** I strongly recommend saying N to IDE TCQ options otherwise this
should hopefully build and run happily.
Doug's scsi changes broke mptfusion. I've not looked into that yet
also u14f/u34f, and the host changes broke all 5380 based devices
This one builds, its not yet had any measurable testing
Linux 2.5.44-ac1
- Resync with Linus 2.5.43/44
o Fix net/ipv4/raw.c build problem (me)
o Fix bluetooth pcmcia builds (me)
o Fix dm includes (me)
| I've not merged any of the DM updates yet
o Fix 3c515, fealnx printk type warnings (me)
o Fix multi-line string literal in olympic driver (me)
o Fix printk type warnings in tulip (me)
o Document core naming sysctl (Randy Dunlap)
o Fix hd_struct size/offset bugs (Mark Lord)
o Further sym53c416 updates (Bjoern Zeeb)
o Fix ramdisk cache flush (Paul Mundt)
o Fix pnp config.in for make Xconfig (Roman Zippel)
o Correct ncpfs marking of executables (Petr Vandrovec)
o Small matroxfb fixes (Petr Vandrovec)
o Small cleanups for i2o_block so Al can clean (Al Viro)
up the core block code
o Fix hang at shutdown with offlined disk (Mike Anderson)
o Fix error reporting on scsi offline (Mike Anderson)
o Fix hang on power off with scsi (Mike Anderson)
o Fix typo in pnp.h (Martin Dahl)
o Remove tqueue.h from cycx_main (Adrian Bunk)
o Fix vlsi irda compile (Adrian Bunk)
o Fix hamradio makefile breakage (Adrian Bunk)
o Fix inia100 build (John Fort)
o Fix AX.25 build for ip_proc (Dave Miller)
o Fix aic7xxx Makefile (Inaky Perez-Gonzalez)
o Fix vga16fb (Ben Pfaff)
o Optimise spinlock to Intel recommendation (Manfred Spraul)
o Fix pipe wakeup bug (Manfred Spraul)
o Fix semop 32bit pid handling (Manfred Spraul)
o Fix qlogic1280 build (Jens Axboe)
o Merge BeOS fs (already in 2.4) (Will Dyson,
Sergey Kostyliov)
o Clean up wan ioctl structures (Krzysztof Halasa)
o Some trident audio takes a long time to (Kenneth Sumrall)
come up (Hitachi webpad)
o Add DVB api and core (Holger Waechtler)
o Add one DVB driver so people can see how it (Holger Waechtler)
all fits together
| This wants further review. There are known things to do yet
| but its important to get the stuff in since Digital TV is
| becoming the norm in western europe.
o Print something clueful if menuconfig explodes (Russell King)
o Move BUG() into asm/bug.h (Russell King)
o Report errors unzipping ramdisks (Russell King)
o Support extra weird numeric key on ARM boxes (Russell King)
o Fix missing devexit_p in tulip (Andrey Panin)
o Kill sr_host (Patrick Mansfield)
o S/390 Makefile and Config updates (Martin Scwidefsky)
o S/390 user access fixes (Martin Scwidefsky)
o 31bit emulation fixes for S/390 (Martin Scwidefsky)
o Make S/390 possible cpu map volatile (Martin Scwidefsky)
o Update dasd drivers for S/390 series (Martin Scwidefsky)
o Update ver_linux (Steven Cole)
o Fix blk ioctls on aacraid (Mark Haverkamp)
o Fix SiS IDE build without procfs (Lionel Bouton)
o i386 verify write fixes (Brian Gerst)
o iphase ATM updates (Francois Romieu)
o Update i810-tco to C99 initializers (Wim Van Sebroeck)
o IDE updates for ARM platform (Russell King)
o Fix megaraid build (Mike Anderson)
| This may change the device order for some folks but it works
| at least
o Fix in2000 to handle scsi host list change (me)
o Fix ncr53c8xx build (me)
o Fix atp870u build (me)
o Fix nsp32 build (me)
o Fix firewire prototypes (me)
Linux 2.5.42-ac1
