Hello,
Here goes -pre2 with a big number of changes, including the new aic7xxx
driver.
I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
short.
Have fun
Summary of changes from v2.4.22-pre1 to v2.4.22-pre2
============================================
<[email protected]>:
o fix bug in drivers/net/cs89x0.c:set_mac_address()
o [IPV4]: Trim the includes used in util.c
<[email protected]>:
o [e1000] TSO fix
o [e1000] Added ethtool test ioctl
o [e1000] Added support for 82546 Quad-port adapter
o [e1000] Removed strong branded device ids
o [e1000] Fixed LED coloring on 82541/82547 controllers
o [e1000] Miscellaneous code cleanup
o [e1000] Whitespace cleanup
<[email protected]>:
o [netdrvr tulip] support DM910x chip from ALi
<[email protected]>:
o [IPV{4,6}]: Fix "slow multicast on 2.5.69" bug
<[email protected]>:
o [NETFILTER]: Really search _backwards_ to find the oldest unreplied connection to evict
<[email protected]>:
o current bk ipmi build fix
<[email protected]>:
o [NET]: Fix OOPSes with RSVP
<[email protected]>:
o [NETFILTER]: Fix two issues in the newnat core, with help from [email protected]
<[email protected]>:
o sd.c: set data direction to SCSI_DATA_NONE for START_STOP
<[email protected]>:
o Add XRAYTEX to SCSI whitelist
o sd.c: Backport wild spin loop mitigation from 2.5
o Backport from 2.5: scsi allow devices to restrict start on add
<[email protected]>:
o [NETFILTER]: Cosmetic changes
o [NETFILTER]: ip{,6}tables enhancement, add new /proc/net files
o [NETFILTER]: Fix conntrack master_ct refcounting
<[email protected]>:
o new eepro100 PDI ID
<[email protected][holtmann]>:
o [Bluetooth] Add CAPI message transport protocol support
<[email protected]>:
o PPC32: Fix /proc/sys/kernel/l2cr on newer CPUs
<[email protected]>:
o [NETFILTER]: Fix processor shifts in lockhelp.h
<[email protected]>:
o ISDN: [PATCH] memory leak in tpam_queues.c
<[email protected]>:
o hfs-readonly-fix.diff
<[email protected]>:
o [NETFILTER]: Fix ip6tables alignment (64bit archs)
o [NETFILTER]: Fix endianness bugs in conntrack
o [NETFILTER]: Fix endianness bugs in ipt_nat
<[email protected]>:
o [netdrvr amd8111e] interrupt coalescing, libmii, bug fixes
o [netdrvr amd8111e] link against mii lib
o [netdrvr amd8111e] bug fix: move stats update after irq free
<[email protected]>:
o [wireless airo] fix end-of-array test
<[email protected]>:
o [NETFILTER]: Add iptables "recent" module
<[email protected]>:
o [bonding] ABI versioning
o [bonding] better 802.3ad mode control, some cleanup
o [bonding] much improved locking
o [bonding] support xmit load balancing mode
o [bonding] add rcv load balancing mode
o [netdrvr bonding] fix long failover in 802.3ad mode
o [netdrvr bonding] fix ABI version control problem
<[email protected]>:
o [IPV4]: Be more verbose about invalid ICMPs sent to broadcast
<[email protected]>:
o make sym53c8xx_2 not reject autosense IWR
<[email protected]>:
o [netdrvr typhoon] s/#if/#ifdef/ for a CONFIG_ var
Adrian Bunk <[email protected]>:
o fix .text.exit error in drivers/net/r8169.c
o add three ACPI Configure.help entries
Alan Cox <[email protected]>:
o [netdrvr tlan] fix 64-bit issues
Andi Kleen <[email protected]>:
o Remove copied inet_aton code in bond_main.c
o ACPI compile fixes for 2.4.22pre1
o Don't enable I2O for AMD64
Andrew Morton <[email protected]>:
o Additional 3c980 device support
Andy Grover <[email protected]>:
o ACPI: Fix config.in (Jeff Garzik)
o ACPI: make it so acpismp=force works (reported by Andrew Morton)
Anton Blanchard <[email protected]>:
o [netdrvr 8139cp] enable MWI via pci_set_mwi, rather than manually
Dave Engebretsen <[email protected]>:
o [netdrvr pcnet32] bug fixes
Dave Kleikamp <[email protected]>:
o Update JFS team members in jfs.txt
o JFS: resize fixes
Douglas Gilbert <[email protected]>:
o sg driver version 3.1.25
Edward Peng <[email protected]>:
o [netdrvr via-rhine] fix promisc mode
