Changes all over the map.
The most noticeable one may well be the new and much improved elevator by
Jens Axboe, this one makes a big difference at least to me.
And Andrew found a nasty SMP deadlock on the tasklist lock.
And Ingo's been busy again, fixing some more threading issues he found
(including the much-talked-about futex thing).
Other stuff all over the map: USB, JFS, XFS, networking, debugging etc.
Linus
---
Summary of changes from v2.5.38 to v2.5.39
============================================
Albert Cranford <[email protected]>:
o i2c core/dev/proc cleanups, and a proc-related fix
<[email protected]>:
o 3ware driver update for 2.5.35
<[email protected]>:
o Update for JMTek USBDrive
<[email protected]>:
o net/sched/sch_htb.c: Verify classid and direct_qlen properly
<[email protected]>:
o Compile fixes for alpha arch
<[email protected]>:
o MODULE_LICENSE for i82092 pcmcia
<[email protected]>:
o Reiserfs: Fix alloc= mount option parser
Christoph Hellwig <[email protected]>:
o XFS: XFS: Use do_gettimeofday() instead of racy direct access to
xtime
o XFS: Small comment corrections/updates
o XFS: Don't include <asm/softirq.h> in page_buf.c
o XFS: XFS: Make pagebuf use the generic xfs ASSERT() instead of it's
own assert()
o XFS: XFS: Sanitize some names in xfs_aops.c, especially a less
offending name for linvfs_pb_bmap
o XFS: XFS: Simplify xfs_dir_lookup_int
o XFS: XFS: Remove some dead prototypes in pagebuf
o XFS: Switch to mpage_readpage
o XFS: More mount code cleanups
o XFS: Fix the mount-cleanup for single-subvolume filesystems
o XFS: Implement readv/writev
<[email protected]>:
o USBLCD updates
<[email protected]>:
o PnP BIOS ESCD sanity check
<[email protected]>:
o SCTP: Resync with LKSCTP tree
Stephen Lord <[email protected]>:
o XFS: Fold some code paths together in the xfs fsync implementation
o XFS: Avoid writing data out to disk twice!
<[email protected]>:
o #include <linux/version.h> missing in drivers/usb/host/ohci-hcd.c
<[email protected]>:
o ISA PnP change Jens Thoms Toerring
<[email protected]>
<[email protected]>:
o USB 2.0 HDD Walker / ST-HW-818SLIM usb-storage fix
<[email protected]>:
o usb-storage: fix return codes
<[email protected]>:
o USB: clean up the error logic for open() in the usb-serial driver
o usb whiteheat driver update
Ted Ts'o <[email protected]>:
o loop device broken in 2.5.38
Matthew Wilcox <[email protected]>:
o flock_lock_file livelock fix
o fix file_lock_cache leak
o Remove NetNews.html
<[email protected]>:
o [IPv6]: Verify ND options properly
<[email protected]>:
o USB storage: Another (!) patch for the abort handler
Adrian Bunk <[email protected]>:
o gendisk typo fixes
Alexander Viro <[email protected]>:
o Re: Linux 2.5.38
o tapeblock blk_size removal
o kills CURRENT in floppy.c
o Lindent pd.c
o cleanup of pd.c
o gendisk for amiflop
o gendisk for ataflop
o gendisk for z2ram
o gendisk for mtdblock
o compile fixes for ftl
o blk_size[] is gone
Andi Kleen <[email protected]>:
o Minor ACPI changes for x86-64
o Fix ELF name for x86-64
o Hammer aperture driver for 2.5.38
o PCI ID for AMD 8151 AMD bridge
o disable early console in console_init
Andrew Morton <[email protected]>:
o 64-bit type correctness in filemap.c
o fix ext3 in data=writeback mode
o don't hold mapping->private_lock while marking a page dirty
o infrastructure for monitoring queue congestion state
o use the queue congestion API in ext2_preread_inode()
o use the congestion APIs in pdflush
o low-latency page reclaim
o direct-io bandaid
o hugetlb fix
o mprotect_fixup fix
o prepare_to_wait/finish_wait sleep/wakeup API
o use prepare_to_wait in VM/VFS
o slab reclaim balancing
o increase traffic on linux-kernel
o speed up sys_sync()
o tighter locking in pdflush
o export test_clear_page_dirty() to modules
Arnaldo Carvalho de Melo <[email protected]>:
o [LLC] remove not needed functions, use llc_sap_find in llc_sock
o [LLC] kill llc_ui_bh_find_sk_by_addr
o [UDPv6] fix udp_v6_get_port introduced by the sock splitup
o [LLC] use sk->state_change when p_flag is cleared or core state
changes
o [LLC] use the core lists to get info for /proc/net/llc
o [LLC] Make llc_save_primitive ready for dataunit/xid/test DGRAM
packets
o [LLC] move reason to the {station,sap,conn}_ev structs
o [LLC] kill sap->{ind,conf}, finally!