Merge with Linus 2.5.42
o Merge the LVM2 device mapper (Joe Thornber)
- Drop uid16 S/390 bits pending resolution (me)
* Fix iphase build (Adrian Bunk)
* Fix UML build (Kai Germaschewski)
* Fix cpufreq compile (Adrian Bunk)
o Move dead verify_area code from sh port (Brian Gerst)
* Fix missing AIO symbols (Ben LaHaise)
o Fix ATM makefile (Sam Ravnborg)
* Fix esp build (Andres Salmon)
* Fix cifs/jfs symbol name collision (Steve F)
* Update CPIA to match 2.4 tree (Duncan Haldane)
* Fix cifs 64bit and cifs scsi name collision (Steve F)
* Fix a compile of missing sysrq updates (James Simmons)
* Fix sparc timer build (Pete Zaitcev)
* Fix comile without networking (Miles Bader)
* Remove some left over _ret functions (SL Baur)
* Update syncppp code (Paul Fulghum)
* Fix n_hdlc leak (Paul Fulghum)
* Make synclink_cs build again (Paul Fulghum)
* Make synclinkmp build again (Paul Fulghum)
* Make synclink build again (Paul Fulghum)
o Fix NFS symbols for NFS as a module (Olaf Dietsche)
* Fix problem with scsidriver docbook (Joaquim Fellmann)
* Kill dead suspend code in IDE (Pavel Machek)
* Kill unreferenced workqueue define (Pavel Machek)
* Fix swsuspend with discontiguous memory bits (Pavel Machek)
* Fix cpqfc ioctl sense buffer handling (Francis Wiran)
* Sym53c416 from cli to real locking (Bjoern Zeeb)
* Fix a case where sd uses freed memory (Patrick Mansfield)
o Fix p4-clockmod on HT processors (Dominic Brodowski)
o CPUfreq interface update (Dominic Brodowski)
* Fix eicon build (me)
* Restore disconnect field in devices for (me)
driver use
Linux 2.5.41-ac2
* Fix jffs/jffs2 properly this time (bpbb) (me)
* Fix jffs2 for workqueues (me)
* Next set of i2o_scsi update work (me)
* Do the 2.5 checkup pass on the 3c501 driver (me)
o Add missing exports for file system modules (Nikita Danilov)
on UML
* Fix ipx proc permission bogosity (Arnaldo Carvalho de Melo)
* Switch appletalk spinlocks to rwlocks (Arnaldo Carvalho de Melo)
* Correct sys_getpid docs (Robert Love)
* SubmittingPatches indent fix (John Levon)
* cciss, cpqarray. rd. hd fixes (Al Viro)
* Fix cpia with gcc 3.2 (Randy Dunlap)
* Use C99 structure initializers in IDE (Art Haas)
* Use C99 structure initializers in HFS (Art Haas)
* Update DMI scanner (Jean Delvare)
* Fix bogus types in ide-cd.h (Skip Ford)
* ns83820 updates (Ben LaHaise)
* AIO updates (Ben LaHaise)
* Beeping and sysrq on m68k (Vojtech Pavlik)
* Improve hid naming (Vojtech Pavlik)
* LSM docs (Greg Kroah-Hartmann)
* Merge UML updates (Jeff Dike)
* Final superblock union cleanup (Brian Gerst)
- Fix atm build/makefile breakage (Adrian Bunk)
* Brlock optimisation (Robert Love)
* Miscellaneous USB updates (Greg Kroah-Hartmann)
* MPT Fusion update (Pam Delaney)
- Back out sched.c change - seem,s to cause hangs (me)
* Serial compile fix (Russell King)
* S/390 compile fixes (Martin Schwidefsky)
* S/390 workqueue updates (Martin Schwidefsky)
* Switch 3215/3270 from work queue to tasklet (Martin Schwidefsky)
* Update S/390 link scripts (Martin Schwidefsky)
* Remove duplicate S/390 memset (Martin Schwidefsky)
* Fix S/390 syscall tracing (Martin Schwidefsky)
* Multiple 3270 fixes (Martin Schwidefsky)
* Configurable core names (Jes Rahbek Klinke)
X Clean up s/390x 16bit uid calls (Martin Schwidefsky)
* Fix EH locking on NCR5380 (me)
| Should now work on SMP boxes (badly admittedly)
