparations and Cleanups
o This patch contains an improved resource management algorithm. It
is capable of resolving nearly any conflict between two or more PnP
devices.
o Moves the resource parsing functions to a new location "support.c".
These resource parsing functions contain many improvements
including the ability to set resources according to actual value
rather than dependent functions.
o Interface Updates
o ISAPnP Updates
o PnPBIOS Updates
o Trivial Card Service Fix
o Radio-Cadet PnP Update
o IDE PnP Update
o Trivial C99 Update
o PnP Bug Fixes
o OSS Sound Blaster Update from Paul Laufer
o Large Stack Usage Fix
o Resource Management Performance Fix
Adrian Bunk <[email protected]>:
o remove an unneeded #if from net/ipv6/af_inet6.c
Alan Cox <[email protected]>:
o ide resync
o add generic ide iops
o eliminate use of ide_ioreg_t on ARM
o update ide.c
o remove old style and unused bad drive list
o clean up the IDE iops, add ones for a dead iface
o fix ide_ioreg_t and ifdefs in iops
o add ide_execute_command but do not use it yet
o remove ide_ioreg_t
o ide-probe updates
o ide-proc - fix crash on identify
o add new settings locks to ide-proc
o ide-tape no longer needs this ifdef
o fix path of file
o path/ide_ioreg_t fixes for legacy drivers
o fix int for i/o in pcmcia ide_cs
o fix the rest of the names/ide_ioreg_t in ide legacy
o rmeove ide_ioreg_t from PCI ide
o fix path names and printks in IDE pci
o add a 'NO_IRQ' definition to IDE
o exterminate unused io_ops structures and switch to ulong
o add pio_speed
o kill more ioregs, add OUTBSYNC
o resync externs, add execute command remove is_flashcard
o copy idesync
o use ide_execute_command for CD
o add a reminder for vdma on non disk
o clean up DMA reference, new style ONLYDISK
o ide-dma, fix bogus inc of waiting_for_dma
o update ide-floppy for new style onlydisk
o fix ALi 32bitisms, fix ALi FIFO, fix ALi IRQ crash
o fix some escaped globals
o don't force enable generic IDE controllers
o part fix the highpoint timing/overclock bug
o clean up siimage, use generic mmio ops
o update sis driver
o make the sl82c105 work again
o ndelay() for x86
Alan Stern <[email protected]>:
o USB: Patches for the ECONNRESET error (2.5)
Andi Kleen <[email protected]>:
o x86-64 update for 2.5.62-bk4
o Fix some 64bit warnings
o Allow xtime_lock declaration in arch specific code for x86-64
o [NET]: Make skbuff.h -W clean, skb_headlen should return unsigned
quantity
o Fix x86-64 loose ends
Andrew Morton <[email protected]>:
o signal warning and uninitialised variable fix
o MPT Fusion build fix
o fix for uninitialized timer in drm_drv.h
o export add_to_page_cache() and __pagevec_lru_add to
o Move mk_pte_huge() into pgtable.h
o fix kirq for clustered apic mode
o Remove MAX_BLKDEV from nfsd
o Fix warnings for XFS
o Fix warnings for NTFS
o allow SMP kernel build without io_apic.c
o export some functions from i8259.c
o make startup_32 kernel entry point
o export boottime gdt descriptor
o visws: boot changes
o visws: move header file into asm/arch-visws
o visws: add missing mach_apic.h file
o visws: pci support
o visws: core
o visws: framebuffer driver update
o visws: sound update
o visws: MAINTAINERS file update
o visws: i386/KConfig update
o fix a visws compile warning
o consolidate and cleanup profiling code
o more ia32 profiler cleanups
o TTY module refcounting fix
o remove (start|end)_lazy_tlb()
o lib/idr.c 64-bit fixes
o posix-timers: fix callback address truncation
o Keep interrupts enabled in exit path
o Don't call mmdrop under a spinlock
o ppc64: Someone removed NR_syscalls from <linux/sys.h>
o ppc64: fix the build for posix timer changes
o pnp compile fix
o make drivers/pnp/interface.c compile
Andries E. Brouwer <[email protected]>:
o remove BSD_PARTITION
Andy Grover <[email protected]>:
o ACPI: Use extended IRQ resource type when setting IRQs on link
devices to more than IRQ 15 (Juan Quintela)
o ACPI: Properly handle an ISO reassigning the ACPI interrupt. Big
thanks to John Stultz.