o [LLC] clean up the ui sending routines and core
o [LLC] use struct sock list members
o [LLC] remove sap->mac_pdu_q, not used at all
o [LLC] keep the skb in llc_sap_state_process
o [LLC] move sap->rcv_func call to llc_rcv
o [SNAP] make SNAP work again
o [X25] remove unneeded typedef x25_address
o [X25] make search functions that grab locks have just one exit
o [LLC] stop using the BKL
o [X25] handle return codes and code reoganization to have only one
exit in functions
o [X25] assorted code cleanup
o [X25] convert sysctl_net_x25 to use designated initializers
o [X25] code reorganization, eliminate duplicated code
Ben Collins <[email protected]>:
o RCS files exclusion (and add subversion)
Dave Kleikamp <[email protected]>:
o JFS: Fix off-by-one error in dbNextAG
o JFS: Fix problems with NFS
o JFS: detect and fix invalid directory index values
o JFS: Remove assert(i < MAX_ACTIVE)
David Brownell <[email protected]>:
o ohci-hcd, queue fault recovery + rm DEBUG
o ehci-hcd: update
o USB shutdown oopser
David Gibson <[email protected]>:
o Orinoco driver update
David S. Miller <[email protected]>:
o [BRIDGE]: Missing headers from ebtables merge
o net/ipv4/netfilter/ip_conntrack_proto_tcp.c: Include linux/string.h
Eric Sandeen <[email protected]>:
o XFS: Remove unused function xfs_vn_iget()
Greg Kroah-Hartman <[email protected]>:
o USB: fix for ezusb firmware download
o USB: made port_softint global for other usb-serial drivers to use
o USB: convert the usb-skeleton.c driver to work with the latest USB
core changes
o USB: fix ifnum usage that was missed in the previous irda-usb patch
o add hotplug support to the driver core for devices, if their bus
type supports it
o converted USB to use the driver core's hotplug call
o converted PCI to use the driver core's hotplug call
Harald Welte <[email protected]>:
o [NETFILTER]: Trivial fixes
Ingo Molnar <[email protected]>:
o pidhash cleanups, tgid-2.5.38-F3
o de-xchg fork.c
o pgrp-fix-2.5.38-A2
o thread-flock-2.5.38-A3
o pidhash-2.5.38-A0
o exit-fix-2.5.38-E3
o exit-fix-2.5.38-F0
o kksymoops-2.5.38-C9
o virtual => physical page mapping cache
Ivan Kokshaysky <[email protected]>:
o Re: 2.5.36 IDE fixes
o another alpha update
o ALi and Cypress IDE fixes
James Morris <[email protected]>:
o net/ipv4/netfilter/ipchains_core.c: Use GFP_ATOMIC under ip_fw_lock
Javier Achirica <[email protected]>:
o airo wireless: use ETH_ALEN constant where appropriate
o airo wireless: disable card while prom flashing is in progress
[note - more work needs to be done here, but this is better than
nothing -jgarzik]
o airo wireless: more verbose MAC-enable errors
o airo wireless: power down on if down, define local 'ai' to fix
build
o airo wireless: fix "non-probe mode" setup
o airo wireless: Fixes signal level retrieval in SPY mode (releases
