There it is (or will soon be - the tar-ball and patches are still
uploading, and mirroring can obviously take some time after that).
It's slightly bigger than I'd like, but that's partly because I had (once
more) missed a merge that was actually sent in well before the -rc1
cut-off, so the x86-64 merge is from there and was delayed due to yours
truly, not Andi Kleen.
Apart from the x86-64 merge and various fixups, I've let MIPS, PARISC and
PowerPC merge up some more.
The shortlog speaks for itself.
Linus
----
Adrian Bunk:
arch/i386/mm/init.c: small cleanups
Al Viro:
m68k: introduce task_thread_info
m68k: introduce setup_thread_stack() and end_of_stack()
m68k: thread_info header cleanup
m68k: m68k-specific thread_info changes
Albert Lee:
libata: honor the transfer cycle time speficied by the EIDE device
Alexey Dobriyan:
alim15x3: use KERN_WARNING
Amit Gud:
cs5520: fix return value of cs5520_init_one()
Andi Kleen:
x86_64: Update defconfig
x86_64: Add 4GB DMA32 zone
x86_64: Set compatibility flag for 4GB zone on IA64
x86_64: Make i386 compile again with fourth DMA32 zone
x86_64: When cpu_up fails clean up page allocator properly
x86_64: Account mem_map in VM holes accounting
x86_64: Fix up outdated pfn_to_page comment
x86_64: Remove obsolete ARCH_HAS_ATOMIC_UNSIGNED and page_flags_t
x86_64: Use the DMA32 zone for dma_alloc_coherent()/pci_alloc_consistent
x86_64: Fix gcc 4 warning in aperture.c
x86_64: Speed up numa_node_id by putting it directly into the PDA
x86_64: Don't apply __PHYSICAL_MASK to page frame numbers
x86_64: Only use asm/sections.h to declare section symbols
x86_64: Replace cpu_pda extern with include
x86_64: Replace swiotlb extern with include
x86_64: Some clarifications for Documention/x86_64/mm.txt
x86_64: Use int operations in spinlocks to support more than 128 CPUs spinning.
x86_64: New heuristics to find out hotpluggable CPUs.
AGP: Support ULI/ALI 1689 bridge on AMD64
AGP: Try unsupported AGP chipsets on x86-64 by default
AGP: Make gart iterator in K8 AGP driver SMP safe
x86_64: Allow modular build of ia32 aout loader
x86_64: Formatting fixes for arch/x86_64/kernel/process.c
x86_64: Don't enable interrupt unconditionally in reboot path
x86_64: Fix NUMA node lookup debug code which had bitrotted
x86_64: Reduce number of retries for reset through keyboard controller
x86_64: Remove optimization for B stepping AMD K8
x86_64: Remove asm-x86_64/rwsem.h
x86_64: Log machine checks from boot on Intel systems
x86_64: Remove CONFIG_CHECKING and add command line option for pagefault tracing
x86_64: Increase the maximum number of local APICs to the maximum
Andrew Morton:
nv_of.c build fix
rpaphp_pci build fix
pciehp_hpc build fix
shpchp_hpc build fix
powerpc-xmon-build-fix
acct.h needs jiffies.h
v4l-944-added-driver-for-saa7127-video-tidy
hfc_usb: fix usb device table
USB: usbdevfs_ioctl 32bit fix
usb devio warning fix
git-netdev-all-ieee80211_get_payload-warning-fix
Andrey Volkov:
[SERIAL] Fix mpc52xx_uart.c
[DRIVER MODEL] Fix typo in ohci-ppc-soc.c
Fix copy-paste bug in ohci-ppc-soc.c
Andy Whitcroft:
ppc64 need HPAGE_SHIFT when huge pages disabled
Antonino A. Daplas:
fbdev: fix module dependency loop
nvidiafb: Fix bug in nvidiafb_pan_display
Antti Andreimann:
USB: Maxtor OneTouch button support for older drives
Arnaud Giersch:
[MIPS] IP32: Export mace symbol.
[MIPS] IP32 Fix and complete IP32 parport definitions
[MIPS] IP32: Fix sparse warnings.
[MIPS] Add const qualifier to writes##bwlq.
[MIPS] Fix documentation typos.
Ashok Raj:
x86_64: Remove duplicate __cpuinit define
Aurelien Jarno:
sis5513: enable ATA133 for the SiS965 southbridge
Bartlomiej Zolnierkiewicz:
ide: remove duplicate documentation for ide_do_drive_cmd()
ide: remove unused ide_action_t:ide_next
ide: remove dead DEBUG_TASKFILE code
ide: remove dead code from flagged_taskfile()
ide: add missing __init tags to device drivers
Ben Collins:
Add missing EXPORT_SYMBOLS() for __ide_mm_* functions on powerpc
Update location of ll_rw_blk.c in docs
Ben Dooks:
[ARM] 3161/1: BAST - fix commas on end of structs
[ARM] 3162/1: S3C2410 - updated defconfig
Benjamin Herrenschmidt:
powerpc: Always rebuild arch/powerpc/include/asm symlink
powerpc: vdso fixes (take #2)
powerpc: kill ppc64 rtc.c, use genrtc instead
powerpc: update defconfigs
powerpc: pci_64 fixes & cleanups
ppc: Fix boot with yaboot with ARCH=ppc
ppc: Fix build with CONFIG_CHRP not set
powerpc: Make the vDSO functions set error code (#2)
powerpc: Workaround for offb on 64 bits platforms
powerpc: merge align.c
powerpc: Fix setting MPIC priority
Bill Pechter:
v4l:: (936) Support for sabrent bt848 version
Bjorn Helgaas:
[SERIAL] Claim Wacom tablet device on HP tc1100 tablet
Bob Picco:
cpuset: fix return without releasing semaphore
x86_64: Fix sparse mem
Bryan Ford:
x86_64: Save/restore CS in 64bit signal handlers and force __USER_CS for CS
Carlos O'Donell:
[PARISC] Document some register usages in assembly files
Chen, Kenneth W:
ia64: cpu_idle performance bug fix
[IA64] 4 level page table bug fix in vhpt_miss
[IA64] polish comments for tlb fault handler in ivt.S
Chris Wright:
VFS: local denial-of-service with file leases
Christoph Hellwig:
[SPARC]: Fix RTC compat ioctl kernel log spam.
[SBUSFB]: implement ->compat_ioctl
fix task_struct leak in ptrace
v850: use generic hardirq code
[PARISC] move PA perf driver over to ->compat_ioctl
[PARISC] remove drm compat ioctls handlers
Christoph Lameter:
slab: remove alloc_pages() calls
Clemens Buchacher:
arch/mips/au1000/common/usbdev.c: don't concatenate __FUNCTION__ with strings
Constantine Gavrilov:
x86: fix sigaddset() inline asm memory constraint
Corey Minyard:
ipmi: bump-driver-version
Coywolf Qi Hunt:
[BLOCK] new block/ directory comment tidy
Daniel Drake:
usb-storage: Fix detection of kodak flash readers in shuttle_usbat driver
via82cxxx IDE: remove /proc/via entry
via82cxxx IDE: support multiple controllers
Daniel Jacobowitz:
[ARM] 3168/1: Update ARM signal delivery and masking
Dave Jones:
v4l: saa711x driver doesn't need segment.h
oops-tracing: mention extended VGA
David Brownell:
USB: onetouch doesn't suspend yet
David Gibson:
powerpc: Remove imalloc.h
David S. Miller:
[DVB] cinergyT2: cinergyt2_register_rc() should return 0 on success
[DVB]: Add compat ioctl handling.
[COMPAT]: Add ext3 ioctl translations.
[LLC]: Fix compiler warnings introduced by TX window scaling changes.
[IPV6]: Fib dump really needs GFP_ATOMIC.
[COMPAT]: EXT3_IOC_SETVERSION is _IOW() not _IOR().
David Woodhouse:
Avoid use of uninitialised spinlock in EEH.
Deepak Saxena:
Fix IXP4xx I2C driver build breakage
Denis Lunev:
ext3: journal handling on error path in ext3_journalled_writepage()
Diego Calleja:
oops-tracing: mention digital photos
Dmitry Torokhov:
I8K: fix /proc reporting of blank service tags
USB: fix 'unused variable' warning
Dominik Brodowski:
[PCMCIA] i82365: use new platform_device helpers
[PCMCIA] inform user of insertion and ejection events
Eric Dumazet:
reorder struct files_struct
x86_64: Optimize NUMA node hash function
Florin Malita:
[SERIAL] sa1100_start_tx spinlock recursion
Francois Romieu:
r8169: fix printk_ratelimit in the interrupt handler
r8169: do not abort when the power management capabilities are disabled
Gabriel A. Devenyi:
drivers/net/wireless/hermes.c unsigned int comparision
George Anzinger:
timespec: normalize off by one errors
Grant Coady:
cciss_scsi warning fix
Grant Grundler:
[PARISC] Disable nesting of interrupts
[PARISC] irq_affinityp[] only available for SMP builds
[PARISC] Remove unused variable in signal.c
Greg Kroah-Hartman:
USB: fix build breakage in dummy_hcd.c
USB Serial: rename ChangeLog.old
USB: move CONFIG_USB_DEBUG checks into the Makefile
USB: delete the nokia_dku2 driver
USB: add the anydata usb-serial driver
Add HOWTO do kernel development document to the Documentation directory
update Documentation/00-INDEX
Guido Guenther:
PowerBook 6,1: headphone not detected after suspend
[SPARC64]: Oops in pci_alloc_consistent with cingergyT2
Hanna Linder:
alim15x3: replace pci_find_device() with pci_dev_present()
Hans Reiser:
re-export clear_page_dirty_for_io()
Hans Verkuil:
v4l: (944) added driver for saa7127 video decoder
v4l: (945) adds a new include for internal v4l2 ioctls and api
v4l: (946) adds support for cx25840 video decoder
v4l: (948) adds support for saa7115 video decoder
v4l: (966) Authorship fixes for new Modules
v4l: 976: ensure consistent v4l firmware prefixes
Harald Welte:
New Omnikey Cardman 4040 driver
New Omnikey Cardman 4000 driver
[NETFILTER] nfnetlink: unconditionally require CAP_NET_ADMIN
Make sysctl.h (again) usable from userspace
[NETFILTER] ip_conntrack: fix ftp/irc/tftp helpers on ports >= 32768
Hartmut Hackmann:
v4l: (949) Added support for secam l'
Heiko Carstens:
signal handling: revert sigkill priority fix
Herbert Xu:
[IPV6]: Fix rtnetlink dump infinite loop
USB: fix race in kaweth disconnect
Ingo Molnar:
rcutorture: renice to low priority
J. Bruce Fields:
VFS: Fix memory leak with file leases
Jacob Shin:
x86_64: Support for AMD specific MCE Threshold.