o ACPI: Factor common code out of an if/else
o ACPI: *really* fix ISO SCI override support (thanks again to John
Stultz)
o ACPI: Change NUMA maintainer email
o ACPI: Eliminate use of acpi_gpl_gpe_number_info (Matthew Wilcox)
o ACPI: Support translation attribute (Bjorn Helgaas)
o ACPI: Add ability to override predefined object values (Ducrot
Bruno)
o ACPI: Decrease size of override's static array, add a define for
the length, and print a msg if used
o ACPI: Fix printk output (Jochen Hein)
o ACPI: Misc interpreter improvements
o ACPI: misc cleanups
o ACPI: Change license from GPL to dual GPL and BSD-style
o ACPI: Toshiba ACPI device update (John Belmonte)
Anton Blanchard <[email protected]>:
o ppc64: IPIs must run with interrupts off so tag them with
SA_INTERRUPT
o ppc64: fix some -Wundef warnings
o ppc64: allow pSeries LPAR insert_pte to fail
o ppc64: fix compile warning
o ppc64: update for recent changes that require switch_to to return
prev
o ppc64: add some bluetooth ioctls and clean up some warnings
o ppc64: Add posix timer syscalls
o ppc64: zero all registers in ELF_PLAT_INIT now we dont in
start_thread
o ppc64: quieten new boot wrapper to match old 2.5 one
o ppc64: defconfig update
Art Haas <[email protected]>:
o [IPV4]: Reformat ipv4_route_table
o C99 initializer for net/ipv6/icmp.c
o C99 initializer for net/rose/sysctl_net_rose.c
o C99 initializer for net/netrom/sysctl_net_netrom.c
o C99 initializer for net/rxrpc/sysctl.c
Brian Gerst <[email protected]>:
o Use mempool_alloc/free_slab
o Clean up list head usage in sysrq.c
o Trival patch to i386 enter_lazy_tlb()
o Remove checkhelp.pl and header.tk
o remove old double fault handler
o Unused variable warning in ac97_codec.c
o Better test for GCC alignment options
o Fix up slabinfo code
Christoph Hellwig <[email protected]>:
o further sim710 updates
o fix wd7000 for scsi command block changes
o remove cpqioctl.c
o remove remaining queueing-related code from scsi.c to
o remove eata_dma driver
o remove scsi_set_pci_device
o remove some dead mtrr code
o handles possible failures in scsi initialization
o two new device list entries
o wd7000 updates
o remove an escaped __MOD_DEC_USE_COUNT
o eata_pio updates
o fix that devfs mess
o drop scsi_register_blocked_host()
o move over exposing host attributes from sg/procfs to sysfs
o Coding Style police for scsi_error.c
o [NET]: Remove __NO_VERSION__ from networking code
Christoph Hellwig <[email protected]>:
o [XFS] make pagebuf_delwri_queue static
o [XFS] insert dirty buffers at the tail of the inode queue
o [XFS] Under heavy load, there are not enough hash buckets to deal
with the number of metadata buffers. Use the same techniques as the
regular linux buffer cache here.