memory block after read out)
Jean Tourrilhes <[email protected]>:
o USB: convert the irda-usb driver to work properly with the new USB
core changes
Jean Tourrilhes <[email protected]>:
o More IrDA __FUNCTION__ cleanups, merged from -ac
o More IrDA __FUNCTION__ cleanups (contributed by Philipp Matthias
Hahn)
o 64-bitness fixes for IrDA irlan protocol code (fixing the new
hashbin code)
o Minor Wavelan wireless net driver fixes
Jeff Garzik <[email protected]>:
o s/schedule_timeout(0)/yield/ in cdu31a, sonycd535, sb1000, and
sis900 drivers
Jens Axboe <[email protected]>:
o more bio updates
o trm compile
o pdc4030
o bio_get_nr_vecs
o ide io scheduler thing
o deadline scheduler
o remove elevator_linus
o deadline ioscheduler cleanups
o io scheduler update
o more io scheduler updates
Kai Germaschewski <[email protected]>:
o ISDN: per net device auto hangup function
o ISDN: Switch isdn_net_hangup() to work on isdn_net_local *
o kbuild: clean up i386 subarch build
o kbuild: Fix modversions generation glitch
o ISDN: per-interface hangup timer
o kbuild: Clean up tags/TAGS targets
o kbuild: arch/alpha cleanup / O_TARGET removal
o kbuild: arch/arm cleanup / O_TARGET removal
o kbuild: arch/cris cleanup / O_TARGET removal
o kbuild: arch/ia64 cleanup / O_TARGET removal
o kbuild: arch/m68k cleanup / O_TARGET removal
o kbuild: arch/mips cleanup / O_TARGET removal
o kbuild: arch/mips64 cleanup / O_TARGET removal
o kbuild: arch/parisc cleanup / O_TARGET removal
o kbuild: arch/ppc cleanup / O_TARGET removal
o kbuild: arch/ppc64 cleanup / O_TARGET removal
o kbuild: arch/s390 cleanup / O_TARGET removal
o kbuild: arch/s390x cleanup / O_TARGET removal
o kbuild: arch/sh cleanup / O_TARGET removal
o kbuild: arch/sparc cleanup / O_TARGET removal
o kbuild: arch/sparc64 cleanup / O_TARGET removal
o kbuild: arch/um cleanup / O_TARGET removal
o ISDN: PPP cleanups
o ISDN: Kill isdn_net_autohup()
o ISDN: ISDN_GLOBAL_STOPPED cleanup
o ISDN: Use <linux/list.h> for list of phone numbers
o ISDN: Lock list of phone numbers appropriately
o kbuild: Convert missed L_TARGET references
o ISDN: Fix build when CONFIG_ISDN_TTY_FAX is not set
Linus Torvalds <[email protected]>:
o IDE: Try to use PCI dma_mask only if the device actually _is_ PCI
o Terminate a failed IO properly
o Merge with DRI CVS tree
o Avoid possibly busy-looping in mouse read
o Remove busy-wait for short RT nanosleeps. It's a random special
case and does the wrong thing for higher HZ values anyway.
o Make the ACPI SCI interrupt get the right polarity when it is
explicitly overridden in the MADT
o Avoid NULL ptr dereference on module names by always having a valid
name (base kernel: "").
o Update x86 defconfig to reflect new config options
o Add CVS files to the list of files ignored by "find".and make the
same ignore rules for "tar" as well.