o [netdrvr sundance] bug fixes, VLAN support
o [netdrvr sundance] fix flow control bug
o [netdrvr sundance] fix another flow control bug
Hideaki Yoshifuji <[email protected]>:
o IPv6 over ARCnet (RFC2497) support, driver part
o IPv6 over ARCnet (RFC2497) support, IPv6 part
Hugh Dickins <[email protected]>:
o remove unsafe BUG() in __remove_inode_page()
Ivan Kokshaysky <[email protected]>:
o alpha: Lynx platform support (Jay Estabrook)
o alpha: initrd fix (Wiedemeier, Jeff)
o alpha: nautilus poweroff
Jay Vosburgh <[email protected]>:
o [bonding] small cleanups
o Bonding 2.4 update patch 1
o Bonding 2.4 update patch 2
o Bonding 2.4 update patch 3
o Bonding 2.4 update patch 4
o Bonding 2.4 update patch 5
o Bonding 2.4 update patch 6
Jean Tourrilhes <[email protected]>:
o irda: static init fixes
o irda: Export CRC routine to drivers
o irda: Mask C/R bit from connection
o irda-usb driver fixes
o IrCOMM chat fixes
o QoS interoperability fixes
o IrLMP timer race fix
o Fix IrIAP skb leak
o irda: Secondary nack code fixes
Jeff Garzik <[email protected]>:
o [net] store physical device a packet arrives in on
o [bonding] fix comment to prevent future merge difficulties
o [bonding] add support for getting slave's speed and duplex via ethtool
o [bonding] Moved setting slave mac addr, and open, from app to the driver
o [bonding] move driver into new drivers/net/bonding directory
o [bonding] move private decls into new drv/net/bonding/bonding.h file
o [bonding] add support for IEEE 802.3ad Dynamic link aggregation
o [netdrvr sundance] small cleanups from 2.5
o Remove duplicate CONFIG_TULIP_MWI entry in Configure.help
o [netdrvr eepro] update MODULE_AUTHOR per old-author request
o [netdrvr tlan] backport fixes and cleanups from 2.5
o [netdrvr] s/init_etherdev/alloc_etherdev/ in code comments, in 8139too and pci-skeleton drivers.
o [netdrvr 8139too] add comment, whitespace cleanup
o [netdrvr olympic] fix build with gcc 3.3
o [netdrvr r8169] use alloc_etherdev (fix race), pci_disable_device
o [netdrvr r8169] sync with 2.5 (backport whitespace cleanups)
o [netdrvr amd8111e] remove out-of-tree feature that snuck in
o [netdrvr] gcc 3.3 cleanups
o [netdrvr sis900] minor fixes from 2.5
Justin T. Gibbs <[email protected]>:
o Update the aic7xxx driver to 6.2.10 and add the aic79xx driver version 1.1.1
o Correct building of aicasm
o Update to aic7xxx version 6.2.22 and aic79xx 1.3.0_ALPHA2
o Integrate 2.5.X aic7xxx and aic79xx changes
o Misc driver updates
o Integrate changes from Christoph Hellwig <[email protected]>
o Update to aic7xxx version 6.2.24 and aic79xx version 1.3.0_ALPHA5
o Preface the "asserting atn" diagnostic with controller/target information
o aic7xxx Driver
o Aic7xxx Driver
o Aic7xxx & Aic79xx Drivers Correct 2.5.X declaration for aic_sector_div().
o Aic7XXX Firmware Assembler
o Aic7XXX and Aic79XX drivers Use down_interruptable() rather than down() to avoid having our DV threads counted toward the load average.
o Aic7XXX and Aic79XX drivers
o Aic79XX and Aic7xxx Drivers
o Aic7XXX and Aic79XX Drivers
o Aic7XXX and Aic79xx Drivers
o aic7xxx/aic79xx firmware assembler
o aic7xx and aic79xx drivers - Correct several DV issues
o aic7xxx and aic79xx driver updates
o Aic7xxx and Aic79xx DV fix
o Aic79xx Driver Update Enable abort and bus device reset handlers for both legacy and packetized connections.
o Aic7xxx Driver Update
o Aic7xxx and Aic79xx Driver Update Force an SDTR after a rejected WDTR if the syncrate is unkonwn.
o Aic7xxx Driver Update 6.2.28
o Update Aic7xxx and Aic79xx Driver Documentation
o Bump aic79xx version number to 1.3.0 now that it has passed functional testing.
o Aic7xxx Driver Update to verstion 6.2.29
o Update aic7xxx/Makefile
o Update aicasm/Makefile so that link specifications are specified after all object files. This seems to be required in order to link correctly in some cases.