o i2c sanity
o get rid of some kdevname abuse
o kill EXPORT_NO_SYMBOLS
o try_module_get(THIS_MODULE) is bogus
o fix module refcounting of pcmcia socket drivers
Dave Hansen <[email protected]>:
o make io_apic.c use named initializers
David Brownell <[email protected]>:
o USB: sg_complete() warning downgrade
o USB: USB keyboard works after reboot (ehci-hcd)
David S. Miller <[email protected]>:
o [NET]: Fix length in skb_padlen
o [SPARC]: Add timer_t and clockid_t
o [SOUND]: ac97_codec.c needs linux/pci.h
o [SPARC]: Fixup asm/ide.h headers for Alans recent IDE merge
o [IPSEC]: Move xfrm6 policy code to net/ipv4/xfrm_policy.c
o [IPSEC]: Export xfrm6 type registry interfaces
o [IPSEC]: Remove xfrm6 exports from ipv6_syms.c
o [SPARC64]: oprofile/timer_int.c needs linux/profile.h
o [SPARC64]: Be like Alpha and turn on -Werror in sparc64
subdirectories
o [SPARC64]: Implement STICK synchronization using ia64 port
algorithms
o [NETLINK]: Remove buggy and useless rcv queue wakeup tests
o [BLUETOOTH]: net/bluetooth/bnep/sock.c needs linux/init.h
o [IPV6]: Cure typo in ipv6_addr_prefix
o [IPV6]: ipv6_count_addresses is static
o [IPV{4,6}]: Make icmp_socket per-cpu and simplify locking
Dominik Brodowski <[email protected]>:
o cpufreq: move Kconfig entries (Marc-Christian Petersen)
o cpufreq: x86 driver updates (speedstep, longrun, p4-clockmod)
o pcmcia: add socket_offset for multiple pci_sockets, correct
suspend&resume
Douglas Gilbert <[email protected]>:
o aic79xxx_osm.c in 2.5.59-bk3
o scsi_mid_low_api.txt
Duncan Sands <[email protected]>:
o USB speedtouch: trivial speedtouch changes
o USB speedtouch: Fix atmsar memory leak
o USB speedtouch: fix speedtouch micro race
o USB speedtouch: Fix speedtouch maxi race
o USB speedtouch: Update CREDITS and MAINTAINERS
o USB speedtouch: add help text for speedtouch
o USB speedtouch: speedtouch 330 support
o USB speedtouch: minor speedtouch changes
o USB speedtouch: replace speedtouch crc routines
o USB speedtouch: speedtouch stability fix fix
o USB speedtouch: speedtouch cleanups
Duncan Sands <[email protected]>:
o USB speedtouch: Even more trivial speedtouch change
o USB speedtouch: yet another trivial speedtouch change
o USB speedtouch: infrastructure for new speedtouch send path
o USB speedtouch: expose crc defs to speedtouch
o USB speedtouch: More infrastructure for new speedtouch send path
o USB speedtouch: on-the-fly AAL5/ATM encoding for speedtouch
o USB speedtouch: new speedtouch send path
o USB speedtouch: speedtouch dead code elimination
o USB speedtouch: Missing speedtouch bits
Eric Sandeen <[email protected]>:
o [XFS] Remove unused init_spinlock #define
George Anzinger <[email protected]>:
o POSIX clocks & timers
Greg Kroah-Hartman <[email protected]>:
o USB: add "present" flag to usb_device structure
o USB serial: fix locking logic
o USB: serial core fix to solve ordering issues when destroying our
objects
o USB: added sched.h to usb.h
o USB: usbnet driver also needs the crc32 code
o PCI: remove large stack usage in pci_do_scan_bus()
o PCI i386: remove large stack usage in pci_sanity_check()
o PCI i386: remove large stack usage in pcibios_fixup_peer_bridges()
Greg Ungerer <[email protected]>:
o init sighand in m68knommu init_task
o add exception table support for m68knommu architecture
o m68knommu cacheflush.h cleanup
o fixup use of sighand in m68knommu signal.c
o bounds check and no argv/envp support for binfmt_flat load
o add extable.c to Makefile for m68knommu architecture
o fix m68knommu/ColdFire serial port hang
o add m68knommu serial console support into tty_io.c
o add missing m68knommu/68VZ328/ucdimm/config.c