James Bottomley:
[PARISC] Make sure timer and IPI execute with interrupts disabled
[PARISC] Fix our interrupts not to use smp_call_function
[PARISC] Add IRQ affinities
[PARISC] Fix our spinlock implementation
ide: fix ide_toggle_bounce() to not try to bounce if we have an IOMMU
James Cleverdon:
i386/x86-64: Share interrupt vectors when there is a large number of interrupt sources
James Ketrenos:
ipw2100: Fix 'Driver using old /proc/net/wireless...' message
Jan Beulich:
i386: NMI pointer comparison fix
make vesafb build without CONFIG_MTRR
x86_64: Adjust, correct, and complete the HPET definitions for x86-64.
Jeff Garzik:
[libata ahci, qstor] fix miscount of scatter/gather entries
[libata ahci] set port ATAPI bit correctly
[libata sata_mv] minor fixes
[libata sata_mv] trim trailing whitespace
[libata sata_mv] note driver is "HIGHLY EXPERIMENTAL" in Kconfig
[libata sata_mv] implement a bunch of errata workarounds
[libata sata_mv] move code around
[libata sata_mv] mv_hw_ops for hardware families; new errata
[libata sata_mv] hardware initialization work
[libata sata_mv] move code around
[libata sata_mv] call phy fixups during init, as well as phy reset
[libata sata_mv] fix tons of 50XX bugs
move pm_register/etc. to CONFIG_PM_LEGACY, pm_legacy.h
[libata ahci] error handling fixes
[libata] fix bugs in ATAPI padding DMA mapping code
[libata] minor fixes, new helpers
[libata] REQUEST SENSE handling fixes
[libata ahci] command completion fixes, improved debug msgs
[libata ahci] tone down ATAPI errors
[libata] bump versions
[libata] add timeout to commands for which we call wait_completion()
[libata sata_mv] SATA probe, DMA boundary fixes
[libata sata_mv] handle lack of hardware nIEN support
[libata sata_mv] update copyright, driver version
[wireless hermes] build fix
siimage: docs urls
Jens Axboe:
[BLOCK] Document the READ/WRITE splitup of the disk stats
VM: fix zone list restart in page allocatate
[PATCH 2/3] cciss: bug fix for BIG_PASS_THRU
Jesper Juhl:
README: add info about -stable to README and point at applying-patches.txt
Jesse Brandeburg:
e100: re-enable microcode with more useful defaults
Jochen Friedrich:
[LLC]: Fix TX window scaling
[LLC]: Make core block on remote busy.
[LLC]: Fix typo
Jody McIntyre:
Add SCM info to MAINTAINERS
Johann Lombardi:
ext2: remove duplicate newlines in ext2_fill_super
John W. Linville:
i82593.h: make header comment GPL-compatible
fec_8xx: make CONFIG_FEC_8XX depend on CONFIG_8xx
Josef Balatka:
USB: cp2101.c: Jablotron usb serial interface identification
Karsten Wiese:
x86_64 two timer entries in /sys
Kirill Korotaev:
mm: __GFP_NOFAIL fix
stop_machine() vs. synchronous IPI send deadlock
KOVACS Krisztian:
[NETFILTER] nf_conntrack: Add missing code to TCP conntrack module
[NETFILTER] Remove nf_conntrack stat proc file when cleaning up
[NETFILTER] Free layer-3 specific protocol tables at cleanup
Krzysztof Halasa:
Generic HDLC WAN drivers - disable netif_carrier_off()
Krzysztof Oledzki:
[NETFILTER]: link 'netfilter' before ipv4
Kumar Gala:
Update email address for Kumar
ppc32: Add support for handling PCI interrupts on MPC834x PCI expansion card
powerpc: replace page_to_virt() with lowmem_page_address() for Book-E
ppc: Fix warnings related to seq_file
ppc: Fix MPC83xx device table
ppc: Fix warnings related to seq_file
Kyle McMartin:
[PARISC] Fix uniprocessor build by dummying smp_send_all_nop()
[PARISC] Make superio.c initialize before any driver needs it
[PARISC] Update CREDITS entries
Kylene Jo Hall:
tpm: necessary PPC64 function exports
tpm: updates for new hardware
tpm: dev_mask handling fix
tpm: locking fix
tpm: use flush_scheduled_work()
tpm: use ioread8 and iowrite8
tpm: remove PCI kconfig dependency
Laurent Riffard:
ide: remove ide_driver_t.owner field
Lennert Buytenhek:
[SERIAL] don't disable xscale serial ports after autoconfig
Linus Torvalds:
Revert "fbcon: Add rl (Roman Large) font"
x86: Fix silly typo in recent <asm/signal.h> fixes
Fix ACPI processor power block initialization
Linux v2.6.15-rc2
Luiz Capitulino:
[IPV6]: Fixes sparse warning in ipv6/ipv6_sockglue.c
Luiz Fernando Capitulino:
Fix sparse warning in proc/task_mmu.c
USB: pl2303: adds new IDs.
USB: pl2303: updates pl2303_update_line_status()
Maciej W. Rozycki:
[MIPS] zs.c: Resurrect the deceased zs.c for now.
Magnus Damm:
x86_64: Make node boundaries consistent
Marcel Holtmann:
USB: Delete leftovers from bluetty driver
Marcelo Tosatti:
ppc32 8xx: update_mmu_cache() needs unconditional tlbie
Mark Lord:
libata: fix comments on ata_tf_from_fis()
[libata passthru] address slave devices correctly
Mark Weaver:
v4l: (939) Support for nebula rc5 based gpio remote
Martin Schwidefsky:
s390: fix class_device_create calls in 3270 the driver
Martin Waitz:
DocBook: allow to mark structure members private
DocBook: include printk documentation
DocBook: comment about paper type
DocBook: revert xmlto use for .ps and .pdf documentation
Mathias Kretschmer:
via82cxxx: add VIA VT6410 IDE support
Matt Domsch:
ipmi: missing NULL test for kthread
Matthew Wilcox:
[PARISC] Return PDC_OK when alloc_pa_dev fails to enumerate all devices
[PARISC] Improve the error message when we get a clashing mod path
[PARISC] Fix some compile problems in ptrace.c
[PARISC] Always spinlock tlb flush operations to ensure preempt safety
[PARISC] Fix compile warning caused by conflicting types of expand_upwards()
[PARISC] Make Serial MUX depend on a specific bus type.
[PARISC] Mention PA-RISC in NS87415 help
[PARISC] Mark hisax and pcbit ISDN drivers as not for parisc
matthieu castet:
fix leaks in request_firmware_nowait
Mauro Carvalho Chehab:
v4l: (926.1) Added compiling options for wm8775 and cs53l32a chips
v4l: (943) added secam l video standard
v4l: (950) Added compiler options for cx25840 saa7115 and saa7127
v4l: (963) em28xx IR fixup
v4l: (966.1) Removes Obsoleted i2c-compat.h from newer drivers
v4l: 977: fix broken dependency needed for sa7134 module
Michael Ellerman:
powerpc: Merge page.h
powerpc: Turn cpu_irq_down into kexec_cpu_down
powerpc: Export htab start/end via device tree
powerpc: Fixup debugging in lmb.c
powerpc: More debugging fixups
powerpc: Fix typo in topology.h
Michael Krufky:
v4l: (963.1) hybrid v4l/dvb: remove duplicated code
v4l: 974: saa7134 shouldn't DEPEND on SND_PCM_OSS. Instead, SELECT it.
Michael S. Tsirkin:
IB/mthca: Safer max_send_sge/max_recv_sge calculation
Mike Kravetz:
Remove SPAN_OTHER_NODES config definition
Mike Krufky:
v4l: prevent saa7134 alsa undefined warnings
mikem:
[PATCH 1/3] cciss: bug fix for hpacucli
[PATCH 3/3] cciss: add put_disk into cleanup routines
Miles Bader:
v850: Fix show_interrupts
v850: Add missing include in hardirq.h
Neil Brown:
md: don't pass a NULL file* into ->prepare_write()
NeilBrown:
knfsd: make sure nfsd doesn't hog a cpu forever
md: mark START_ARRAY deprecated with a date
md: make md threads interruptible again
md: fix is_mddev_idle calculation now that disk/sector accounting happens when request completes
Nick Piggin:
mm: highmem watermarks
i386: generic cmpxchg
atomic: cmpxchg
atomic: inc_not_zero
powerpc: Fix database regression due to scheduler changes
Nickolay V. Shmyrev:
v4l: (937) Included missing interrupt.h at saa7134-alsa.c
Nicolas Pitre:
[ARM] 3165/1: fix atomic_cmpxchg() implementation for ARMv6+
smc91x: fix one source of spurious interrupts
OGAWA Hirofumi:
usbfs: usbfs_dir_inode_operations cleanup
Olaf Hering:
ppc boot: replace string labels with numbers
Oliver Neukum:
USB: Adapt microtek driver to new scsi features
Olof Johansson:
ppc: add support for new powerbooks
powerpc: add new powerbooks to feature table
Pablo Neira Ayuso:
[NETFILTER] ctnetlink: use size_t to make gcc-4.x happy
[NETFILTER] nfnetlink: skip size check if size not specified (== 0)
[NETFILTER] ctnetlink: More thorough size checking of attributes
Pantelis Antoniou:
[MIPS] Alchemy: Console output fixup
Paolo 'Blaisorblade' Giarrusso:
Kbuild: index asm-$(SUBARCH) headers for UML
uml: remove bogus WARN_ON, triggerable harmlessly on a page fault race
uml: micro fixups to arch Kconfig
uml: fixups for "reuse i386 cpu-specific tuning"
uml: fix mcast network driver error handling
uml console channels: remove console_write wrappers
uml console channels: fix the API of console_write
uml: fix access_ok
uml: fix daemon transport exit path bug
x86_64: Set ____cacheline_maxaligned_in_smp alignment to 128 bytes
x86_64: Use common sys_time64
Patrick McHardy:
[NETFILTER]: Fix nf_conntrack compilation with CONFIG_NETFILTER_DEBUG
Paul E. McKenney:
add success/failure indication to RCU torture test
Paul Fulghum:
synclink: update to use DMA mapping API
Paul Jackson:
mm: gfp_noreclaim cleanup
Paul Mackerras:
powerpc: Move a bunch of ppc64 headers to include/asm-powerpc
powerpc: Move most remaining ppc64 files over to arch/powerpc
powerpc: Export a couple of prom functions
powerpc: Mark PREP and embedded as broken for now
powerpc: Fix 32-bit compile: PPC_MEMSTART was undeclared
powerpc: Fix clearing of the FPSCR when invoking a signal handler
powerpc: Remove an extraneous and incorrect declaration of pmac_nvram_init.
powerpc: Remove __init from a function used in suspend/resume.
powerpc: Fix sparsemem with memory holes [was Re: ppc64 oops..]
powerpc: Move ppc64 boot wrapper code over to arch/powerpc
powerpc: Fix delay functions for 601 processors
powerpc: Move remaining .c files from arch/ppc64 to arch/powerpc
powerpc: Fix compile error on pSeries arising from delay.h changes
powerpc: time-of-day fixes for 32-bit CHRP systems
powerpc: Fix a couple of compile warnings for 32-bit compiles
powerpc: Move defconfig over and remove remaining arch/ppc64 files
offb: Fix compile error on ppc32 systems
powerpc: Trivially merge several headers from asm-ppc64 to asm-powerpc
powerpc: Merge pci.h
powerpc: move include/asm-ppc64/ptrace-common.h to arch/powerpc/kernel
powerpc: Merge spinlock.h
powerpc: Fix bug in timebase synchronization on 32-bit SMP powermac
Paul Mundt:
Shut up per_cpu_ptr() on UP
Pavel Machek:
[ARM] Fix collie for -rc1
USB: kill unneccessary usb-storage blacklist entries
Pekka Enberg:
slab: convert cache to page mapping macros
ipw2200: disallow direct scanning when device is down
Peter Osterlund:
packet writing oops fix
Ping Cheng:
USB: add new wacom devices to usb hid-core list
USB: wacom tablet driver update
Prakash Punnoor:
fix b2c2 dvb undefined symbol
Ralf Baechle:
[SERIAL] dz: Nuke trailing whitespace
[SERIAL] dz: Use CKSEG1ADDR to setup mappings.