* Indent wd7000 (no code changes) (me)
* First pass at the in2000 scsi driver (me)
| New locking, new_eh, address conversion
Linux 2.5.41-ac1
- Merge with Linus 2.5.41
- Drop S/390 drivers subtree for Linus
- Drop task queue fixes for schedule_work
- TODO: merge two sets of conflicting UML changes
- TODO: double check bluetooth merge
* Fix aacraid makefile (Mark Haverkamp)
* Fix ips compile (Paul Larson)
* Fix aha152x compile (Michel Eyckmans)
* Fix orinoco_cs compile (Wichert Akkerman, Martin Waitz)
* Fix i2o_core compiler (Gregoire Favre)
* Fix missing exports for netfilter
* Fix compile failure in jffs (me)
* Fix compile failure in jffs2 (me)
* Fix Divas_Mod compile (me)
* Fix hisax compile (me)
* Fix ipacx compile (me)
* Fix pcbit compile (me)
* Fix tpam compile (me)
* Fix i2o_lan build (me)
* Fix i2o_proc build (me)
* Fix ppa compile (me)
* Fix imm compile (me)
* Fix ipv6 compile (me)
Linux 2.5.40-ac6
* Cadet_wake can be static (me)
* Bluetooth configuration cleanups (Marcel Holtmann)
o Hardwired empty bar handling fix take two (Ivan Kokshaysky)
* Use kernel crc32 lib for bluetooth (Marcel Holtmann)
* Make scsi cdrom honour passed timeouts (Peter Osterlund)
* Make aironet4500_cs compile (me)
* Fix bugs where ibmtr unmapped the wrong address (me)
* Fix crash problem in oss dmabuf.c (me)
| Its still very broken but ALSA should replace it
* Fix opl3sa2 warnings (me)
* Make tcic compile again (me)
* Make i82365 also use del_timer_sync (me)
* Fix warnings in fpu emulator (me)
* Fix t128 for NCR5380 changes (me)
* Fix pas16 for NCR5380 changes (me)
* Fix dmx3191 for NCR538 changes (me)
* First pass seagate st02 cleanups (me)
* Clean up de600 driver. Switch to spinlocks (me)
remove crud, formatting junk etc
| Still needs rewriting to use parport
o Remove extra unlock in wd7000 (Matthew Wilcox)
* First basic pass at qlogicgas (me)
* Clean up the fdomain isa scsi (me)
* Clean up max_thread setting limits (Matthew Wilcox)
* Ricoh cardbus performance fix (KOMURO)
* Switch appletalk to seq_file /proc (Arnaldo Carvalho de Melo)
* Switch X.25 to seq_file (Arnaldo Carvalho de Melo)
* Fix bugs in the above (Arnaldo Carvalho de Melo)
Linux 2.5.40-ac5
* Rework S/390 driver init sequences (Martin Schwidefsky)
* Swap immediate_bh for tasklets for s/390 3215 (Martin Schwidefsky)
* UML updates - crash fixes, driver cleanup (Jeff Dike)
pcap transport
* Switch fmi radio card to sleeping waits (me)
* Fixing missing printk \n in fmi radio (me)
o Update to newer uclinux patch (Greg Ungerer)
| Unresolved now:
| fs/exec.c kernel/fork.c procfs sysctl
| can nommu be folded in (Hch)
* Remove surplus irq_disable from mpt fusion (Carlos Gorges)
* Export gdt for APM (Carlos Gorges)
| Marked as _GPL because its deep internals stuff
* Merge the add/put disk gendisk changes for i2o (Al Viro)
* Switch NCR5380/g_NCR5380 to new_eh (me)
* Fix cs89x0 netdevice init as module (me)
* Change some of the wd7000 code to use
udelay and do other cleanups
* Switch wd7000 to new_eh (me)
* Serial driver updates (Russell King)
* Sync bluetooth with 2.4, fix SMP, hotplug (Maksim Krasnyanskiy)
support L2CAP, BNEP, HCI filter etc
* Move firmwareloading to hotplug for bluetooth (Maksim Krasnyanskiy)
* Pull hpfs out of shared struct superblock (Brian Gerst)
X Fix sleep with pre-empt disabled in (Manfred Spraul)
set_cpus_allowed
Linux 2.5.40-ac4