o reformat m68knommu 68360/uCquicc crt0_rom.S
o create common vector setup code for m68knommu/ColdFire
o reformat m68knommu 68328/pilot crt0_rom.S
o create an architecture specific flat header for v850
o include the architecture flat file header in common flat header
o inline unsued functions for MMUless configuration
o reformat m68knommu 68360/uCquicc crt0_ram.S
o use local RODATA setup for m68knommu linker script
o clean up compiler warnings in m68knommu machdep.h
o remove duplicate memory size option in m68knommu Kconfig
o fix text and data sizing in MMUless task_nommu.c
o create an architecture specific flat header for m68knommu
o add missing page_referenced() for MMUless configs
Hideaki Yoshifuji <[email protected]>:
o [AF_KEY]: Add missing credit
o [NET]: Convert dst->{input,output}() fully to dst_{input,output}()
o [IPV6]: Export ip6_route_me_harder for netfilter and add
ipv6_addr_prefix
o [IPV6]: Privacy Extensions for Stateless Address Autoconfiguration
Hirofumi Ogawa <[email protected]>:
o [IPV4]: rt_cache_stat initialization fix
Ivan Kokshaysky <[email protected]>:
o numa fixes
James Bottomley <[email protected]>:
o Minor fixes to scsi/sim710.c
o Add #include <linux/pci.h> to scsi_lib.c (needed for
PCI_DMA_BUS_IS_PHYS)
o Add back SCSI subsystem initialisation prints
o Remove dead code from 53c700
o Update 53c700 error handling
o Correct uninitialised timer in scsi_error.c
o update sim710.c for new eisa sysfs registration returns
James Morris <[email protected]>:
o [NET]: Add myself as co-maintainer
o [NETLINK]: Un-duplicate rcv wakeup logic
Jay Vosburgh <[email protected]>:
o [BONDING]: Add MAINTAINERS entry
Jeff Garzik <[email protected]>:
o report luns default
Jeff Garzik <[email protected]>:
o [netdrvr 8139too] add to the list of supported boards
o [netdrvr] Remove superceded wireless driver aironet4500
o Move the old wireless drivers into drivers/net/wireless
Johannes Erdfelt <[email protected]>:
o usb_get_driver_np() gives wrong driver name (usb_mouse)
John Levon <[email protected]>:
o oprofile author needs to learn C
Keith Owens <[email protected]>:
o [XFS] XFS patches from 2.5.60-mm1
Kunihiro Ishiguro <[email protected]>:
o [IPSEC]: Add ipv6 support to ipsec netlink sockets
Linus Torvalds <[email protected]>:
o Add doublefault handling with a task gate
o Fix x86 "switch_to()" to properly set the previous task
information, which is needed to keep track of process usage counts
correctly and efficiently.
o Fix "make clean" to remove scripts/elfconfig.h
o We don't need to wait for task in-activity in release_task() any
more, since we now properly reference-count the allocations and
thus can't be freeing the thread structures from underneath the
task running on another CPU.
o Fix up incorrect __exit marking for SCSI functions that are called
from non-exit code.
o Add support for forcing default signal handlers to
flush_signal_handlers()
o Add ndelay() compatibility macro. If the architecture doesn't
define ndelay(), fall back on udelay().
Maksim Krasnyanskiy <[email protected]>:
o Add BCSP TXCRC option to drivers/bluetooth/Kconfig
o Bluetooth Kconfigs. Cleanup things missed by automatic converter
o l2cap_do_connect() should be static
o Fix L2CAP client/server PSM clash
o Fix hci_get_dev_list() for big endian machines
o Ordinary users are not allowed to use raw L2CAP sockets
o BNEP extension headers handling fix
o Remove duplicated include in HCI H4 driver
o Convert Bluetooth HCI devices to new module refcounting
o arch/sparc64/ioctl32.c Put Bluetooth ioctls at the end, right
before the translation table.
o Kill old BNEP ioctls
o [Bluetooth] Cleanup and fix __init and __exit functions.