[MIPS] Delete duplicate definitions of break codes.
[MIPS] feature-removal-schedule.txt: Schedule au1x00_uart for removal.
[MIPS] Add missing arch defines for the Alchemy MTD driver.
Add definitions for the Dallas DS17287 RTC.
Add definitions for the Dallas DS1742 RTC / non-volatile memory.
[MIPS] IP32: No need to include <asm/io.h>.
[MIPS] DDB5477: Fix unused variable warning.
[MIPS] JMR3927: Undo accidental rename.
[MIPS] JMR3927: Fix syntax error.
[IDE] Add driver for Sibyte Swarm evaluation board
[MIPS] JMR3927: It's ops-tx3927.o not ops-jmr3927.o
[MIPS] JMR3927: need include/asm-mips/mach-jmr3927 in it's include path.
[MIPS] JMR3927: Fix compilation by including <linux/ds1742rtc.h>.
[MIPS] JMR3927: Fix include wrapper symbol.
[MIPS] Ocelot G: Use CPU_MASK_NONE instead of 0 to initialize cpu mask.
[MIPS] SEAD: Delete seadint_init() prototype.
[MIPS] TX3927: Try to glue the PCI code.
[MIPS] SEAD: More build fixes.
[MIPS] Update defconfigs
IOC3: Replace obsolete PCI API
au1000_eth: Include <linux/config.h>
SAA9730: Add missing header bits.
ide: make comment match reality
Ravikiran G Thirumalai:
x86_64: Make ACPI NUMA and NUMA emulation peers of K8_NUMA in Kconfig
Ricardo Cerqueira:
v4l: (930) Alsa fixes and improvements
v4l: (935) Moved common IR stuff to ir-common.c
v4l: (951) Make saa7134-oss as a stand-alone module
v4l: (962) Added new saa7134 card (MSI TV@anywhere plus)
v4l: 975: apply saa7134-alsa fixes
Richard Purdie:
[ARM] 3149/1: SharpSL: Add Akita (SL-C1000) machine support
[ARM] 3154/1: SharpSL PM Driver updates
[ARM] 3158/1: SharpSL: Add PM device driver for the SL-C7x0 machines.
[ARM] 3159/1: SharpSL: Add PM device driver for the SL-Cx00 machines.
[ARM] 3160/1: SharpSL: Add driver for Akita specific GPIOs
w100fb: platform device conversion fixup
USB: OHCI lh7a404 platform device conversion fixup
Robin Holt:
mm: ZAP_BLOCK causes redundant work
Roger While:
prism54 : Remove extraneous udelay/register read
Rohit Seth:
mm: __alloc_pages cleanup
Roland Dreier:
[IB] srp: increase max_luns
[IB] srp: don't post receive if no send buf available
[IB] mthca: don't disable RDMA writes if no responder resources
IB/umad: make sure write()s have sufficient data
Roman Zippel:
m68k: convert thread flags to use bit fields
[NET]: Sanitize NET_SCHED protection in /net/sched/Kconfig
Russell King:
[ARM] Fix Footbridge-based machines
[ARM] Fix broken sl82c105 DMA prevention
[ARM] Restore apparant pointless change in arch/arm/kernel/smp.c
[MMC] mmci doesn't need asm/irq.h
[ARM] Ensure sl82c105 IDE interfaces are serialized when using DMA
[ARM] Use correct IO operations for Pleb
[ARM] Re-fix footbridge
[SERIAL] Fix Bug 4900: S3 resume oops with irattach - Thinkpad A21m
[ARM] Use kernel/power/Kconfig
[ARM] Initialise SA1111 core before SA1111 PCMCIA
[ARM] Fix arch-realview/system.h to use __io_address()
[ARM] Include asm/hardware.h instead of asm/arch/hardware.h
[ARM] compressed/head.S debugging defaults to asm/arch/debug-macro.S
[ARM] Add linux/compiler.h includes where required
[ARM] Move zone adjustment for SA1111 on SA11x0 platforms
[ARM] Use unsigned long not u32 in atomic_cmpxchg
[ARM] sa1111.c needs asm/sizes.h
[ARM] No need to include asm/proc-fns.h into asm/system.h
[DRIVER MODEL] Fix merge clashes with ARM ixp2000 / ixp4xx platforms
[ARM] Improve comment about ASSERT()s in vmlinux.lds.S
[ARM] Drivers should not make use of architecture private __ioremap
[ARM] __ioremap doesn't use 4th argument
[ARM] Fix some corner cases in new mm initialisation
[ARM] Fix get_user when passed a const pointer
smc91x: fix bank mismatch
[SERIAL] Fix status reporting with PL011 serial driver
[SERIAL] Remove unused variable in sa1100.c
Ryan Bradetich:
[PARISC] Make redirecting irq messages less noisy
[PARISC] Compile fixups for serial/mux.c
[PARISC] Define port->timeout to fix a long msleep in mux.c
Sean Young:
[MTD] maps: Replace dependency on non existing config option
Segher Boessenkool:
powerpc: Maple: request I/O resource.
Shaohua Li:
x86_64: Force correct address space size for MTRR on some 64bit Intel Xeons
Siddha, Suresh B:
x86_64: fix tss limit
x86_64: Unmap NULL during early bootup
x86-64/i386: Intel HT, Multi core detection fixes
x86_64: x86_64/i386 fix Intel cache detection code assumption about threads sharing
Stephen Hemminger:
[TCP]: More spelling fixes.
[TCP]: TCP highspeed build error
Stephen Rothwell:
powerpc: make iSeries use generic virtual irq mapping
powerpc: have only one definition of __irq_offset_value
powerpc: iSeries build fixes
ppc32: move some dma routines
powerpc: merge dma-mapping.h
Suresh Siddha:
x86-64/i386: Fix CPU model for family 6
Tejun Heo:
[BLOCK] elevator: run queue in elevator_switch
[BLOCK] cfq-iosched: cfq forced dispatching fix
[BLOCK] Implement elv_drain_elevator for improved switch error detection
[BLOCK] fix string handling in elv_iosched_store
[BLOCK] cfq-iosched: fix slice_left calculation
[BLOCK] noop-iosched: reimplementation of request dispatching
[BLOCK] elevator: elv_latter/former_request update
sil24: add missing ata_pad_free()
sil24: add constants
sil24: add sil24_restart_controller
sil24: use SRST for phy_reset
sil24: add ATAPI support
sil24: make error_intr less verbose
Thibaut VARENE:
pmac IDE: don't release empty interfaces
aec62xxx: remove all dead (#if0'd) code
Thomas Gleixner:
[JFFS2] Remove broken and useless debug code
Thomas Graf:
[IPV6]: Fix unnecessary GFP_ATOMIC allocation in fib6 dump
Tim Mann:
x86: fix cpu_khz with clock=pit
Toni Mueller:
sdladrv.c build fix
Tyler Trafford:
v4l: (958) Make cx25840 use firmware image named 'cx25840.fw'
Vitaly Bordug:
ppc32: add missing define for fs_enet Ethernet driver
Vivek Goyal:
drop "i386 kexec-on-panic: Don't shutdown the apics"
Vlad Drukker:
[NETFILTER] {ip,nf}_conntrack TCP: Accept SYN+PUSH like SYN
Yan Zheng:
[IPV6]: small fix for ipv6_dev_get_saddr(...)
Yasuyuki Kozakai:
[NETFILTER]: cleanup IPv6 Netfilter Kconfig
[NETFILTER]: fix type of sysctl variables in nf_conntrack_ipv6
[NETFILTER] nf_conntrack: fix possibility of infinite loop while evicting nf_ct_frag6_queue
[NETFILTER] fix leak of fragment queue at unloading nf_conntrack_ipv6
[IPV4,IPV6]: replace handmade list with hlist in IPv{4,6} reassembly
Yoichi Yuasa:
Add GT64111 PCI ID back
Zach Brown:
aio: remove kioctx from mm_struct
aio: replace locking comments with assert_spin_locked()
aio: don't ref kioctx after decref in put_ioctx
Zachary Amsden:
[BLOCK] elevator init fixes
[BLOCK] elevator init fixes #2
Zhu Yi:
ipw2200: fix error log offset calculation
On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
>There it is (or will soon be - the tar-ball and patches are still
>uploading, and mirroring can obviously take some time after that).
First breakage report, tvtime, blue screen no audio. Trying slightly
different .config for next build. My tuner (OR51132) seems to be
permanently selected in an xconfig screen. Dunno if thats good or bad
ATM.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
Hi,
i get this after plugged in dvb-t/Cinergy T2 with Kernel 2.6.15-git*/rc*.
usb 3-5.5: new high speed USB device using ehci_hcd and address 7
DVB: registering new adapter (TerraTec/qanu USB2.0 Highspeed DVB-T Receiver).