* Make ibm partition code compile again (Martin Schwidefsky)
* Remove unneeded config options on S/390 (Martin Schwidefsky)
* Update DASD drivers (Martin Schwidefsky)
* Update S/390 xpram driver (Martin Schwidefsky)
* Replace S/390 BH code by tasklets (Martin Schwidefsky)
* Fix S/390 bitops bugs (Martin Schwidefsky)
* S/390x 31bit emulation fixes (Martin Schwidefsky)
* Update S/390 link scripts (Martin Schwidefsky)
* Add S/390 pre-empt support (Martin Schwidefsky)
* Inline some S/390 old compilers couldnt handle (Martin Schwidefsky)
* Use diag 44 for S/390x spinlocks (Martin Schwidefsky)
* Better S/390 timer handling (Martin Schwidefsky)
* S/390 code cleanups (Martin Schwidefsky)
* Clean up S/390 fpu load/stores (Martin Schwidefsky)
* DECnet updates for testing (Steve Whitehouse)
* Add console shutdown handling to S/390 (Martin Schwidefsky)
* Remove some bogus S/390 sanity checks (Martin Schwidefsky)
* Clean up S/390 process irq (Martin Schwidefsky)
* Fix/simplify chpids handling on S/390 (Martin Schwidefsky)
* No /proc/interrupts on S/390 (Martin Schwidefsky)
* Remove now unneeded S/390 hack in init/main.c (Martin Schwidefsky)
* Clean up all the S/390 ptrace handling (Martin Schwidefsky)
o Fix build with local apic enabled (James Bottomley)
* Initial i2o_block merge of 2.4/2.5 code (me)
| Not yet functional
* Initial i2o_scsi merge of 2.4/2.5 code (me)
| Needs dma mapping, 64bit, be and new_eh
- Revert Ivan's pci change (breaks serverworks)
* PCI serial oops fix (William Irwin)
* Remove dead wood from unistd.h (Brian Gerst)
o Fix bug in capget (Chris Wright)
* Switch qnxfs to new style initializers (Art Haas)
o Recongize qnx v6 file systems (Anders Larsen)
* Kill off remaining pcibios_ users (Greg "Ninja Turtle" Kroah-Hartmann)
* Fix scsi debug for scsi scan changes (Mike Anderson)
* Fix some bugs in scsi error handling (Mike Andersen)
* Forward port RMK's 2.4 scsi fixes (Mike Andersen)
* Allow longer settle times for scsi reset (Mike Andersen)
* Hopefully improve error policies a bit (Mike Andersen)
Linux 2.5.40-ac3
* Resync telephony drivers with 2.4 (me)
| Forward port security and other minor fixes
* Fix aironet4500 build for tq changes (me)
* Fix keyspan USB warnings with gcc 3 (me)
* Switch to the newer 2.4 depca driver (me)
* Re-merge depca fixes from 2.5.0->2.5.40]
* Fix depca spinning waiting for irq probe (me)
* Fix depca copy with interrupts off (me)
* Fix depca clash with other ALIGN macros (me)
* Initial port of NCR5380/g_NCR5380 to new locks (me)
| This still needs new_eh, further clean up
| and possibly making NCR5380_main a thread
* Initial locking rework for the wd7000 scsi (me)
| Still needs new_eh
* Update jffs to the dequeue_signal changes (me)
* Update jffs2 to the dequeue_signal changes (me)
* Fix shpnt misuse in NCR53c406a, wrong free_irq (me)
* Update NCR53c406a to new style sglist (me)
| Still needs new_eh
* Architecture updates for S/390 (Martin Schwidefsky)
* Include updates for S/390 (Martin Schwidefsky)
* Base S/390 driver updates (Martin Schwidefsky)
* Add the new syscalls to S/390 (Martin Schwidefsky)
* Fix sleeping with locks in sound_core (Jaroslav Kysela)
* Fix oops on shutdown of cs4281 (Suresh Siddha)
* Fix cdrom paths in devfs (Jordan Breeding)
* Fix missing cache tag entry in intel cpu table (Jean Delvare)
* Remove old 2.2 compatibility pci functions (Greg Kroah-Hartmann)