Marcel Holtmann <[email protected]>:
o [Bluetooth] Don't do wakeup if protocol is not set
o [Bluetooth] Fix operator precedence for modem status
o [Bluetooth] Check for signals while waiting for DLC
o [Bluetooth] Fix some bits of the modem status handling
o [Bluetooth] Free skbs with kfree_skb() instead of kfree()
o [Bluetooth] Add HCI id for Bluetooth PCI cards
o [Bluetooth] Fix another operator precedence for modem status
o [Bluetooth] Add HCI UART PC Card driver
o [Bluetooth] Fix return with a value, in function returning void
o [Bluetooth] Add the needed call of init_timer()
o [Bluetooth] Remove EXPORT_NO_SYMBOLS
o [Bluetooth] Don't use %d notation for non devfs name field of
tty_driver
o [Bluetooth] Another cleanup of the Kconfig files
o [Bluetooth] Convert dlci and channel variables to u8
o [Bluetooth] Add some COMPATIBLE_IOCTL for SPARC64
o [Bluetooth] Make READ_VOICE_SETTING available for normal users
o [Bluetooth] Replace info message about SCO MTU with BT_DBG
o [Bluetooth] Remove wrong check for size value in rfcomm_wmalloc()
o [Bluetooth] Disable HCI flow control for vendor commands
o [Bluetooth] Get rid of hci_send_raw()
Mark W. McClelland <[email protected]>:
o USB: ov511 bugfixes/cleanup
Martin Schwidefsky <[email protected]>:
o s390: base fixes
o s390: common i/o layer
o s390: ctc network driver
o s390: dasd block device driver
o s390: documentation
o s390: unified extable code
o s390: gcc 3.3 adaptions
o s390: iucv network driver
o s390: configuration
o s390: bogus zfcp header, small scsi change
o s390: trivial bug fixes
o s390: kernel module loader
o s390: 31 bit compatability layer
Matthew Wilcox <[email protected]>:
o Two small fixes for sym53c8xx_2
o [wireless airo] call pci_enable_device, pci_set_master as needed
Mike Anderson <[email protected]>:
o fix scsi/ppa.c
o scsi_error handler update. (1-4)
o scsi/pcmcia compile fix
o scsi_error update take 2
Miles Bader <[email protected]>:
o Handle null OLD argument in nb85e_uart's nb85e_uart_set_termios
function
o Fix up some left-over sig->sighand issues on the v850
o Add v850 version of `init_irq_proc' for sysctl
o Set child process initial stack-pointers correctly on the v850
o Remove unused compile-time configuration options on v850
o Use .balign rather than .align for v850 asm funcs
o v850 kernel entry fixes and cleanup
o Implement <asm/bug.h> for v850
o Add a v850 config option to pass illegal insn traps to the kernel
o Force v850 interrupt vector parts into their correct locations
Nathan Scott <[email protected]>:
o [XFS] Extra check on the mount path - ensure we don't attempt to
mount XFS fs's with sector sizes smaller than those the device
supports. Tripped a BUG in pagebuf, should now be resolved.
Neil Brown <[email protected]>:
o Bounds checking for NFSv3 readdirplus
o Keep track of which page is the 'tail' of an NFSd reply
o Fix handling of error code in NFSv4 replies
o Fix problem where knfsd wouldn't release filesystem on unexport
o Make kNFSd pre/post_[acm]time use struct timespec
o Convert fs/nfsctl.c to use C99 named initiailzers
o Fix bug in md superblock sanity checking
o linear.c fix for gcc bug
o Small bug fix for multipath
o C99 initializers for drivers/md/md.c
o Add name of md device to name of thread managing that device
o Provide a 'safe-mode' for soft raid
Oliver Neukum <[email protected]>:
o USB: added device id for kaweth
o USB: kaweth fix
Patrick McHardy <[email protected]>:
o [IPV{4,6}]: lru queue for ip_fragment evictor
Paul Mackerras <[email protected]>:
o PPC32: add system calls for POSIX timer stuff
o PPC32: provide __ide_mm_insw etc
o PPC32: Add definition of ndelay()
o PPC32: Make switch_to return the previous task in the `last'
argument
Pete Zaitcev <[email protected]>:
o [SPARC]: Kconfig help update
o [SPARC]: Add rtc_lock
Petko Manolov <[email protected]>:
o USB: pegasus update (2.5)
Petri Koistinen <[email protected]>:
o update README file to current realities