Unable to handle kernel paging request at virtual address 0483b400
printing eip:
f9060f01
*pde = 00000000
Oops: 0002 [#1]
SMP
Modules linked in: cinergyT2 dvb_core w83627hf hwmon_vid eeprom i2c_isa
snd_seq_midi snd_emu10k1_synth snd_emux_synth snd_seq_virmidi
snd_seq_midi_emul snd_pcm_oss snd_mixer_oss snd_seq_oss snd_seq_midi_event
snd_seq usbhid usb_storage ehci_hcd ohci_hcd ohci1394 ieee1394 snd_emu10k1
snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm snd_timer snd_ac97_bus
snd_page_alloc snd_util_mem snd_hwdep snd soundcore e1000 nvidia i2c_i801
intel_agp usbcore
CPU: 0
EIP: 0060:[<f9060f01>] Tainted: P VLI
EFLAGS: 00010206 (2.6.15-rc2)
EIP is at cinergyt2_register_rc+0xc8/0xde [cinergyT2]
eax: 6f635f32 ebx: ec07b3dc ecx: c1806f60 edx: 0000006f
esi: f6974800 edi: ec07b000 ebp: ec07b234 esp: f5a0eea8
ds: 007b es: 007b ss: 0068
Process modprobe (pid: 10806, threadinfo=f5a0e000 task=ec7b2030)
Stack: ec07b3dc 00000040 f90619c4 c197e0f0 f3c4f004 ec07b000 ec07b208 00000000
f906119c ec07b000 00000003 00000001 f6885200 f9062e80 ffffffed f8de8ce0
f8dcf0a0 f9062ea0 f6885214 00000000 c0243768 00000000 ec3d8400 c040c200
Call Trace:
[<f906119c>] cinergyt2_probe+0x19c/0x2f3 [cinergyT2]
[<f8dcf0a0>] usb_probe_interface+0x6f/0x9a [usbcore]
[<c0243768>] driver_probe_device+0x36/0xa0
[<c0243886>] __driver_attach+0x4d/0x4f
[<c0242f96>] bus_for_each_dev+0x42/0x57
[<c024389e>] driver_attach+0x16/0x1a
[<c0243839>] __driver_attach+0x0/0x4f
[<c024338f>] bus_add_driver+0x7b/0xb9
[<f8dcf198>] usb_register+0x51/0xb4 [usbcore]
[<f9021011>] cinergyt2_init+0x11/0x58 [cinergyT2]
[<c012c6cc>] kthread_stop_sem+0x70/0xa8
[<c0132d9a>] sys_init_module+0x154/0x1ea
[<c0102b33>] sysenter_past_esp+0x54/0x75
Code: 00 00 00 00 00 00 8d 87 34 04 00 00 e8 11 14 0c c7 31 d2 c7 46 18 02 00
10 00 c7 46 04 c0 1b 06 f9 89 5e 08 8b 04 95 48 16 06 f9 <f0> 0f ab 46 1c 83
c2 03 eb ef 83 c4 14 b8 f4 ff ff ff 5b 5e 5f
Thanks,
M. Geithe
On Sun, Nov 20, 2005 at 02:20:54PM +0100, Michael Geithe wrote:
> i get this after plugged in dvb-t/Cinergy T2 with Kernel 2.6.15-git*/rc*.
> usb 3-5.5: new high speed USB device using ehci_hcd and address 7
> DVB: registering new adapter (TerraTec/qanu USB2.0 Highspeed DVB-T Receiver).
> Unable to handle kernel paging request at virtual address 0483b400
> printing eip:
> f9060f01
> *pde = 00000000
> Oops: 0002 [#1]
> SMP
> Modules linked in: cinergyT2 dvb_core w83627hf hwmon_vid eeprom i2c_isa
> snd_seq_midi snd_emu10k1_synth snd_emux_synth snd_seq_virmidi
> snd_seq_midi_emul snd_pcm_oss snd_mixer_oss snd_seq_oss snd_seq_midi_event
> snd_seq usbhid usb_storage ehci_hcd ohci_hcd ohci1394 ieee1394 snd_emu10k1
> snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm snd_timer snd_ac97_bus
> snd_page_alloc snd_util_mem snd_hwdep snd soundcore e1000 nvidia i2c_i801
> intel_agp usbcore
> CPU: 0
> EIP: 0060:[<f9060f01>] Tainted: P VLI
^^^^^^^^^^
Can you reproduce it with clean kernel?
Hi,
On Sunday 20 November 2005 15:52, you wrote:
> Can you reproduce it with clean kernel?
the same with clean kernel.
usb 3-4.1: new high speed USB device using ehci_hcd and address 7
DVB: registering new adapter (TerraTec/qanu USB2.0 Highspeed DVB-T Receiver).
Unable to handle kernel paging request at virtual address 0452fc00
printing eip:
f8feaf01
*pde = 00000000
Oops: 0002 [#1]
SMP
Modules linked in: cinergyT2 dvb_core w83627hf hwmon_vid eeprom i2c_isa
snd_seq_midi snd_emu10k1_synth snd_emux_synth snd_seq_virmidi
snd_seq_midi_emul snd_pcm_oss snd_mixer_oss snd_seq_oss snd_seq_midi_event
snd_seq usbhid usb_storage ehci_hcd ohci_hcd ohci1394 ieee1394 snd_emu10k1
snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm snd_timer snd_ac97_bus
snd_page_alloc snd_util_mem snd_hwdep snd soundcore e1000 i2c_i801 intel_agp
usbcore
CPU: 1
EIP: 0060:[<f8feaf01>] Not tainted VLI
EFLAGS: 00010206 (2.6.15-rc2)
EIP is at cinergyt2_register_rc+0xc8/0xde [cinergyT2]
eax: 6f635f32 ebx: f7ac83dc ecx: c180ef60 edx: 0000006f
esi: f6669000 edi: f7ac8000 ebp: f7ac8234 esp: f6611ea8
ds: 007b es: 007b ss: 0068
Process modprobe (pid: 10497, threadinfo=f6611000 task=f7b79030)
Stack: f7ac83dc 00000040 f8feb9c4 c197e0f0 f7be5c04 f7ac8000 f7ac8208 00000000
f8feb19c f7ac8000 00000003 00000001 f508aa00 f8fece80 ffffffed f8de8ce0
f8dcf0a0 f8fecea0 f508aa14 00000000 c0243768 00000000 f7bd4400 c040c200
Call Trace:
[<f8feb19c>] cinergyt2_probe+0x19c/0x2f3 [cinergyT2]
[<f8dcf0a0>] usb_probe_interface+0x6f/0x9a [usbcore]
[<c0243768>] driver_probe_device+0x36/0xa0
[<c0243886>] __driver_attach+0x4d/0x4f
[<c0242f96>] bus_for_each_dev+0x42/0x57
[<c024389e>] driver_attach+0x16/0x1a
[<c0243839>] __driver_attach+0x0/0x4f
[<c024338f>] bus_add_driver+0x7b/0xb9
[<f8dcf198>] usb_register+0x51/0xb4 [usbcore]
[<f8fc1011>] cinergyt2_init+0x11/0x58 [cinergyT2]
[<c012c6cc>] kthread_stop_sem+0x70/0xa8
[<c0132d9a>] sys_init_module+0x154/0x1ea
[<c0102b33>] sysenter_past_esp+0x54/0x75
Code: 00 00 00 00 00 00 8d 87 34 04 00 00 e8 11 74 13 c7 31 d2 c7 46 18 02 00
10 00 c7 46 04 c0 bb fe f8 89 5e 08 8b 04 95 48 b6 fe f8 <f0> 0f ab 46 1c 83
c2 03 eb ef 83 c4 14 b8 f4 ff ff ff 5b 5e 5f
Thanks,
M. Geithe
-
On Sunday 20 November 2005 08:20, Michael Geithe wrote:
> Hi,
> i get this after plugged in dvb-t/Cinergy T2 with Kernel 2.6.15-git*/rc*.
>
Hm, is there one driver in drivers/media that I left working? Please
try the patch below.
--
Dmitry
Subjtect: Fix an OOPS is CinergyT2
Fix an OOPS is CinergyT2 driver when registering IR remote
Signed-off-by: Dmitry Torokhov <[email protected]>
---
drivers/media/dvb/cinergyT2/cinergyT2.c | 2 +-
1 files changed, 1 insertion(+), 1 deletion(-)
Index: work/drivers/media/dvb/cinergyT2/cinergyT2.c
===================================================================
--- work.orig/drivers/media/dvb/cinergyT2/cinergyT2.c
+++ work/drivers/media/dvb/cinergyT2/cinergyT2.c
@@ -772,7 +772,7 @@ static int cinergyt2_register_rc(struct
input_dev->name = DRIVER_NAME " remote control";
input_dev->phys = cinergyt2->phys;
input_dev->evbit[0] = BIT(EV_KEY) | BIT(EV_REP);
- for (i = 0; ARRAY_SIZE(rc_keys); i += 3)
+ for (i = 0; i < ARRAY_SIZE(rc_keys); i += 3)
set_bit(rc_keys[i + 2], input_dev->keybit);
input_dev->keycodesize = 0;
input_dev->keycodemax = 0;
On Sunday 20 November 2005 00:18, Gene Heskett wrote:
>On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
>>There it is (or will soon be - the tar-ball and patches are still
>>uploading, and mirroring can obviously take some time after that).
>
>First breakage report, tvtime, blue screen no audio. Trying slightly
>different .config for next build. My tuner (OR51132) seems to be
>permanently selected in an xconfig screen. Dunno if thats good or bad
>ATM.
Update, I may be sticking my finger in the dike and hollering wolf or
however that old saw goes. I've now rebooted to 3 kernels where tvtime
was known to work, but it doesn't. Turning off the signal detection
shows that all I'm getting is some sort of digital noise. So its time
to drag in another receiver and see if its the DISH convertor or my
pcHDTV-3000.
More later when I've done that.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Sunday 20 November 2005 12:02, Gene Heskett wrote:
>On Sunday 20 November 2005 00:18, Gene Heskett wrote:
>>On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
>>>There it is (or will soon be - the tar-ball and patches are still
>>>uploading, and mirroring can obviously take some time after that).
>>
>>First breakage report, tvtime, blue screen no audio. Trying slightly
>>different .config for next build. My tuner (OR51132) seems to be
>>permanently selected in an xconfig screen. Dunno if thats good or bad
>>ATM.
>
>Update, I may be sticking my finger in the dike and hollering wolf or
>however that old saw goes. I've now rebooted to 3 kernels where tvtime
>was known to work, but it doesn't. Turning off the signal detection
>shows that all I'm getting is some sort of digital noise. So its time
>to drag in another receiver and see if its the DISH convertor or my
>pcHDTV-3000.
>
>More later when I've done that.
Ok, the Dish is working. My tv card, a pcHDTV-3000, worked less than 5
minutes before I rebooted to 2.6.15-rc2 from 2.6.14.2, and now it
doesn't even when booted back to 2.6.14.2. These were not powerdown
reboots so that may have a bearing on this.
That leaves two possibilities.
1) The card has died (doubtfull)
2) something in the i2c probing for 2.6.15-rc2 put it into some odd
mode, from which tvtime seems unable to recover from.
I'd turned on the nxt-200x stuff and have now turned it off, which was
the only diff in an lsmod listing between the boots. If that module
does something to the card as it inits, how can I undo that?
Comments anyone?