* Clean up some dead devfs bits (Greg Kroah-Hartmann)
* Fix an oops in the hugetblpage stuff (Andrew Morton)
| Its still a stupid idea but now it doesnt oops
o Handle read only BARs with type bits set (Ivan Kokshaysky)
Linux 2.5.40-ac2
* Fix a cut and paste error in the amd rng docs (Troels Hansen)
* Forward port OSS maestro3 fixes for toughbook
o Forward port ramdisk cache coherency
* RTL8150 USB updates (Petko Manalov)
* Fix corega USB ident (Petko Manalov)
* USB keyboard driver fix (Dave Miller)
* USB prototype fix (Luc Vanoostenryck)
* USB string fixes ([email protected])
* USB test driver (David Brownell)
* Speedtouch USB driver fixes (Greg Kroah-Hartmann)
* Clean environment for hotplug (Greg Kroah-Hartmann)
* Fix mprotect oops (Hugh Dickins)
o NUMA-Q cleanups (Martin Dobson)
* Split timers into one x86 timer type per file (John Stultz)
* Cyclone timer support for x440 etc (John Stultz)
* Fix sleeping from illegal context for ioperm (Andrew Morton)
* Fix imm compile ([email protected])
* Fix irda for tq changes (Carlos Gorges)
* Fix xjack telephony build (Carlos Gorges)
* Fix ppa compile (Carlos Gorges)
* Fix aha152x compile for tq changes (Carlos Gorges)
* Fix hamradio drivers for tq changes (Carlos Gorges)
* Fix plip driver for tq changes (Carlos Gorges)
* Fix mpt fusion for tq changes (Carlos Gorges)
* Fix isdn for tq changes (Carlos Gorges)
* Fix ieee1394 for tq changes (Carlos Gorges)
* Fix new timer code to build with cpufreq on (me)
* Fix capi build for new tq_ code (me)
| ISDN still needs moving to real locks
| this just cleans up one item
* Fix missing header in mtdblock_ro (Carlos Gorges)
* Fix a typo and other header (me)
* Fix up ixj_pcmcia for 2.5 (me)
| Note for janitors - it looks like a lot of the pcmcia release
| code people "fixed" should be using del_timer_sync not del_timer
* Fix missing header in longhaul cpu speed driver (me)
* Pipe read/write cleanup (Manfred Spraul)
* Make IDE PCI config text clearer (Andrzej Krzysztofowicz)
Linux 2.5.40-ac1
* Initial port of aacraid driver to 2.5 (me)
* vfat corruption fix (Petr Vandrovec)
* Clean up firestream warnings (Francois Romieu)
+ Voyager support (James Bottomley)
* Fix split_vma (Hugh Dickins)
+ Fix config in video subdirectory (John Levon)
* Update olympic driver to 2.5 (Mike Phillips)
* Fix sg init error (Mike Anderson)
* Fix Rules.make
o Merge most of ucLinux stuff (Greg Ungerer)
| It needs putting somewhere so we can pick over the
| hard bits left
| Q: Wouldn't drivers/char/mem-nommu.c be better
| Q: How to do the procfs stuff tidily
| Q: Wouldn't it be nicer to move all mm or mmnommu specific ksyms
| int the relevant mm/*.c file area instead of kernel/ksyms
| Q: Why ifdef out overcommit - its even easier to account on
| MMUless and useful info
* Stick tulip back under 10/100 ethernet (me)
* Correct docs for IBM touchpad back to how (me)
they were before
o Fix abuse of set_bit in winbond-840 (me)
* Fix abuse of set_bit in atp (me)
Alan Cox wrote:
[SNIP]
> o Fix dm includes (me)
> | I've not merged any of the DM updates yet
Would it be possible to provide the patch you used to merge them in the
first place? Joe has since junked the ones on his webpage which you used,
so it would be helpful for those of us who might want to back it out and
try the newer updates. Thanks in advance!