o Aic79xx Driver Update to 1.3.2
o Update Aic7xxx to version 6.2.29
o AICLIB Update
o Update Aic7xxx driver [Rev 6.2.31]
o Aic79XX Driver Update [Rev 1.3.5]
o Change the callback argument for aic brace option parsing to u_long to avoid casting problems with different architectures.
o Aic7xxx Driver Update (version 6.2.32)
o Aic79xx Driver Update (version 1.3.6)
o Complete merge of AC aic7xxx and aic79xx bits
o Remove the CONFIG_AIC7XXX_ALLOW_MEMIO option. It has been supplanted by the MEMIO probe/test code.
o Aic79xx Driver Update
o Aic7xxx and Aic79xx driver Update
o Aic7xxx and Aic79xx Driver Update
o Aic7xxx and Aic79xx driver updates
o Aic7xxx and Aic79xx driver updates
o Aic7xxx and Aic79xx driver Update
o Aic7xxx and Aic79xx Driver Updates
o Aic7xxx Driver Update
o Aic79xx Driver Update
o Use absolute path to drivers/scsi in the aic7xxx Makefile
o Aic79xx Driver Update
o Aic79xx Driver Update
o Aic79xx Driver Upate
o Remove pre-2.2.X kernel support. Pre-2.2.X support requires
o Aic79xx Driver Update
o Aic7xxx and Aic79xx Driver Updates
o Update Aic79xx and Aic7xxx Documenation
o Aic79xx Driver Update (version 1.3.8)
o Aic7xxx Driver Update (6.2.33)
o Aic7xxx Driver Update
o Aic7xxx and Aic79xx Driver Updates
o Aic7xxx and Aic79xx Driver Update
o Aic7xxx and Aic79xx Driver Update
o Aic79XX Driver Update
o Aic7xxx Driver Update
o Aic7xxx Driver README update
o Aic79xx and Aic7xxx Driver Updates
o Cset exclude: [email protected]|ChangeSet|20030508192559|45150 Cset exclude: [email protected]|ChangeSet|20030507201543|47130 Cset exclude: [email protected]|ChangeSet|20030507200707|47153
o Aic7xxx and Aic79xx Updates
o Aic79xx Update
o Aic79xx Driver Update
o Aic7xxx Driver version 6.2.35
o Aic7xxx Driver Update
o Aic7xxx and Aic79xx Driver Updated
o Aic7xxx Driver Update
o Aic7xxx Driver Update
o Aic79xx Driver Update
o Aic7xxx and Aic79xx Driver Updates
o Bump aic79xx driver version to 1.3.9
o Aic7xxx Driver Update
o Aic7xxx Driver Update
o Aic79xx Driver Update
o Aic79xx Driver Update
o Aic7xxx Driver Update
o Aic7xxx and Aic79xx Driver Update
o Aic7xxx and Aic79xx driver Update
o Aic7xxx Driver Update
o Aic7xxx Driver Update
o Aic79xx Driver Update
o Aic7xxx and Aic79xx Driver Update
o Aic7xxx Driver Update
o Aic79xx Driver Update
o Update Aic79xx Readme
Kai Germaschewski <[email protected]>:
o ISDN: Fix Fritz!PCI v2 xmit irq underrun recovery
o ISDN: Fix bug in ST5481 D-Channel state machine
Karsten Keil <[email protected]>:
o ISDN: [PATCH] Fix problem with external hisax drivers
Maksim Krasnyanskiy <[email protected]>:
o L2CAP config req/rsp handling fixes
o [Bluetooth] Detect and log error condition when first L2CAP fragment is too long
o [Bluetooth] RFCOMM must wait for MSC exchange to complete before sending the data
o [Bluetooth] L2CAP sockets can now set LM_RELIABLE flag and get notification when we detect reliablity problem with the ACL connection.
o [Bluetooth] Add support for SO_LINGER option to all Bluetooth protocols
o Bluetooth: RFCOMM must send MSC when DLC was opened by SABM
o [Bluetooth] Fix RFCOMM C/R and Direction bit handling
o [Bluetooth] L2CAP qualification spec mandates sending additional config request if we receive config response with unacceptable parameters error code.