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
Hi,
On Sunday 20 November 2005 17:13, you wrote:
> On Sunday 20 November 2005 08:20, Michael Geithe wrote:
> > Hi,
> > i get this after plugged in dvb-t/Cinergy T2 with Kernel 2.6.15-git*/rc*.
>
> Hm, is there one driver in drivers/media that I left working? Please
> try the patch below.
> Subjtect: Fix an OOPS is CinergyT2
> Fix an OOPS is CinergyT2 driver when registering IR remote
> Signed-off-by: Dmitry Torokhov <[email protected]>
> drivers/media/dvb/cinergyT2/cinergyT2.c | 2 +-
> 1 files changed, 1 insertion(+), 1 deletion(-)
> Index: work/drivers/media/dvb/cinergyT2/cinergyT2.c
> ===================================================================
> --- work.orig/drivers/media/dvb/cinergyT2/cinergyT2.c
> +++ work/drivers/media/dvb/cinergyT2/cinergyT2.c
> @@ -772,7 +772,7 @@ static int cinergyt2_register_rc(struct
> input_dev->name = DRIVER_NAME " remote control";
> input_dev->phys = cinergyt2->phys;
> input_dev->evbit[0] = BIT(EV_KEY) | BIT(EV_REP);
> - for (i = 0; ARRAY_SIZE(rc_keys); i += 3)
> + for (i = 0; i < ARRAY_SIZE(rc_keys); i += 3)
> set_bit(rc_keys[i + 2], input_dev->keybit);
> input_dev->keycodesize = 0;
> input_dev->keycodemax = 0;
Thanks, it works. :-)
M. Geithe
-
On Sunday 20 November 2005 12:37, Gene Heskett wrote:
>On Sunday 20 November 2005 12:02, Gene Heskett wrote:
>>On Sunday 20 November 2005 00:18, Gene Heskett wrote:
>>>On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
>>>>There it is (or will soon be - the tar-ball and patches are still
>>>>uploading, and mirroring can obviously take some time after that).
>>>
>>>First breakage report, tvtime, blue screen no audio. Trying slightly
>>>different .config for next build. My tuner (OR51132) seems to be
>>>permanently selected in an xconfig screen. Dunno if thats good or
>>> bad ATM.
>>
>>Update, I may be sticking my finger in the dike and hollering wolf or
>>however that old saw goes. I've now rebooted to 3 kernels where
>> tvtime was known to work, but it doesn't. Turning off the signal
>> detection shows that all I'm getting is some sort of digital noise.
>> So its time to drag in another receiver and see if its the DISH
>> convertor or my pcHDTV-3000.
>>
>>More later when I've done that.
>
>Ok, the Dish is working. My tv card, a pcHDTV-3000, worked less than 5
>minutes before I rebooted to 2.6.15-rc2 from 2.6.14.2, and now it
>doesn't even when booted back to 2.6.14.2. These were not powerdown
>reboots so that may have a bearing on this.
>
>That leaves two possibilities.
>1) The card has died (doubtfull)
>2) something in the i2c probing for 2.6.15-rc2 put it into some odd
>mode, from which tvtime seems unable to recover from.
>
>I'd turned on the nxt-200x stuff and have now turned it off, which was
>the only diff in an lsmod listing between the boots. If that module
>does something to the card as it inits, how can I undo that?
See above, a powerdown for 20 seconds, followed by a reboot to
2.6.14.2, and tvtime works just fine.
Here is the lsmod output for this boot:
[root@coyote linux-2.6.15-rc2]# lsmod
Module Size Used by
cx88_dvb 7428 0
cx8800 27276 2
v4l1_compat 13188 1 cx8800
v4l2_common 4864 1 cx8800
cx8802 9092 1 cx88_dvb
cx88xx 53024 3 cx88_dvb,cx8800,cx8802
i2c_algo_bit 8456 1 cx88xx
ir_common 7556 1 cx88xx
btcx_risc 3976 3 cx8800,cx8802,cx88xx
tuner 37032 0
tveeprom 12304 1 cx88xx
videodev 7296 4 cx8800,cx88xx
mt352 6148 1 cx88_dvb
or51132 9220 1 cx88_dvb
video_buf_dvb 4612 1 cx88_dvb
dvb_core 75304 1 video_buf_dvb
video_buf 17540 5
cx88_dvb,cx8800,cx8802,cx88xx,video_buf_dvb
lgdt330x 7580 1 cx88_dvb
cx22702 5892 1 cx88_dvb
dvb_pll 7812 3 cx88_dvb,or51132,cx22702
radeon 101888 1
drm 64020 2 radeon
nvidia_agp 6044 1
agpgart 29000 2 drm,nvidia_agp
w83627hf 24592 0
hwmon_vid 2176 1 w83627hf
i2c_isa 3584 1 w83627hf
i2c_nforce2 5760 0
i2c_core 17808 12
cx88_dvb,cx88xx,i2c_algo_bit,tuner,tveeprom,mt352,or51132,lgdt330x,cx227
02,w83627hf,i2c_isa,i2c_nforce2
Pardon the wrapping of the last line in mid-word.
Now, the difference that I can see is that I cannot build a 2.6.15-rc2
without the nxt200x stuffs without depmod spitting out a whole menu of
squawks about dvb_stuffs.
This card apparently doesn't need the nxt2000x stuffs, so how do I go
about building a working 2.6.15-rc2 kernel without it?
.configs available on request.
>Comments anyone?
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Sun, 2005-11-20 at 00:18 -0500, Gene Heskett wrote:
> On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
> >There it is (or will soon be - the tar-ball and patches are still
> >uploading, and mirroring can obviously take some time after that).
>
> First breakage report, tvtime, blue screen no audio. Trying slightly
> different .config for next build. My tuner (OR51132) seems to be
> permanently selected in an xconfig screen. Dunno if thats good or bad
> ATM.
if it needs to be loaded with a parameter you will need to build it as a
module.. my saa7134 chip needs card=9.
i am experiencing same problems with saa7134, no video, however i do get
audio.
this is a way to (incorrectly according to v4l devs) "fix" it:
drivers/media/video/video-buf.c
change line 1233 to this:
vma->vm_flags |= VM_DONTEXPAND;
>
On Sunday 20 November 2005 17:36, Kasper Sandberg wrote:
>On Sun, 2005-11-20 at 00:18 -0500, Gene Heskett wrote:
>> On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
>> >There it is (or will soon be - the tar-ball and patches are still
>> >uploading, and mirroring can obviously take some time after that).
>>
>> First breakage report, tvtime, blue screen no audio. Trying slightly
>> different .config for next build. My tuner (OR51132) seems to be
>> permanently selected in an xconfig screen. Dunno if thats good or
>> bad ATM.
>
>if it needs to be loaded with a parameter you will need to build it as
> a module.. my saa7134 chip needs card=9.
Its never needed an argument before.
>i am experiencing same problems with saa7134, no video, however i do
> get audio.
I wasn't, total digital gibberish on screen.
A full powerdown reboot to 2.6.14.2 fixed it.
>this is a way to (incorrectly according to v4l devs) "fix" it:
>drivers/media/video/video-buf.c
>change line 1233 to this:
> vma->vm_flags |= VM_DONTEXPAND;
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Sun, Nov 20, 2005 Gene Heskett wrote:
> First breakage report, tvtime, blue screen no audio. Trying slightly
> different .config for next build.
Probably v4l breakage due to VM changes. For me xawtv overlay works,
grabdisplay doesn't (with different cards). This was reported before.
> My tuner (OR51132) seems to be
> permanently selected in an xconfig screen. Dunno if thats good or bad
> ATM.
Works for me in menuconfig. You probably have
CONFIG_VIDEO_SAA7134_DVB_ALL_FRONTENDS selected?
Johannes
PS: don't trim Cc: on lkml
On Sunday 20 November 2005 19:15, Johannes Stezenbach wrote:
>On Sun, Nov 20, 2005 Gene Heskett wrote:
>> First breakage report, tvtime, blue screen no audio. Trying slightly
>> different .config for next build.
>
>Probably v4l breakage due to VM changes. For me xawtv overlay works,
>grabdisplay doesn't (with different cards). This was reported before.
>
>> My tuner (OR51132) seems to be
>> permanently selected in an xconfig screen. Dunno if thats good or
>> bad ATM.
>
>Works for me in menuconfig. You probably have
>CONFIG_VIDEO_SAA7134_DVB_ALL_FRONTENDS selected?
Nope. its off. Or lets put it this way:
# grep SAA7134 .config
# CONFIG_VIDEO_SAA7134 is not set
The longer string above doesn't exist in my .config, made from the
2.6.14.2 .config with a make oldconfig. Is this a bug in the patchfile?
>Johannes
>
>PS: don't trim Cc: on lkml
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Sun, Nov 20, 2005 at 08:07:44PM -0500, Gene Heskett wrote:
> On Sunday 20 November 2005 19:15, Johannes Stezenbach wrote:
> >On Sun, Nov 20, 2005 Gene Heskett wrote:
> >> First breakage report, tvtime, blue screen no audio. Trying slightly
> >> different .config for next build.
> >
> >Probably v4l breakage due to VM changes. For me xawtv overlay works,
> >grabdisplay doesn't (with different cards). This was reported before.
> >
> >> My tuner (OR51132) seems to be
> >> permanently selected in an xconfig screen. Dunno if thats good or
> >> bad ATM.
> >
> >Works for me in menuconfig. You probably have
> >CONFIG_VIDEO_SAA7134_DVB_ALL_FRONTENDS selected?
>
> Nope. its off. Or lets put it this way:
> # grep SAA7134 .config
> # CONFIG_VIDEO_SAA7134 is not set
>
> The longer string above doesn't exist in my .config, made from the
> 2.6.14.2 .config with a make oldconfig. Is this a bug in the patchfile?
Maybe you should've included the actual hardware you use in your initial
posting. How about CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS, or
$ grep DVB_ALL_FRONTENDS .config
Johannes
On Sunday 20 November 2005 20:32, Johannes Stezenbach wrote:
>grep DVB_ALL_FRONTENDS .config
[root@coyote linux-2.6.15-rc2]# grep DVB_ALL_FRONTENDS .config
# CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS is not set
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Sun, 2005-11-20 at 18:58 -0500, Gene Heskett wrote:
> On Sunday 20 November 2005 17:36, Kasper Sandberg wrote:
> >On Sun, 2005-11-20 at 00:18 -0500, Gene Heskett wrote:
> >> On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
> >> >There it is (or will soon be - the tar-ball and patches are still
> >> >uploading, and mirroring can obviously take some time after that).
> >>
> >> First breakage report, tvtime, blue screen no audio. Trying slightly
> >> different .config for next build. My tuner (OR51132) seems to be
> >> permanently selected in an xconfig screen. Dunno if thats good or
> >> bad ATM.
> >
> >if it needs to be loaded with a parameter you will need to build it as
> > a module.. my saa7134 chip needs card=9.