Randy Dunlap <[email protected]>:
o USB: cdc-acm memory leak
o scsi/imm.c compile errors in 2.5.60
o fix scsi/aha15*.c for 2.5.60
o convert /proc/io{mem,ports} to seq_file
Richard Henderson <[email protected]>:
o [PCI] Correctly size hardwired empty BARs
o [OPROFILE] First cut at oprofile support for Alpha
o [ALPHA] Mirror i386 change to include asm-generic/ide_iops.h
o [ALPHA] Add clockid_t and timer_t for posix clocks
o [OPROF] Update for change to cpu_type interface
o [OPROF] Fix arguments to oprofile_add_sample
o [OPROF] Display banner message on startup
o [OPROF] Fix alpha/ev6 proc_mode setting
o [OPROF] Tidy model handle_interrupt handling
o [ALPHA] Sync with i386 ptrace changes
o [ALPHA] Implement ndelay
o [ALPHA] Collection of warning fixes
o [ALPHA] Turn on -Werror in alpha subdirectories
o [ALPHA] Use more compiler builtins instead of inline assembly
o [ALPHA] Fix typo in __kernel_cmpbge
o [ALPHA] More CFLAGS a-la Sam
o [ALPHA] Fix switch_to semantics wrt LAST
o eliminate warnings in generated module files
Rob Radez <[email protected]>:
o [SPARC] Fix compilation of sunsu.c and sunzilog.c
Russell King <[email protected]>:
o Alternative tty fasync fix
o kernel/pm.c requires <linux/init.h>
o Remove dummy cb_config() and cb_release()
o Remove unused "dev" argument from cb_setup_cis_mem
o Remove "fn" argument from read_cb_mem()
o Remove pci_{read,write}[bwl]
o Remove stack allocation of struct pci_dev
o Always re-read vendor for each function
Rusty Russell <[email protected]>:
o Enable signals for usermode helpers
Sam Ravnborg <[email protected]>:
o Alpha CFLAGS fix
Sridhar Samudrala <[email protected]>:
o [IPV4/IPV6]: Fix to avoid overriding TCP/UDP with a new protocol of
same type
Stelian Pop <[email protected]>:
o sonypi and input subsystem
o meye suspend/resume capabilities
o use correct gcc flags when compiling for Crusoe
Stephen Lord <[email protected]>:
o [XFS] cleanup delayed allocate write path a little and fix some
small bugs in there.
o [XFS] fix a couple of memory leaks found by stanford checker
Stephen Rothwell <[email protected]>:
o [COMPAT]: compat_sys_futex sparc64
Steve French <[email protected]>:
o Remove compiler warnings and allow reconnection of tids after
temporary tcp session failure
Steve French <[email protected]>:
o Fix caching of directory inodes (was always querying server on
directory inodes). Reconnect fids (file ids) after session
reconnection e.g. due to temporary server or network failure
o Fix freeing of captive thread at unmount time (which was causing
unmount warning).
o Fix search handle leak on search rewind. Fix setting of uid and
gid in local inode (not just remote)
Steven Cole <[email protected]>:
o USB: trivial help texts for drivers/usb/serial/Kconfig
o spelling fix accessable -> accessible
o spelling fix adress/addres -> address
o spelling fix for interupt -> interrupt
o spelling fix for compatable -> compatible
o spelling fix for propogate -> propagate
o various spelling fixes
> <[email protected]>:
> o Fusion Driver 2.05.00.03 against 2.5.62bk3
This update is broken and strange in many ways, it would have been
nice if you actually sent this to some list for review before submitting..
(1) there's stuff like:
#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,5,58)
#define mpt_inc_use_count()
#define mpt_dec_use_count()
#else
#define mpt_inc_use_count() MOD_INC_USE_COUNT
#define mpt_dec_use_count() MOD_DEC_USE_COUNT
#endif
but even if the old-style refcounting is deprecated now you can't
just simply remove it! you need to rearchitecture your code to
work with try_module_get/module_put
(2) and like this:
+#if LINUX_VERSION_CODE < KERNEL_VERSION(2,5,59)
MODULE_PARM(PortIo, "0-1i");
MODULE_PARM_DESC(PortIo, "[0]=Use mmap, 1=Use port io");
+#endif
Again. just because old-style module paramters are deprecated you can't
just remove them without replacement, use module_param() instead.