Marcel Holtmann <[email protected]>:
o [Bluetooth] Send the correct values in RPN response
o [Bluetooth] Handle priority bits in parameter negotiation
o [Bluetooth] Implement rfcomm_tty_put_char() function
o [Bluetooth] Send correct RPN response for accepted values
o [Bluetooth] Set EA bit for V.24 signals parameter
o [Bluetooth] Handle bit rate in remote port negotiation
o [Bluetooth] Quirk for devices with no ISOC endpoints
Marcelo Tosatti <[email protected]>:
o Changed EXTRAVERSION to -pre2
o Cset exclude: [email protected]|ChangeSet|20030620200318|50799
Mark A. Greer <[email protected]>:
o PPC32: Fix the gen550 infrastructure for baud rates other than 9600
Olaf Hering <[email protected]>:
o remove TIOCGDEV from asm/ioctls.h
o RAID_AUTORUN is a compatible ioctl
Patrick McHardy <[email protected]>:
o ISDN: [PATCH] missing cli() in isdn_net.c
o ISDN: [PATCH] don't unlock lp if there is nothing to unlock
o ISDN: Add CONFIG_IPPP_FILTER
o [NETFILTER]: Dont call helpers expectfn() for unconfirmed connections
Paul Mackerras <[email protected]>:
o PPC32: Update for PPC 4xx TLB and exception handling
o PPC32: Add a new framework for on-chip peripherals for the IBM 4xx embedded processors.
o PPC32: Introduce a new config symbol, CONFIG_40x, used for PPC 40x cpus
o PPC32: Add generic IBM PPC405GP support and use it on the walnut platform
o PPC32: Update the support for the "Walnut" 405GP platform
o PPC32: Make debug exceptions usable on 4xx-class processors, and improve trap handling.
o PPC32: Add support for PPC 405GP interrupt controller
o PPC32: Extra register and other definitions for the PPC 405GP processor
o PPC32: Move PC-style serial port definitions out to asm/pc_serial.h
o PPC32: remove ppc4xx_serial.h, it is no longer used
o PPC32: Cleanups for PPC 405GP-based systems; add file of OCP ids
o PPC32: Don't run `checks' program on make zImage
o PPC32: Add definitions for the UIC interrupt controller on the 405GP processor
o PPC32: Add support for PCI and time-of-day clock on 405GP-based systems
o PPC32: Allow for PCI host bridges that need explicit type 1 cycle indication
Randy Dunlap <[email protected]>:
o unexpected IO-APIC code update
Rusty Russell <[email protected]>:
o [irda] module refcounts for irlan
o [patch, 2.5] dgrs doesn't free on error path
o namespace pollution in cosa driver
o [2.4 patch] fix wavelan_cs compile warning
o Clear up GFP confusion in rcpci45.c
o [patch, 2.5] fix errorpath in apne.c
o Remove naked GFP_DMA from drivers_net_macmace.c
o namespace pollution in skfddi driver
o improve signal-to-noise ratio in atm code
o 2.4.20 wait.h doc typo
o fs_autofs4_root.c unused variable
o [TRIVIAL PATCH 2.4] update README file to current
o fix documentation in include_asm-i386_bitops.h
o missing headers in i82092.c
o fix linewrap in Documentation_power_pci.txt
o include_asm-ia64_sal.h, typo: the the
o Typos in drivers_s390_net_iucv.h
o [TRIVIAL PATCH] include_asm-i386_dma.h: wrong lowest DMA
o redundant declarations (#1_15)
o add some missing init.h inclusions
o remove superflous if in wait_kio
o Squash warning in ppc64 addnote tool
o fix linewrap in Documentation_filesystems_sysv-fs.txt
o set b_page to null in fake buffer_head for O_DIRECT
o fix linewrap in Documentation_pci.txt
o misc_register audit fix of wdt_pci
o misc register fix on ds1286
o reorganize for unreachable code
Sam Ravnborg <[email protected]>:
o [netdrvr sis900] make function headers readable by kernel-doc tool
Scott Feldman <[email protected]>:
o [netdrvr e1000] add support for NAPI
o [netdrvr e1000] add TSO support -- disabled
o 10GbE ethtool support
o remove ethtool privileged references
o [e100] Remove "Freeing alive device" warning
o [e100] move e100_asf_enable under CONFIG_PM to avoid warning
o [e100] Add ethtool parameter support
o [e100] Add ethtool cable diag test
o [e100] Add MDI/MDI-X status to ethtool reg dump
o [e100] cleanup Tx resources before running ethtool diags
o [e100] full stop/start on ethtool set speed/duplex/autoneg
o [e100] fixed stalled stats collection
o [e100] VLAN configuration was lost after ethtool diags run
o [e100] use skb_headlen() rather than rolling own
o [e100] set netdev members before registration
o [e100] misc
Tom Rini <[email protected]>:
o PPC32: Clean up the cpu_idle() code a bit
o PPC32: Fix a multicast bug in the MPC 8xx / 8260 enet drivers
o PPC32: Correct the DTLB miss handler on MPC8xx
o PPC32: Fix a problem with MDIO requests on reset in MPC 8xx enet
o PPC32: Minor cleanups to the MPC 8xx FEC driver
o PPC32: Fix a small problem in the 8xx / 8260 uart code
o PPC32: Important fixes in the MPC8xx FEC and MPC826x enet driver
o PPC32: Describe when we want to do a CPM reset on MPC8xx
o Add /proc/sys/kernel/l3cr
Zwane Mwaikambo <[email protected]>:
o Remove warning due to comparison in drivers/net/pcnet32.c
On 06.27, Marcelo Tosatti wrote:
>
> Hello,
>
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
>
[...]