>
> Its never needed an argument before.
then you have a good card, mine is a cheap cheap cheap one which
apparently doesent have the nessecary embedded info to do proper
autodetection, so i gotta manually specify which card i have.
>
> >i am experiencing same problems with saa7134, no video, however i do
> > get audio.
>
> I wasn't, total digital gibberish on screen.
>
> A full powerdown reboot to 2.6.14.2 fixed it.
>
> >this is a way to (incorrectly according to v4l devs) "fix" it:
> >drivers/media/video/video-buf.c
> >change line 1233 to this:
> > vma->vm_flags |= VM_DONTEXPAND;
>
On Monday 21 November 2005 16:17, Kasper Sandberg wrote:
>On Sun, 2005-11-20 at 18:58 -0500, Gene Heskett wrote:
>> On Sunday 20 November 2005 17:36, Kasper Sandberg wrote:
>> >On Sun, 2005-11-20 at 00:18 -0500, Gene Heskett wrote:
>> >> On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
>> >> >There it is (or will soon be - the tar-ball and patches are still
>> >> >uploading, and mirroring can obviously take some time after
>> >> > that).
>> >>
>> >> First breakage report, tvtime, blue screen no audio. Trying
>> >> slightly different .config for next build. My tuner (OR51132)
>> >> seems to be permanently selected in an xconfig screen. Dunno if
>> >> thats good or bad ATM.
>> >
>> >if it needs to be loaded with a parameter you will need to build it
>> > as a module.. my saa7134 chip needs card=9.
>>
>> Its never needed an argument before.
>
>then you have a good card, mine is a cheap cheap cheap one which
>apparently doesent have the nessecary embedded info to do proper
>autodetection, so i gotta manually specify which card i have.
Its a pcHDTV-3000
>> >i am experiencing same problems with saa7134, no video, however i do
>> > get audio.
>>
>> I wasn't, total digital gibberish on screen.
>>
>> A full powerdown reboot to 2.6.14.2 fixed it.
>>
>> >this is a way to (incorrectly according to v4l devs) "fix" it:
>> >drivers/media/video/video-buf.c
>> >change line 1233 to this:
>> > vma->vm_flags |= VM_DONTEXPAND;
And this will actually do what? Elaborate please.
I think my problem is that somehow, the dvb stuff now has a dependency
on the nxt200x thing, whatever it is, as if I force a kernel build
without it, then I get depmod problems at the end of the build telling
me:
WARNING:/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb
.ko needs unknown symbol nxt200x_attach.
This was never a requirement up thru 2.6.14.2, and I didn't try 15-rc1
as the -rc1's are usually a disaster of some kind, and this is also my
main box. I don't need a filesystem muckup again.
So whatever did that seems in error to me, and should be backed out in
favor of some other method to bring in the nxt200x bearing cards.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
Linus Torvalds wrote:
>
> There it is (or will soon be - the tar-ball and patches are still
> uploading, and mirroring can obviously take some time after that).
>
> It's slightly bigger than I'd like, but that's partly because I had (once
> more) missed a merge that was actually sent in well before the -rc1
> cut-off, so the x86-64 merge is from there and was delayed due to yours
> truly, not Andi Kleen.
>
> Apart from the x86-64 merge and various fixups, I've let MIPS, PARISC and
> PowerPC merge up some more.
> <http://www.tux.org/lkml/>
>
I get the following build error:
x86_x64-linux-gcc -Wp,-MD,arch/x86_64/ia32/.ia32_ioctl.o.d -nostdinc
-isystem /usr/lib/gcc/i486-linux-gnu/4.0.3/include -D__KERNEL__
-Iinclude -include include/linux/autoconf.h -Wall -Wundef
-Wstrict-prototypes -Wno-trigraphs -fno-strict-aliasing -fno-common
-ffreestanding -O2 -fomit-frame-pointer -march=k8 -mno-red-zone
-mcmodel=kernel -pipe -fno-reorder-blocks -Wno-sign-compare
-fno-asynchronous-unwind-tables -funit-at-a-time -mno-sse -mno-mmx
-mno-sse2 -mno-3dnow -Wdeclaration-after-statement -Wno-pointer-sign
-Ifs/ -DKBUILD_BASENAME=ia32_ioctl -DKBUILD_MODNAME=ia32_ioctl -c -o
arch/x86_64/ia32/ia32_ioctl.o arch/x86_64/ia32/ia32_ioctl.c
In file included from include/linux/ext3_jbd.h:20,
from fs/compat_ioctl.c:52,
from arch/x86_64/ia32/ia32_ioctl.c:14:
include/linux/ext3_fs.h: In function 'ext3_raw_inode':
include/linux/ext3_fs.h:696: error: dereferencing pointer to incomplete type
include/linux/ext3_fs.h: At top level:
include/linux/ext3_fs.h:734: error: syntax error before '*' token
include/linux/ext3_fs.h:734: warning: function declaration isn't a prototype
include/linux/ext3_fs.h:735: error: syntax error before '*' token
include/linux/ext3_fs.h:736: warning: function declaration isn't a prototype
=== full build log and .config attached
$ gcc --version
gcc (GCC) 4.0.3 20051111 (prerelease) (Debian 4.0.2-4)
Copyright (C) 2005 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
The last successfully build and run version was linux-2.6.14.2 using
this same build environment. I can try linux-2.6.15-rc1 but there's
been many x86_64 changes. Let me know how I can help. I tried with and
without ext3 and/or ext2.
--
Jeffrey Hundstad
Jeffrey Hundstad <[email protected]> wrote:
>
> from fs/compat_ioctl.c:52,
> from arch/x86_64/ia32/ia32_ioctl.c:14:
> include/linux/ext3_fs.h: In function 'ext3_raw_inode':
> include/linux/ext3_fs.h:696: error: dereferencing pointer to incomplete type
This might help?
--- 25/include/linux/ext3_fs.h~ext3_fsh-needs-buffer_headh Tue Nov 22 14:59:17 2005
+++ 25-akpm/include/linux/ext3_fs.h Tue Nov 22 14:59:27 2005
@@ -19,7 +19,7 @@
#include <linux/types.h>
#include <linux/ext3_fs_i.h>
#include <linux/ext3_fs_sb.h>
-
+#include <linux/buffer_head.h>
struct statfs;
_
On Tue, 22 Nov 2005, Andrew Morton wrote:
> Jeffrey Hundstad <[email protected]> wrote:
> >
> > from fs/compat_ioctl.c:52,
> > from arch/x86_64/ia32/ia32_ioctl.c:14:
> > include/linux/ext3_fs.h: In function 'ext3_raw_inode':
> > include/linux/ext3_fs.h:696: error: dereferencing pointer to incomplete type
>
> This might help?
Why does it happen at all, though? And why aren't more people reporting
this? Something strange going on.
Linus
Linus Torvalds <[email protected]> wrote:
>
>
>
> On Tue, 22 Nov 2005, Andrew Morton wrote:
>
> > Jeffrey Hundstad <[email protected]> wrote:
> > >
> > > from fs/compat_ioctl.c:52,
> > > from arch/x86_64/ia32/ia32_ioctl.c:14:
> > > include/linux/ext3_fs.h: In function 'ext3_raw_inode':
> > > include/linux/ext3_fs.h:696: error: dereferencing pointer to incomplete type
> >
> > This might help?
The patch didn't help.
>
> Why does it happen at all, though?
davem recently merged a patch which adds ext3 ioctls to fs/compat_ioctl.c.
That required inclusion of ext3 and jbd header files. Those files explode
unpleasantly when CONFIG_JBD=n.
No trivial fix was apparent - perhaps we should disable the compat wrappers
if CONFIG_EXT3=n and/or CONFIG_JBD=n.
> And why aren't more people reporting
> this? Something strange going on.
Most people use ext3.
On Tue, 2005-11-22 at 16:50 -0800, Linus Torvalds wrote:
>
> On Tue, 22 Nov 2005, Andrew Morton wrote:
>
> > Jeffrey Hundstad <[email protected]> wrote:
> > >
> > > from fs/compat_ioctl.c:52,
> > > from arch/x86_64/ia32/ia32_ioctl.c:14:
> > > include/linux/ext3_fs.h: In function 'ext3_raw_inode':
> > > include/linux/ext3_fs.h:696: error: dereferencing pointer to incomplete type
> >
> > This might help?
>
> Why does it happen at all, though? And why aren't more people reporting
> this? Something strange going on.
I get exactly the same build error on my AMD64 system. I didn't report
it because I figured that someone else had already seen it.
--
Zan Lynx <[email protected]>
On Tue, Nov 22, 2005 at 05:05:07PM -0800, Andrew Morton wrote:
> Linus Torvalds <[email protected]> wrote:
> >
> >
> >
> > On Tue, 22 Nov 2005, Andrew Morton wrote:
> >
> > > Jeffrey Hundstad <[email protected]> wrote:
> > > >
> > > > from fs/compat_ioctl.c:52,
> > > > from arch/x86_64/ia32/ia32_ioctl.c:14:
> > > > include/linux/ext3_fs.h: In function 'ext3_raw_inode':
> > > > include/linux/ext3_fs.h:696: error: dereferencing pointer to incomplete type
> > >
> > > This might help?
>
> The patch didn't help.
>
> >
> > Why does it happen at all, though?
>
> davem recently merged a patch which adds ext3 ioctls to fs/compat_ioctl.c.
> That required inclusion of ext3 and jbd header files. Those files explode
> unpleasantly when CONFIG_JBD=n.
Or use ->compat_ioctl and do it in fs/ext3
>
> No trivial fix was apparent - perhaps we should disable the compat wrappers
> if CONFIG_EXT3=n and/or CONFIG_JBD=n.
That's already done for a lot of other wrappers, so would be fine too
-Andi
On Tue, 22 Nov 2005, Andrew Morton wrote:
> >
> > Why does it happen at all, though?
>
> davem recently merged a patch which adds ext3 ioctls to fs/compat_ioctl.c.
> That required inclusion of ext3 and jbd header files. Those files explode
> unpleasantly when CONFIG_JBD=n.
Oh. How about just making jbd.h do the rigt thing, and not care about the
configuration?
If we include jbd.h, we want the jbd data structures. There's never any
reason to care whether jbd is enabled or not afaik.
Ie maybe just something like this?
(Untested, obviously. I'm just assuming that anything that actually
_needs_ the jbd functionality should have made sure that jdb is enabled.)
Zan, Jeffrey?