Cheers,
Nicholas
On Tue, 2002-10-22 at 18:56, Nicholas Wourms wrote:
> Alan Cox wrote:
>
> [SNIP]
> > o Fix dm includes (me)
> > | I've not merged any of the DM updates yet
>
> Would it be possible to provide the patch you used to merge them in the
> first place? Joe has since junked the ones on his webpage which you used,
> so it would be helpful for those of us who might want to back it out and
> try the newer updates. Thanks in advance!
Im hoping Joe will send me the updates he wants in the tree. If not I'll
have to go digging
On Tue, 2002-10-22 at 11:27, Alan Cox wrote:
[snip]
> This one builds, its not yet had any measurable testing
>
> Linux 2.5.44-ac1
> - Resync with Linus 2.5.43/44
[snip]
> o Move BUG() into asm/bug.h (Russell King)
I got this error building with CONFIG_REISERFS_FS=y:
fs/built-in.o: In function `keyed_hash':
fs/built-in.o(.text+0x84d56): undefined reference to `BUG'
fs/built-in.o(.text+0x84e08): undefined reference to `BUG'
fs/built-in.o(.text+0x84ea4): undefined reference to `BUG'
fs/built-in.o(.text+0x84f0e): undefined reference to `BUG'
make: *** [.tmp_vmlinux1] Error 1
The following patch allows 2.5.44-ac1 to build with reiserfs.
Steven
--- linux-2.5.44-ac1/fs/reiserfs/hashes.c.orig Tue Oct 22 12:16:52 2002
+++ linux-2.5.44-ac1/fs/reiserfs/hashes.c Tue Oct 22 12:17:15 2002
@@ -20,6 +20,7 @@
#include <asm/types.h>
#include <asm/page.h>
+#include <asm/bug.h>
Note that the patch <http://chunk.mp3revolution.net/lvm2/patches/09.patch>
is necessary for Joe's older stuff; otherwise, dm oopses (with
2.5.44, anyways; have not yet tried -ac1). If you don't merge any of
the newer DM stuff, please at least fix the lack of gendisk
initialization...
On Tue, Oct 22, 2002 at 01:27:06PM -0400, Alan Cox wrote:
>
> ** I strongly recommend saying N to IDE TCQ options otherwise this
> should hopefully build and run happily.
>
> Doug's scsi changes broke mptfusion. I've not looked into that yet
> also u14f/u34f, and the host changes broke all 5380 based devices
>
> This one builds, its not yet had any measurable testing
>
> Linux 2.5.44-ac1
> - Resync with Linus 2.5.43/44
> o Fix net/ipv4/raw.c build problem (me)
> o Fix bluetooth pcmcia builds (me)
> o Fix dm includes (me)
> | I've not merged any of the DM updates yet
[...]
> -
> 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/
--
It's not denial. I'm just selective about the reality I accept.