> Marcelo Tosatti <[email protected]>:
> o Changed EXTRAVERSION to -pre2
> o Cset exclude: [email protected]|ChangeSet|20030620200318|50799
>
I suppose it is the db4 change. Why did you remove it ?
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.21-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-2mdk))
On 06.27, Marcelo Tosatti wrote:
>
> Hello,
>
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
>
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
>
This fixes inlining (really, not-inlining) with gcc3. How about next -pre ?
--- 25/include/linux/compiler.h~gcc3-inline-fix 2003-03-06 03:02:43.000000000 -0800
+++ 25-akpm/include/linux/compiler.h 2003-03-06 03:11:42.000000000 -0800
@@ -1,6 +1,13 @@
#ifndef __LINUX_COMPILER_H
#define __LINUX_COMPILER_H
+#if __GNUC__ >= 3
+#define inline __inline__ __attribute__((always_inline))
+#define inline__ __inline__ __attribute__((always_inline))
+#define __inline __inline__ __attribute__((always_inline))
+#define __inline__ __inline__ __attribute__((always_inline))
+#endif
+
/* Somewhere in the middle of the GCC 2.96 development cycle, we implemented
a mechanism by which the user can annotate likely branch directions and
expect the blocks to be reordered appropriately. Define __builtin_expect
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.21-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-2mdk))
On 06.27, Marcelo Tosatti wrote:
>
> Hello,
>
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
>
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
>
I think this is useful and safe. Plz, can you apply for next pre ?
--- linux-2.4.21-bp1/arch/i386/Makefile.orig 2003-06-18 23:40:25.000000000 +0200
+++ linux-2.4.21-bp1/arch/i386/Makefile 2003-06-18 23:59:25.000000000 +0200
@@ -45,7 +45,7 @@
endif
ifdef CONFIG_M586MMX
-CFLAGS += -march=i586
+CFLAGS += $(call check_gcc,-march=pentium-mmx,-march=i586)
endif
ifdef CONFIG_M686
@@ -53,11 +53,11 @@
endif
ifdef CONFIG_MPENTIUMIII
-CFLAGS += -march=i686
+CFLAGS += $(call check_gcc,-march=pentium3,-march=i686)
endif
ifdef CONFIG_MPENTIUM4
-CFLAGS += -march=i686
+CFLAGS += $(call check_gcc,-march=pentium4,-march=i686)
endif
ifdef CONFIG_MK6
@@ -74,11 +74,11 @@
endif
ifdef CONFIG_MWINCHIPC6
-CFLAGS += -march=i586
+CFLAGS += $(call check_gcc,-march=winchip-c6,-march=i586)
endif
ifdef CONFIG_MWINCHIP2
-CFLAGS += -march=i586
+CFLAGS += $(call check_gcc,-march=winchip2,-march=i586)
endif
ifdef CONFIG_MWINCHIP3D
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.21-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-2mdk))
On 06.27, Marcelo Tosatti wrote:
>
> Hello,
>
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
>
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
>
Alredy declared in kernel.h.
--- linux/include/asm-i386/spinlock.h.orig 2002-10-15 10:12:25.000000000 +0100
+++ linux/include/asm-i386/spinlock.h 2002-10-15 10:12:35.000000000 +0100
@@ -6,9 +6,6 @@
#include <asm/page.h>
#include <linux/config.h>
-extern int printk(const char * fmt, ...)
- __attribute__ ((format (printf, 1, 2)));
-
/* It seems that people are forgetting to
* initialize their spinlocks properly, tsk tsk.