Linus
---
diff --git a/include/linux/jbd.h b/include/linux/jbd.h
index aa56172..dcde7ad 100644
--- a/include/linux/jbd.h
+++ b/include/linux/jbd.h
@@ -16,8 +16,6 @@
#ifndef _LINUX_JBD_H
#define _LINUX_JBD_H
-#if defined(CONFIG_JBD) || defined(CONFIG_JBD_MODULE) || !defined(__KERNEL__)
-
/* Allow this file to be included directly into e2fsprogs */
#ifndef __KERNEL__
#include "jfs_compat.h"
@@ -1083,19 +1081,4 @@ extern int jbd_blocks_per_page(struct in
#endif /* __KERNEL__ */
-#endif /* CONFIG_JBD || CONFIG_JBD_MODULE || !__KERNEL__ */
-
-/*
- * Compatibility no-ops which allow the kernel to compile without CONFIG_JBD
- * go here.
- */
-
-#if defined(__KERNEL__) && !(defined(CONFIG_JBD) || defined(CONFIG_JBD_MODULE))
-
-#define J_ASSERT(expr) do {} while (0)
-#define J_ASSERT_BH(bh, expr) do {} while (0)
-#define buffer_jbd(bh) 0
-#define journal_buffer_journal_lru(bh) 0
-
-#endif /* defined(__KERNEL__) && !defined(CONFIG_JBD) */
#endif /* _LINUX_JBD_H */
On Tue, 23 Nov 2005, Andi Kleen wrote:
>
> >
> > No trivial fix was apparent - perhaps we should disable the compat wrappers
> > if CONFIG_EXT3=n and/or CONFIG_JBD=n.
>
> That's already done for a lot of other wrappers, so would be fine too
That may be the right thing, but looking at compat-ioctl.c I don't see
anything that really depends on ext3, it just wants to have the data
structure definitions in _case_ ext3 migth be enabled. Or did I miss
anything?
In general, I don't like code that depends on a module having been marked
as a module. What if you compile the kernel and then decide later that you
need the jbd/ext3 modules, so you compile and install those on an already
running kernel?
So almost all "#ifdef CONFIG_xyzzy_MODULE" usages tend to be fundamentally
buggy: they expect all modules to come pre-configured, which may be ok for
a distro kernel, but it's a bit against the whole point of being a module,
isn't it?
Linus
Gene Heskett wrote:
>WARNING:/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb
>.ko needs unknown symbol nxt200x_attach.
>
Gene has sent me a copy of his .config ... here are the relevant lines:
[snip]
CONFIG_VIDEO_CX88=m
CONFIG_VIDEO_CX88_DVB=m
# CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS is not set
# CONFIG_VIDEO_CX88_DVB_MT352 is not set
CONFIG_VIDEO_CX88_DVB_OR51132=m
# CONFIG_VIDEO_CX88_DVB_CX22702 is not set
# CONFIG_VIDEO_CX88_DVB_LGDT330X is not set
# CONFIG_VIDEO_CX88_DVB_NXT200X is not set
[snip]
CONFIG_DVB_NXT2002=m
# CONFIG_DVB_NXT200X is not set
# CONFIG_DVB_OR51211 is not set
CONFIG_DVB_OR51132=m
# CONFIG_DVB_BCM3510 is not set
CONFIG_DVB_LGDT330X=m
[snip]
A configuration like this should have compiled cx88-dvb without any
references to nxt200x at all.
Gene, do you have v4l-kernel cvs installed on top of kernel 2.6.15-rc2?
Unless I'm missing something, it seems that this is the only way that
you could have nxt200x support in cx88-dvb without having built the
nxt200x module itself. The v4l-kernel cvs build environment has nxt200x
enabled by default when building against kernels 2.6.15 and later.
If this is true, then you can either disable nxt200x in
v4l-kernel/v4l/Makefile, or re-build the cvs modules using the dvb + v4l
merged-tree build environment.
http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS
This would, in effect, give the same result as if you had selected
CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS (selected by default in Kconfig)
...or maybe you initially started a kernel build with the nxt200x module
selected, then stopped the compile, unselected nxt200x, and continued
the build process without cleaning in between, and after cx88-dvb had
already been built? Hmm... I guess that might be a bit far-fetched...
but to be sure, you can wipe out your kernel tree and rebuild it again
from scratch using the same .config .
If my assumption above is incorrect, then this will need to be looked
into a bit deeper. Please, let me know.
Regards,
Michael Krufky
Linus Torvalds <[email protected]> wrote:
>
> On Tue, 22 Nov 2005, Andrew Morton wrote:
> > >
> > > Why does it happen at all, though?
> >
> > davem recently merged a patch which adds ext3 ioctls to fs/compat_ioctl.c.
> > That required inclusion of ext3 and jbd header files. Those files explode
> > unpleasantly when CONFIG_JBD=n.
>
> Oh. How about just making jbd.h do the rigt thing, and not care about the
> configuration?
>
> If we include jbd.h, we want the jbd data structures. There's never any
> reason to care whether jbd is enabled or not afaik.
Yes, that would be the right thing to do.
> ...
> -#if defined(__KERNEL__) && !(defined(CONFIG_JBD) || defined(CONFIG_JBD_MODULE))
> -
> -#define J_ASSERT(expr) do {} while (0)
> -#define J_ASSERT_BH(bh, expr) do {} while (0)
> -#define buffer_jbd(bh) 0
> -#define journal_buffer_journal_lru(bh) 0
> -
> -#endif /* defined(__KERNEL__) && !defined(CONFIG_JBD) */
I guess the core kernel once needed these, but it doesn't (and shouldn't) now.
From: Andi Kleen <[email protected]>
Date: 23 Nov 2005 02:32:44 +0100,Wed, 23 Nov 2005 02:32:44 +0100
> On Tue, Nov 22, 2005 at 05:05:07PM -0800, Andrew Morton wrote:
> > davem recently merged a patch which adds ext3 ioctls to fs/compat_ioctl.c.
> > That required inclusion of ext3 and jbd header files. Those files explode
> > unpleasantly when CONFIG_JBD=n.
>
> Or use ->compat_ioctl and do it in fs/ext3
And reiserfs and ... that's why I didn't do it using ->compat_ioctl(),
several other filesystems make use of some of the ext{2,3} ioctls.
I think the ifdef's are a possible solution for now. But it's very
silly, as we just want some of the ioctl and type definitions.
Perhaps ext3_fs.h can be arranged to still export the ioctl types even
when CONFIG_JBD is disabled?
Linus Torvalds wrote:
>On Tue, 22 Nov 2005, Andrew Morton wrote:
>
>
>>>Why does it happen at all, though?
>>>
>>>
>>davem recently merged a patch which adds ext3 ioctls to fs/compat_ioctl.c.
>>That required inclusion of ext3 and jbd header files. Those files explode
>>unpleasantly when CONFIG_JBD=n.
>>
>>
>
>Oh. How about just making jbd.h do the rigt thing, and not care about the
>configuration?
>
>If we include jbd.h, we want the jbd data structures. There's never any
>reason to care whether jbd is enabled or not afaik.
>
>Ie maybe just something like this?
>
>(Untested, obviously. I'm just assuming that anything that actually
>_needs_ the jbd functionality should have made sure that jdb is enabled.)
>
>Zan, Jeffrey?
>
> Linus
>---
>diff --git a/include/linux/jbd.h b/include/linux/jbd.h
>index aa56172..dcde7ad 100644
>--- a/include/linux/jbd.h
>+++ b/include/linux/jbd.h
>@@ -16,8 +16,6 @@
> #ifndef _LINUX_JBD_H
> #define _LINUX_JBD_H
>
>-#if defined(CONFIG_JBD) || defined(CONFIG_JBD_MODULE) || !defined(__KERNEL__)
>-
> /* Allow this file to be included directly into e2fsprogs */
> #ifndef __KERNEL__
> #include "jfs_compat.h"
>@@ -1083,19 +1081,4 @@ extern int jbd_blocks_per_page(struct in
>
> #endif /* __KERNEL__ */
>
>-#endif /* CONFIG_JBD || CONFIG_JBD_MODULE || !__KERNEL__ */
>-
>-/*
>- * Compatibility no-ops which allow the kernel to compile without CONFIG_JBD
>- * go here.
>- */
>-
>-#if defined(__KERNEL__) && !(defined(CONFIG_JBD) || defined(CONFIG_JBD_MODULE))
>-
>-#define J_ASSERT(expr) do {} while (0)
>-#define J_ASSERT_BH(bh, expr) do {} while (0)
>-#define buffer_jbd(bh) 0
>-#define journal_buffer_journal_lru(bh) 0
>-
>-#endif /* defined(__KERNEL__) && !defined(CONFIG_JBD) */
> #endif /* _LINUX_JBD_H */
>
>
This one compiles and boots just fine. I was also able to loop mount an
ext2 filesystem.
Thanks for all the effort!
BTW: Since I have your ear, this same version DOES seem to have some
other bug as well. I did a "make distclean" and the "rm -f" of all he
object files hung forever in "D" state. I'm using XFS on IDE disks.
I'm using the same config as was posted before. I didn't get anything
in an log files that would indicate a problem. Has this been reported?
If not, what can I do to make a meaningful report?
--
Jeffrey Hundstad
On Tue, 22 Nov 2005, Jeffrey Hundstad wrote:
>
> This one compiles and boots just fine. I was also able to loop mount an
> ext2 filesystem.
Ok. Andrew added it to his queue, I guess I'll get my patch back that way ;)
> BTW: Since I have your ear, this same version DOES seem to have some
> other bug as well. I did a "make distclean" and the "rm -f" of all he
> object files hung forever in "D" state. I'm using XFS on IDE disks.
> I'm using the same config as was posted before. I didn't get anything
> in an log files that would indicate a problem. Has this been reported?
> If not, what can I do to make a meaningful report?
I don't recognize those symptoms, so more info would be nice.
For example, it would be good to know where the threads are that are
waiting uninterruptibly. You should be able to get that info with Sysrq
'T' (or with the old "ctrl + ScrollLock" thing).
That should tell us if they're hung on a semaphore or waiting for disk IO
to complete or what...
Linus
On Tuesday 22 November 2005 20:56, Michael Krufky wrote:
>Gene Heskett wrote:
>>WARNING:/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-d
>>vb .ko needs unknown symbol nxt200x_attach.
>
>Gene has sent me a copy of his .config ... here are the relevant lines:
>
>[snip]
>
>CONFIG_VIDEO_CX88=m
>CONFIG_VIDEO_CX88_DVB=m
># CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS is not set
># CONFIG_VIDEO_CX88_DVB_MT352 is not set
>CONFIG_VIDEO_CX88_DVB_OR51132=m
># CONFIG_VIDEO_CX88_DVB_CX22702 is not set
># CONFIG_VIDEO_CX88_DVB_LGDT330X is not set
># CONFIG_VIDEO_CX88_DVB_NXT200X is not set
>
>[snip]
>
>CONFIG_DVB_NXT2002=m
># CONFIG_DVB_NXT200X is not set
># CONFIG_DVB_OR51211 is not set
>CONFIG_DVB_OR51132=m
># CONFIG_DVB_BCM3510 is not set
>CONFIG_DVB_LGDT330X=m
>
>[snip]
>
>A configuration like this should have compiled cx88-dvb without any
>references to nxt200x at all.