-- Bill Watterson
On Tue, 22 Oct 2002, Alan Cox wrote:
>...
> Linux 2.5.44-ac1
>...
> o Clean up wan ioctl structures (Krzysztof Halasa)
>...
This caused the following compile error:
<-- snip -->
...
gcc -Wp,-MD,drivers/net/wan/.hdlc_x25.o.d -D__KERNEL__ -Iinclude -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing
-fno-common -pipe -mpreferred-stack-boundary=2 -march=k6
-Iarch/i386/mach-generic -nostdinc -iwithprefix include -DKBUILD_BASENAME=hdlc_x25 -c -o
drivers/net/wan/hdlc_x25.o drivers/net/wan/hdlc_x25.c
drivers/net/wan/hdlc_x25.c: In function `hdlc_x25_ioctl':
drivers/net/wan/hdlc_x25.c:187: invalid type argument of `->'
drivers/net/wan/hdlc_x25.c:189: invalid type argument of `->'
drivers/net/wan/hdlc_x25.c:190: warning: unreachable code at beginning of
switch statement
make[3]: *** [drivers/net/wan/hdlc_x25.o] Error 1
<-- snip -->
The follwing patch seems to fix it:
--- linux-2.5.44-full-ac/drivers/net/wan/hdlc_x25.c.old 2002-10-22 21:38:32.000000000 +0200
+++ linux-2.5.44-full-ac/drivers/net/wan/hdlc_x25.c 2002-10-22 21:38:51.000000000 +0200
@@ -184,9 +184,9 @@
struct net_device *dev = hdlc_to_dev(hdlc);
int result;
- switch (ifr->ifr_settings->type) {
+ switch (ifr->ifr_settings.type) {
case IF_GET_PROTO:
- ifr->ifr_settings->type = IF_PROTO_X25;
+ ifr->ifr_settings.type = IF_PROTO_X25;
return 0; /* return protocol only, no settable parameters */
case IF_PROTO_X25:
The next compile error is in a file that seems to need a bit more work:
<-- snip -->
...
gcc -Wp,-MD,drivers/net/wan/.pc300_drv.o.d -D__KERNEL__ -Iinclude -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing
-fno-common -pipe -mpreferred-stack-boundary=2 -march=k6
-Iarch/i386/mach-generic -nostdinc -iwithprefix include -DKBUILD_BASENAME=pc300_drv -c -o
drivers/net/wan/pc300_drv.o drivers/net/wan/pc300_drv.c
drivers/net/wan/pc300_drv.c: In function `cpc_ioctl':
drivers/net/wan/pc300_drv.c:2545: invalid type argument of `->'
drivers/net/wan/pc300_drv.c:2744: invalid type argument of `->'
drivers/net/wan/pc300_drv.c:2748: invalid type argument of `->'
drivers/net/wan/pc300_drv.c:2776: invalid type argument of `->'
drivers/net/wan/pc300_drv.c:2799: invalid type argument of `->'
drivers/net/wan/pc300_drv.c:2747: warning: unreachable code at beginning
of switch statement
make[3]: *** [drivers/net/wan/pc300_drv.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
On Tue, 2002-10-22 at 10:27, Alan Cox wrote:
> o Fix blk ioctls on aacraid (Mark Haverkamp)
Somewhere between 2.5.43 and 2.5.44 the blk ioctls code changed
elsewhere making this one unnecessary.
Mark.
Andres Salomon wrote:
> Note that the patch <http://chunk.mp3revolution.net/lvm2/patches/09.patch>
> is necessary for Joe's older stuff; otherwise, dm oopses (with
> 2.5.44, anyways; have not yet tried -ac1). If you don't merge any of
> the newer DM stuff, please at least fix the lack of gendisk
> initialization...
>
Joe's got a new set of patches on his homepage, but unfortunately they
aren't the same ones that Alan used. I don't suppose you have patches 1-8
of the original set which I could use to back out the old code? I was
palnning on doing that and then patching in the new code.