* Remember to turn this off in 2.4. -ben
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.21-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-2mdk))
On 06.27, Marcelo Tosatti wrote:
>
> Hello,
>
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
>
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
>
--- linux-2.4.20.orig/include/linux/pci_ids.h 2003-05-30 00:43:12.000000000=
+0200
+++ linux-2.4.20/include/linux/pci_ids.h 2003-05-30 00:44:37.000000000 +020=
0
@@ -1029,7 +1029,6 @@
#define PCI_DEVICE_ID_VIA_8233A 0x3147
#define PCI_DEVICE_ID_VIA_P4X333 0x3168
#define PCI_DEVICE_ID_VIA_8235 0x3177
-#define PCI_DEVICE_ID_VIA_8377_0 0x3189
#define PCI_DEVICE_ID_VIA_8377_0 0x3189
#define PCI_DEVICE_ID_VIA_8237 0x3227
#define PCI_DEVICE_ID_VIA_86C100A 0x6100
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.21-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-2mdk))
[CCs slightly trimmed]
Quoth J.A. Magallon:
[snippety]
> This fixes inlining (really, not-inlining) with gcc3. How about next -pre ?
>
> --- 25/include/linux/compiler.h~gcc3-inline-fix 2003-03-06 03:02:43.000000000 -0800
> +++ 25-akpm/include/linux/compiler.h 2003-03-06 03:11:42.000000000 -0800
> @@ -1,6 +1,13 @@
> #ifndef __LINUX_COMPILER_H
> #define __LINUX_COMPILER_H
>
> +#if __GNUC__ >= 3
> +#define inline __inline__ __attribute__((always_inline))
> +#define inline__ __inline__ __attribute__((always_inline))
> +#define __inline __inline__ __attribute__((always_inline))
> +#define __inline__ __inline__ __attribute__((always_inline))
> +#endif
> +
> /* Somewhere in the middle of the GCC 2.96 development cycle, we implemented
> a mechanism by which the user can annotate likely branch directions and
> expect the blocks to be reordered appropriately. Define __builtin_expect
I'm willing to give this a shot. Where, or perhaps what, should I test
to evaluate best the effects, if any, of this patch?
Kurt
--
Decision maker, n.:
The person in your office who was unable to form a task force
before the music stopped.
On Friday 27 June 2003 00:03, Marcelo Tosatti wrote:
Hi Marcelo,
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
so why you don't merge this? This is now the 4th resend.
--------------------------------------------------------------------------
[PATCH 2.4.22-BK] [RESEND 3rd] Fix oom killer braindamage
Date: 21.06.2003 22:04
From: Marc-Christian Petersen <[email protected]> (Working Overloaded
Linux Kernel)
To: Marcelo Tosatti <[email protected]>
CC: [email protected]
Hi Marcelo,
attached patch fixes the oom killer braindamage where it tries to kill
processes again and again and again w/o any ending or successfull killing of
the selected processes in an OOM case.
The attached, very simple but effective, patch fixes it.
All the kudos go to Rik van Riel.
Patch tested and works, and also for a long time in my tree (and maybe also
others?!)
This issue is out there for several years.
Please apply it for 2.4.22-pre2, thanks.
--------------------------------------------------------------------------
ciao, Marc
On Thu, 26 Jun 2003, Marcelo Tosatti wrote:
>
> Hello,
>
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
>
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
>
> Have fun
>
>
> Summary of changes from v2.4.22-pre1 to v2.4.22-pre2
> ============================================
Please pull the new "shortlog" script from
bk://kernel.bkbits.net/torvalds/tools/
On Gwe, 2003-06-27 at 00:08, J.A. Magallon wrote:
> This fixes inlining (really, not-inlining) with gcc3. How about next -pre ?
Benchmark that before you blindly assume its right. Gcc not inlining large
stuff actually appears to be _smarter_ than the authors of the code
On Fri, Jun 27, 2003 at 01:31:17AM +0200, J.A. Magallon wrote:
>
> On 06.27, Marcelo Tosatti wrote:
> >
> > Hello,
> >
> > Here goes -pre2 with a big number of changes, including the new aic7xxx
> > driver.
> >
> > I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> > short.
> >
>
> Alredy declared in kernel.h.
>...
It seems this issue is also present in 2.5.73.
Could you send such cleanup changes for inclusion into 2.5 first?
TIA
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
On 06.27, Alan Cox wrote:
> On Gwe, 2003-06-27 at 00:08, J.A. Magallon wrote:
> > This fixes inlining (really, not-inlining) with gcc3. How about next -pre ?