(3) you remove backward copatiblity code in one place but add lots more
in other places. this doesn't make sense - either you have a nice or
all releases driver, but not one that doesn't work everywhere _and_ looks
horrible.
> George Anzinger <[email protected]>:
> o POSIX clocks & timers
Care to explain what FOLD_NANO_SLEEP_INTO_CLOCK_NANO_SLEEP
is supposed to do? It's always defined in signal.h, so we can
aswell get rid of it..
And what's this:
#ifndef div_long_long_rem
+#include <asm/div64.h>
+
+#define div_long_long_rem(dividend,divisor,remainder) ({ \
+ u64 result = dividend; \
+ *remainder = do_div(result,divisor); \
+ result; })
+
+#endif /* ifndef div_long_long_rem */
Any reason you can't just use do_div directly like everyone else? :)
Reading Oops report from the terminal
kernel BUG at drivers/net/tulip/de2104x.c:925!
Oops: Exception in kernel mode, sig: 4
NIP: C013A454 LR: C013A454 SP: C0277E10 REGS: c0277d60 TRAP: 0700 Not tainted
MSR: 00089032 EE: 1 PR: 0 FP: 0 ME: 1 IR/DR: 11
TASK = c0231530[0] 'swapper' Last syscall: 120
GPR00: C013A454 C0277E10 C0231530 0000002F 00000001 C0277CC8 C0273A00 C02B0000
GPR08: 00001669 00000000 00000000 C0277D40 4000C088 00000000 00000000 00000000
GPR16: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
GPR24: 00000000 00000000 00000002 00001032 C03DA000 00009032 FFFFFFCE C03DA1C0
Call trace: [c013a794] [c0020710] [c001b830] [c0007e80] [c00061c4] [c0007b
Kernel panic: Aiee, killing interrupt handler!
In interrupt handler - not syncing
Using defaults from ksymoops -t elf32-powerpc -a powerpc:common
Warning (Oops_read): Code line not seen, dumping what data is available
>>NIP; c013a454 <de_set_media+48/1f0> <=====
This is on ppc/prep
--
Remco
Compile statistics: 2.5.63
Note that gcc 3.2 was used for all of these statistics.
2.5.62 2.5.63
-------------------- -----------------
bzImage (defconfig) 18 warnings 15 warnings
0 errors 0 errors
bzImage (allmodconfig) 33 warnings 29 warnings
9 errors 9 errors
modules (allmodconfig) 2514 warnings 2426 warnings
105 errors 128 errors
Compile statistics have been for kernel releases from 2.5.46 to 2.5.63
at: http://www.osdl.org/archive/cherry/stability
I am also compiling nightly views of Linus' linux-2.5 bitkeeper tree.
Results can be found at:
http://www.osdl.org/archive/cherry/stability/linus-tree/running.txt
John
On Mon, Feb 24, 2003 at 11:32:07AM -0800, Linus Torvalds wrote:
>...
> Summary of changes from v2.5.62 to v2.5.63
> ============================================
>...
> <[email protected]>:
> o [ATM]: use sock timestamp
>...
This change broke the compilation of fore200e.c:
<-- snip -->
gcc -Wp,-MD,drivers/atm/.fore200e.o.d -D__KERNEL__ -Iinclude -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-pipe -mpreferred-stack-boundary=2 -march=k6
-Iinclude/asm-i386/mach-default -nostdinc -iwithprefix include -g
-DKBUILD_BASENAME=fore200e -DKBUILD_MODNAME=fore_200e -c -o
drivers/atm/fore200e.o drivers/atm/fore200e.c
drivers/atm/fore200e.c: In function `fore200e_push_rpd':
drivers/atm/fore200e.c:1135: structure has no member named `timestamp'
drivers/atm/fore200e.c:1136: structure has no member named `timestamp'
make[2]: *** [drivers/atm/fore200e.o] Error 1
<-- snip -->
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
In message <[email protected]>,Adrian Bunk writes:
>drivers/atm/fore200e.o drivers/atm/fore200e.c
>drivers/atm/fore200e.c: In function `fore200e_push_rpd':
>drivers/atm/fore200e.c:1135: structure has no member named `timestamp'
>drivers/atm/fore200e.c:1136: structure has no member named `timestamp'
it shouldnt be doing that. you only need to set the timestamp in the
skb. i see the eni driver does the same thing. i will see about
a patch for that shortly.