Cheers,
Nicholas
07 and 08 were my own patches; ejt released 0-6. I believe the ones he
sent alan are the same as the ones on
<http://people.sistina.com/~thornber/patches/2.5-stable/>; not
positive, though.
For those interested, his new stuff is in
<http://people.sistina.com/~thornber/patches/2.5-stable/>..
On Tue, Oct 22, 2002 at 06:53:36PM -0400, Nicholas Wourms wrote:
>
> Andres Salomon wrote:
>
> > Note that the patch <http://chunk.mp3revolution.net/lvm2/patches/09.patch>
> > is necessary for Joe's older stuff; otherwise, dm oopses (with
> > 2.5.44, anyways; have not yet tried -ac1). If you don't merge any of
> > the newer DM stuff, please at least fix the lack of gendisk
> > initialization...
> >
>
> Joe's got a new set of patches on his homepage, but unfortunately they
> aren't the same ones that Alan used. I don't suppose you have patches 1-8
> of the original set which I could use to back out the old code? I was
> palnning on doing that and then patching in the new code.
>
> Cheers,
> Nicholas
>
>
> -
> 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/
--
It's not denial. I'm just selective about the reality I accept.
-- Bill Watterson
On Tue, 2002-10-22 at 13:27, Alan Cox wrote:
> ** I strongly recommend saying N to IDE TCQ options otherwise this
> should hopefully build and run happily.
I don't have the IDE TCQ options on and am still having an issue with a
kernel oops in the default kernel code. I did some digging after I
noticed it in 2.5.43 and found a possibility for a null pointer in the
code within the ide_iomio_dma function in ide-dma.c The problem shows
itself if you only enable the secondary channel of your IDE controller.
I understand this is a strange set up, but it could happen in a machine
that boots off of SCSI and uses IDE disks for DATA or a CD Burner. I
came up with a fix, some extra sanity checks before this line in the
code:
hwif->dma_master = (hwif->channel) ? hwif->mate->dma_base : base;
Because it is the hwif->mate variable that is null when it gets here and
if hwif->channel is not 0 I am taking it from the logic on this line
that hwif->mate should not be a null variable. So I applied the sainity
checks in the attached patch and this seems to work. I did this against
the 2.5.43 and 2.5.44 kernels, and have posted patches for each of them
but no one seems interested. If I am completely off base here, just let
me know and I'll go back to the code and look for a different way to
conquer this little bug.
Jason
> code within the ide_iomio_dma function in ide-dma.c The problem shows
> itself if you only enable the secondary channel of your IDE controller.
> I understand this is a strange set up, but it could happen in a machine
> that boots off of SCSI and uses IDE disks for DATA or a CD Burner. I
> came up with a fix, some extra sanity checks before this line in the
> code:
Yes I saw the report. I've not applied it because I want to know how
the slave came not to have a hwif->mate even though it was bios disabled.
There are other things that really mean we should be assigning the hwif
pointers (eg hot plugging)
On Wed, 2002-10-23 at 07:49, Alan Cox wrote:
> > code within the ide_iomio_dma function in ide-dma.c The problem shows
> > itself if you only enable the secondary channel of your IDE controller.
> > I understand this is a strange set up, but it could happen in a machine
> > that boots off of SCSI and uses IDE disks for DATA or a CD Burner. I
> > came up with a fix, some extra sanity checks before this line in the
> > code:
>
> Yes I saw the report. I've not applied it because I want to know how
> the slave came not to have a hwif->mate even though it was bios disabled.
> There are other things that really mean we should be assigning the hwif
> pointers (eg hot plugging)
That is what I thought you might say. I have been digging around in the
code, trying to figure out the answer to that question myself. For some
reason, even though that patch worked, it just didn't 'feel' right. I
was looking for someone to confirm that there was a bigger problem to
search for. Now I'll go back and hunt down the bigger beast and find
out why that hwif->mate is coming out null.
Jason