>
> Benchmark that before you blindly assume its right. Gcc not inlining large
> stuff actually appears to be _smarter_ than the authors of the code
>
Let's be clear, I just collect patches that I think are interesting. The
original post by Andrew Morton <[email protected]> is here:
http://marc.theaimsgroup.com/?l=linux-kernel&m=104695003008386&w=2
Quoting it:
shrinks my 3.2.1-compiled kernel text by about 64 kbytes:
text data bss dec hex filename
3316138 574844 726816 4617798 467646 vmlinux-before
3249255 555436 727204 4531895 4526b7 vmlinux-after
mnm:/tmp> nm vmlinux-before|grep __constant_c_and_count_memset | wc
233 699 9553
mnm:/tmp> nm vmlinux-after|grep __constant_c_and_count_memset | wc
13 39 533
And I also remember other posts, by other author of a similar patch, that
claimed some critical functions not being inlined due to size (memcpys..).
Will try to look original info about that...
...here it is:
http://marc.theaimsgroup.com/?l=linux-kernel&m=103632312702693&w=2
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.21-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-2mdk))
Hi Marcelo,
the patch below contains four trivial fixes for compile failures with
gcc 3.3 stolen from -ac.
I've tested the compilation with 2.4.22-pre2.
diffstat output:
drivers/net/irda/ma600.c | 6 +++---
drivers/net/wan/sdla_chdlc.c | 3 +--
drivers/sound/cs46xx.c | 4 ++--
net/decnet/dn_table.c | 3 +--
4 files changed, 7 insertions(+), 9 deletions(-)
Please apply
Adrian
--- linux.vanilla/drivers/net/irda/ma600.c 2002-11-29 21:27:18.000000000 +0000
+++ linux.21-ac3/drivers/net/irda/ma600.c 2003-05-29 01:40:07.000000000 +0100
@@ -51,9 +51,9 @@
#undef ASSERT(expr, func)
#define ASSERT(expr, func) \
if(!(expr)) { \
- printk( "Assertion failed! %s,%s,%s,line=%d\n",\
- #expr,__FILE__,__FUNCTION__,__LINE__); \
- ##func}
+ printk( "Assertion failed! %s,%s,%s,line=%d\n",\
+ #expr,__FILE__,__FUNCTION__,__LINE__); \
+ func}
#endif
/* convert hex value to ascii hex */
--- linux.vanilla/drivers/net/wan/sdla_chdlc.c 2002-11-29 21:27:18.000000000 +0000
+++ linux.21-ac3/drivers/net/wan/sdla_chdlc.c 2003-05-28 15:35:56.000000000 +0100
@@ -591,8 +591,7 @@
if (chdlc_set_intr_mode(card, APP_INT_ON_TIMER)){
- printk (KERN_INFO "%s:
- Failed to set interrupt triggers!\n",
+ printk (KERN_INFO "%s: Failed to set interrupt triggers!\n",
card->devname);
return -EIO;
}
--- linux.vanilla/drivers/sound/cs46xx.c 2003-06-14 00:11:37.000000000 +0100
+++ linux.21-ac3/drivers/sound/cs46xx.c 2003-06-22 13:36:11.000000000 +0100
@@ -947,8 +950,8 @@
struct InitStruct
{
- u32 long off;
- u32 long val;
+ u32 off;
+ u32 val;
} InitArray[] = { {0x00000040, 0x3fc0000f},
{0x0000004c, 0x04800000},
--- linux.vanilla/net/decnet/dn_table.c 2001-12-21 17:42:05.000000000 +0000
+++ linux.21-ac3/net/decnet/dn_table.c 2003-05-28 15:37:27.000000000 +0100
@@ -836,8 +836,7 @@
return NULL;
if (in_interrupt() && net_ratelimit()) {
- printk(KERN_DEBUG "DECnet: BUG! Attempt to create routing table
-from interrupt\n");
+ printk(KERN_DEBUG "DECnet: BUG! Attempt to create routing table from interrupt\n");
return NULL;
}
if ((t = kmalloc(sizeof(struct dn_fib_table), GFP_KERNEL)) == NULL)
Marcelo Tosatti wrote:
> Hello,
>
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
Fixes my can't-unblank-screen problem reported with 2.4.22-pre1.
Thanks,
--alessandro
"I'm always trying to find some kind of honest emotion
and connection with the most naked human emotions.
It's a dangerous, dark path but also very exciting."
(Steve Wynn, interviewed by 'Mescalina', 2003)
On 06.27, Marcelo Tosatti wrote:
>
> Hello,
>
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
>
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
>
Correct me if I'm wrong. I found this just by chance:
mm/filemap.c:
ssize_t
do_generic_direct_write(struct file *file,const char *buf,size_t count, loff_t *ppos)
...
if (!file->f_flags & O_DIRECT)
BUG();
This fails to trigger the BUG() just when it should. This should be:
if (!(file->f_flags & O_DIRECT))
BUG();
To check my C, I tried this:
#include <stdio.h>
#define O_D 2
int main()
{
int flags;
flags = O_D +4;
printf("%d\n",!flags & O_D);
printf("%d\n",!(flags & O_D));
flags = 1+4;
printf("%d\n",!flags & O_D);
printf("%d\n",!(flags & O_D));
return 0;
}
Output:
werewolf:~# ./kk
0
0
0
1
So it gives 0 when 'flags' != 0, not regarding if O_D is there or not....