o Simplify elevator algorithm, make it prefer reads heavily
[email protected] <[email protected]>:
o NUMA-Q fixes
Mikael Pettersson <[email protected]>:
o fix UP_APIC linkage problem in 2.5.3[78]
Nathan Scott <[email protected]>:
o XFS: XFS: Cleanup mount argument manipulation, sanitize
xfs_cmountfs and move the
Patrick Mochel <[email protected]>:
o Driver model: improve support for system devices
o Driver model: handle devices registered with ->driver set
o USB: fixup handling of generic USB driver
o driver model: add support for CPUs
o driver model: add support for multi-board systems
o driver model: add better platform device support
o add disk device class
Paul Mackerras <[email protected]>:
o fix null dereference in sys_mprotect
Petr Vandrovec <[email protected]>:
o [NCPFS]: 32->64bit sparc64 conversions
o Fix matroxfb compile when G450 support is not selected
o Fix matroxfb compile on m68k
Robert Love <[email protected]>:
o s/preempt_count()/in_atomic() in do_exit()
o remove preempt workaround in slab.c
o per-cpu data preempt-safing
Rusty Russell <[email protected]>:
o UP cpu_possible
o export cpu_callout_map for SMP modules
Tim Schmielau <[email protected]>:
o sb1000 net driver: kill float constant, time_after_eq() jiffies
cleanup
o trivial typo in drivers/ide/pci/sl82c105.c
o fix compares of jiffies
On Fri, 27 Sep 2002 15:02:48 -0700 (PDT) Linus Torvalds (LT) wrote:
LT> Summary of changes from v2.5.38 to v2.5.39
LT> ============================================
Following compile error with 2.5.39 has been in 2.5.38, too.
gcc -Wp,-MD,./.sleep.o.d -D__KERNEL__ -I/usr/src/linux-2.5.39/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing
-fno-common -pipe -mpreferred-stack-boundary=2 -march=athlon
-I/usr/src/linux-2.5.39/arch/i386/mach-generic -nostdinc -iwithprefix include
-D_LINUX -I/usr/src/linux-2.5.39/drivers/acpi/include -DKBUILD_BASENAME=sleep
-c -o sleep.o sleep.c
sleep.c: In function `acpi_system_restore_state':
sleep.c:63: warning: implicit declaration of function `acpi_restore_state_mem'
sleep.c: In function `acpi_system_save_state':
sleep.c:132: warning: implicit declaration of function `acpi_save_state_mem'
sleep.c:135: warning: implicit declaration of function `acpi_save_state_disk'
sleep.c: In function `acpi_system_suspend':
sleep.c:209: warning: implicit declaration of function `do_suspend_lowlevel'
sleep.c: In function `acpi_suspend':
sleep.c:237: `acpi_wakeup_address' undeclared (first use in this function)
sleep.c:237: (Each undeclared identifier is reported only once
sleep.c:237: for each function it appears in.)
sleep.c: In function `acpi_sleep_init':
sleep.c:707: `sysrq_acpi_poweroff_op' undeclared (first use in this function)
make[3]: *** [sleep.o] Error 1
-Udo.
On Sat, Sep 28, 2002 at 12:27:03AM +0200, Udo A. Steinberg wrote:
> On Fri, 27 Sep 2002 15:02:48 -0700 (PDT) Linus Torvalds (LT) wrote:
>
> LT> Summary of changes from v2.5.38 to v2.5.39
> LT> ============================================
>
> Following compile error with 2.5.39 has been in 2.5.38, too.
>
[zap]
Try compiling with a clean tree, I saw the same error here too, but it
disappeared after cleaning the tree.
--
Jasper Spaans
http://jsp.ds9a.nl/contact/
``Got no clue? Too bad for you.''
On Sat, 2002-09-28 at 01:02, Linus Torvalds wrote:
>
> Changes all over the map.
Looks like the attached patch didn't work its way in there. Please
apply. The patch is good for 2.5.39 as well.
The current code in sound/oss/audio.c tries to access spinlocks through
a garbage pointer in many places. This causes bad things to happen on
SMP machines.
Regards,
MikaL
On Fri, 27 Sep 2002, Linus Torvalds wrote:
>
> Changes all over the map.
The following patch is still needed:
--- linux-2.5-tm/sound/oss/v_midi.h.old Sun Sep 22 10:01:43 2002
+++ linux-2.5-tm/sound/oss/v_midi.h Sun Sep 22 10:02:48 2002
@@ -11,5 +11,6 @@
int input_opened;
int intr_active;
void (*midi_input_intr) (int dev, unsigned char data);
+ spinlock_t lock;
} vmidi_devc;