Index: linux/drivers/atm/fore200e.c
===================================================================
RCS file: /home/chas/CVSROOT/linux/drivers/atm/fore200e.c,v
retrieving revision 1.1.1.1
diff -u -d -b -w -r1.1.1.1 fore200e.c
--- linux/drivers/atm/fore200e.c 20 Feb 2003 13:45:03 -0000 1.1.1.1
+++ linux/drivers/atm/fore200e.c 25 Feb 2003 14:42:06 -0000
@@ -1132,8 +1132,7 @@
return;
}
- do_gettimeofday(&vcc->timestamp);
- skb->stamp = vcc->timestamp;
+ do_gettimeofday(&skb->stamp)
#ifdef FORE200E_52BYTE_AAL0_SDU
if (cell_header) {
On Tue, Feb 25, 2003 at 09:53:57AM -0500, chas williams wrote:
> missed a semicolon--it should be:
>
> Index: linux/drivers/atm/fore200e.c
> ===================================================================
> RCS file: /home/chas/CVSROOT/linux/drivers/atm/fore200e.c,v
> retrieving revision 1.1.1.1
> diff -u -d -b -w -r1.1.1.1 fore200e.c
> --- linux/drivers/atm/fore200e.c 20 Feb 2003 13:45:03 -0000 1.1.1.1
> +++ linux/drivers/atm/fore200e.c 25 Feb 2003 14:42:06 -0000
> @@ -1132,8 +1132,7 @@
> return;
> }
>
> - do_gettimeofday(&vcc->timestamp);
> - skb->stamp = vcc->timestamp;
> + do_gettimeofday(&skb->stamp);
>
> #ifdef FORE200E_52BYTE_AAL0_SDU
> if (cell_header) {
Thanks, this patch fixed it.
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
Christoph Hellwig wrote:
>>George Anzinger <[email protected]>:
>> o POSIX clocks & timers
>
>
> Care to explain what FOLD_NANO_SLEEP_INTO_CLOCK_NANO_SLEEP
> is supposed to do? It's always defined in signal.h, so we can
> aswell get rid of it..
It is there in case some one might want nanosleep to NOT be folded
into clock_nanosleep. For a while this was a moving target and I got
a bit paranoid :) I see no real reason to keep it...
>
> And what's this:
>
> #ifndef div_long_long_rem
> +#include <asm/div64.h>
> +
> +#define div_long_long_rem(dividend,divisor,remainder) ({ \
> + u64 result = dividend; \
> + *remainder = do_div(result,divisor); \
> + result; })
> +
> +#endif /* ifndef div_long_long_rem */
>
> Any reason you can't just use do_div directly like everyone else? :)
Actually I have coded a better function as part of the expanded
high-res-timers which does a 64-bit/32-bit div in a much cleaner way.
Again, it is part of the full high-res-timers patch. I have been
thinking of submitting the complete set of these math routines outside
of the high-res patch. They are designed to make scaled math easy. I
have both a generic and a i386 header file, but they still need a
bit of testing.
The issue is getting around the C limitation of not being able to use
the div and mpy instructions that take 64-bits/32-bits and return 2
32-bit results and the mpy which takes 2 32-bit operands and returns a
64-bit result.
For scaled operations, they also roll a shift into the mix in an
efficient way (i.e. a small inline asm function).
--
George Anzinger [email protected]
High-res-timers: http://sourceforge.net/projects/high-res-timers/
Preemption patch: http://www.kernel.org/pub/linux/kernel/people/rml