Correct ?
One other job for the Stanford checker... ?
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.21-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-2mdk))
Hi,
Marcelo Tosatti <[email protected]> writes:
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
>
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
What's wrong with the generic HDLC update then? Are you going to apply it?
ftp://ftp.pm.waw.pl/pub/linux/hdlc/hdlc-2.4.21pre7-1.14.patch
or http://ftp.pm.waw.pl/pub/linux/hdlc/hdlc-2.4.21pre7-1.14.patch
Yes, it applies to 2.4.21-pre7 and later kernels, including 2.4.22-pre2.
I hope it will require "-R" to apply it to pre3...
TIA.
--
Krzysztof Halasa
Network Administrator
Hi Marcello,
please can we have generic hdlc in 2.4.22. I was very disappointed
that it didn't get into 2.4.21
Kevin Curtis
Linux Development
FarSite Communications Ltd
[email protected]
tel: +44 1256 330461
fax: +44 1256 854931
http://www.farsite.co.uk
-----Original Message-----
From: Marc-Christian Petersen [mailto:[email protected]]
Sent: 27 June 2003 06:59
To: Marcelo Tosatti; lkml
Subject: Re: Linux 2.4.22-pre2
On Friday 27 June 2003 00:03, Marcelo Tosatti wrote:
Hi Marcelo,
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
so why you don't merge this? This is now the 4th resend.
--------------------------------------------------------------------------
[PATCH 2.4.22-BK] [RESEND 3rd] Fix oom killer braindamage
Date: 21.06.2003 22:04
From: Marc-Christian Petersen <[email protected]> (Working Overloaded
Linux Kernel)
To: Marcelo Tosatti <[email protected]>
CC: [email protected]
Hi Marcelo,
attached patch fixes the oom killer braindamage where it tries to kill
processes again and again and again w/o any ending or successfull killing of
the selected processes in an OOM case.
The attached, very simple but effective, patch fixes it.
All the kudos go to Rik van Riel.
Patch tested and works, and also for a long time in my tree (and maybe also
others?!)
This issue is out there for several years.
Please apply it for 2.4.22-pre2, thanks.
--------------------------------------------------------------------------
ciao, Marc
Marcello,
Yes please do that. There are people who want it in.
Kevin Curtis
Linux Development
FarSite Communications Ltd
[email protected]
tel: +44 1256 330461
fax: +44 1256 854931
http://www.farsite.co.uk
-----Original Message-----
From: Krzysztof Halasa [mailto:[email protected]]
Sent: 28 June 2003 00:39
To: Marcelo Tosatti
Cc: lkml
Subject: Re: Linux 2.4.22-pre2
Hi,
Marcelo Tosatti <[email protected]> writes:
> Here goes -pre2 with a big number of changes, including the new aic7xxx
> driver.
>
> I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> short.
What's wrong with the generic HDLC update then? Are you going to apply it?
ftp://ftp.pm.waw.pl/pub/linux/hdlc/hdlc-2.4.21pre7-1.14.patch
or http://ftp.pm.waw.pl/pub/linux/hdlc/hdlc-2.4.21pre7-1.14.patch
Yes, it applies to 2.4.21-pre7 and later kernels, including 2.4.22-pre2.
I hope it will require "-R" to apply it to pre3...
TIA.
--
Krzysztof Halasa
Network Administrator
On 06.28, J.A. Magallon wrote:
>
> On 06.27, Marcelo Tosatti wrote:
> >
> > Hello,
> >
> > Here goes -pre2 with a big number of changes, including the new aic7xxx
> > driver.
> >
> > I wont accept any big changes after -pre4: I want 2.4.22 timecycle to be
> > short.
> >
>
> Correct me if I'm wrong. I found this just by chance:
>
> mm/filemap.c:
> ssize_t
> do_generic_direct_write(struct file *file,const char *buf,size_t count, loff_t *ppos)
> ...
>
> if (!file->f_flags & O_DIRECT)
> BUG();
>
> This fails to trigger the BUG() just when it should. This should be:
>
> if (!(file->f_flags & O_DIRECT))
> BUG();
Someone can confirm this / flame me , plz ?
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.22-pre2-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-2mdk))