>
>Gene, do you have v4l-kernel cvs installed on top of kernel 2.6.15-rc2?
No, not since back at about 2.6.4 or so when I needed the ieee1394
stuff. This is a 2.6.14, with the 2.6.15-rc2 patch applied straight
out of the box.
>Unless I'm missing something, it seems that this is the only way that
>you could have nxt200x support in cx88-dvb without having built the
>nxt200x module itself. The v4l-kernel cvs build environment has
> nxt200x enabled by default when building against kernels 2.6.15 and
> later.
>
Well, I just went thru it again, and turned off everything but the
cx8800 and ORv51132 stuffs, and now I get this at the and of the
'makeit' script I use here:
WARNING:
/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol mt352_attach
WARNING:
/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol nxt200x_attach
WARNING:
/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol mt352_write
WARNING:
/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol lgdt330x_attach
WARNING:
/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol cx22702_attach
So that definitely was a non-no. That .config is also attached.
>If this is true, then you can either disable nxt200x in
>v4l-kernel/v4l/Makefile, or re-build the cvs modules using the dvb +
> v4l merged-tree build environment.
>http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS
>
>This would, in effect, give the same result as if you had selected
>CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS (selected by default in Kconfig)
>
>...or maybe you initially started a kernel build with the nxt200x
> module selected, then stopped the compile, unselected nxt200x, and
> continued the build process without cleaning in between, and after
> cx88-dvb had already been built? Hmm... I guess that might be a bit
> far-fetched... but to be sure, you can wipe out your kernel tree and
> rebuild it again from scratch using the same .config .
The first thing my makeit script does is a make clean. So it should be
self-cleaning in that scenario.
My buildit26 script also starts everything from scratch, moving old
base version trees out of the way until the new one can be renamed
properly. But theres nothing in it precious, so I can nuke 2.6.15-rc2
and retry again. This time makeing no adjustments to the make xconfig
it finishes up with. So that build from scratch is underway.
And I note that when its going thru the make oldconfig, that nxt2002 is
set as a module, however it is not now loaded into 2.6.14.2, and tvtime
works ok without it, so somehow its getting turned on. Humm, grepping
the .config in the 2.6.14.2 tree shows that it is set as a module. So
thats where that came from. Not sure why I turned it on in the first
place, scratching aching head. (I've got a cold, good thing you can't
catch this virus by email :)
And, I just noticed this go by during the compile of the modules:
Building modules, stage 2.
MODPOST
*** Warning: "nxt200x_attach" [drivers/media/video/cx88/cx88-dvb.ko]
undefined!
It was probably there all the time and I just missed it as its many
pages back up in the history. The depmod at the end of the compile
also reports this again:
WARNING:
/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol nxt200x_attach
>If my assumption above is incorrect, then this will need to be looked
>into a bit deeper. Please, let me know.
Looks like the magnifying glass needs to come out.
>Regards,
>
>Michael Krufky
Thanks Michael.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Tuesday 22 November 2005 19:50, Linus Torvalds wrote:
>On Tue, 22 Nov 2005, Andrew Morton wrote:
>> Jeffrey Hundstad <[email protected]> wrote:
>> > from fs/compat_ioctl.c:52,
>> > from arch/x86_64/ia32/ia32_ioctl.c:14:
>> > include/linux/ext3_fs.h: In function 'ext3_raw_inode':
>> > include/linux/ext3_fs.h:696: error: dereferencing pointer to
>> > incomplete type
>>
>> This might help?
>
>Why does it happen at all, though? And why aren't more people reporting
>this? Something strange going on.
What would I have to turn on to get that one built? Its not building at
all with my current .config.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Tue, 2005-11-22 at 17:39 -0800, Linus Torvalds wrote:
>
> On Tue, 22 Nov 2005, Andrew Morton wrote:
> > >
> > > Why does it happen at all, though?
> >
> > davem recently merged a patch which adds ext3 ioctls to fs/compat_ioctl.c.
> > That required inclusion of ext3 and jbd header files. Those files explode
> > unpleasantly when CONFIG_JBD=n.
>
> Oh. How about just making jbd.h do the rigt thing, and not care about the
> configuration?
>
> If we include jbd.h, we want the jbd data structures. There's never any
> reason to care whether jbd is enabled or not afaik.
>
> Ie maybe just something like this?
>
> (Untested, obviously. I'm just assuming that anything that actually
> _needs_ the jbd functionality should have made sure that jdb is enabled.)
>
> Zan, Jeffrey?
[snip patch]
Yes, that also works for me. It compiled and is running just fine.
--
Zan Lynx <[email protected]>
On Mon, 2005-11-21 at 18:52 -0500, Gene Heskett wrote:
> On Monday 21 November 2005 16:17, Kasper Sandberg wrote:
> >On Sun, 2005-11-20 at 18:58 -0500, Gene Heskett wrote:
> >> On Sunday 20 November 2005 17:36, Kasper Sandberg wrote:
> >> >On Sun, 2005-11-20 at 00:18 -0500, Gene Heskett wrote:
> >> >> On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
> >> >> >There it is (or will soon be - the tar-ball and patches are still
> >> >> >uploading, and mirroring can obviously take some time after
> >> >> > that).
> >> >>
> >> >> First breakage report, tvtime, blue screen no audio. Trying
> >> >> slightly different .config for next build. My tuner (OR51132)
> >> >> seems to be permanently selected in an xconfig screen. Dunno if
> >> >> thats good or bad ATM.
> >> >
> >> >if it needs to be loaded with a parameter you will need to build it
> >> > as a module.. my saa7134 chip needs card=9.
> >>
> >> Its never needed an argument before.
i do not know, but it was the only changeset in that file from a kernel
where it worked, and to 2.6.15-rc1 where it did not work (video, i just
got no images.)
the v4l developers are working on it, i just thought this might help.
> >
> >then you have a good card, mine is a cheap cheap cheap one which
> >apparently doesent have the nessecary embedded info to do proper
> >autodetection, so i gotta manually specify which card i have.
>
> Its a pcHDTV-3000
>
> >> >i am experiencing same problems with saa7134, no video, however i do
> >> > get audio.
> >>
> >> I wasn't, total digital gibberish on screen.
> >>
> >> A full powerdown reboot to 2.6.14.2 fixed it.
> >>
> >> >this is a way to (incorrectly according to v4l devs) "fix" it:
> >> >drivers/media/video/video-buf.c
> >> >change line 1233 to this:
> >> > vma->vm_flags |= VM_DONTEXPAND;
>
> And this will actually do what? Elaborate please.
>
> I think my problem is that somehow, the dvb stuff now has a dependency
> on the nxt200x thing, whatever it is, as if I force a kernel build
> without it, then I get depmod problems at the end of the build telling
> me:
>
> WARNING:/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb
> .ko needs unknown symbol nxt200x_attach.
>
> This was never a requirement up thru 2.6.14.2, and I didn't try 15-rc1
> as the -rc1's are usually a disaster of some kind, and this is also my
> main box. I don't need a filesystem muckup again.
>
> So whatever did that seems in error to me, and should be backed out in
> favor of some other method to bring in the nxt200x bearing cards.
>
Linus Torvalds wrote:
>>BTW: Since I have your ear, this same version DOES seem to have some
>>other bug as well. I did a "make distclean" and the "rm -f" of all he
>>object files hung forever in "D" state. I'm using XFS on IDE disks.
>>I'm using the same config as was posted before. I didn't get anything
>>in an log files that would indicate a problem. Has this been reported?
>>If not, what can I do to make a meaningful report?
>>
>>
>
>I don't recognize those symptoms, so more info would be nice.
>
>For example, it would be good to know where the threads are that are
>waiting uninterruptibly. You should be able to get that info with Sysrq
>'T' (or with the old "ctrl + ScrollLock" thing).
>
>That should tell us if they're hung on a semaphore or waiting for disk IO
>to complete or what...
>
It has been pointed out to me that the XFS team does know about this and
is looking into it. The thread is "unable to use dpkg 2.6.15-rc2".
--
Jeffrey Hundstad
On Tue, Nov 22, 2005 at 11:36:48PM -0500, Gene Heskett wrote:
>...
> Well, I just went thru it again, and turned off everything but the
> cx8800 and ORv51132 stuffs, and now I get this at the and of the
> 'makeit' script I use here:
>
> WARNING:
> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> needs unknown symbol mt352_attach
> WARNING:
> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> needs unknown symbol nxt200x_attach
> WARNING:
> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> needs unknown symbol mt352_write
> WARNING:
> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> needs unknown symbol lgdt330x_attach
> WARNING:
> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> needs unknown symbol cx22702_attach
>...
Nice catch and thanks for your report.
The bug is obvious. A possible patch is below (and at least
drivers/media/video/saa7134/Makefile contains the same bug),
but I'd really prfer getting rid of the -DHAVE_* stuff in the
Makefiles and using Kconfig variables instead.
Would such a patch be accepted?
> Cheers, Gene
cu
Adrian
BTW: Please don't strip the Cc whenreplying to linux-kernel.
--- linux-2.6.15-rc2/drivers/media/video/cx88/Makefile.old 2005-11-23 18:34:07.000000000 +0100
+++ linux-2.6.15-rc2/drivers/media/video/cx88/Makefile 2005-11-23 18:34:18.000000000 +0100
@@ -9,21 +9,21 @@
EXTRA_CFLAGS += -I$(src)/..
EXTRA_CFLAGS += -I$(srctree)/drivers/media/dvb/dvb-core
EXTRA_CFLAGS += -I$(srctree)/drivers/media/dvb/frontends
-ifneq ($(CONFIG_VIDEO_BUF_DVB),n)
+ifneq ($(CONFIG_VIDEO_BUF_DVB),)
EXTRA_CFLAGS += -DHAVE_VIDEO_BUF_DVB=1
endif
-ifneq ($(CONFIG_DVB_CX22702),n)
+ifneq ($(CONFIG_DVB_CX22702),)
EXTRA_CFLAGS += -DHAVE_CX22702=1
endif
-ifneq ($(CONFIG_DVB_OR51132),n)
+ifneq ($(CONFIG_DVB_OR51132),)
EXTRA_CFLAGS += -DHAVE_OR51132=1
endif
-ifneq ($(CONFIG_DVB_LGDT330X),n)
+ifneq ($(CONFIG_DVB_LGDT330X),)
EXTRA_CFLAGS += -DHAVE_LGDT330X=1
endif
-ifneq ($(CONFIG_DVB_MT352),n)
+ifneq ($(CONFIG_DVB_MT352),)
EXTRA_CFLAGS += -DHAVE_MT352=1
endif
-ifneq ($(CONFIG_DVB_NXT200X),n)
+ifneq ($(CONFIG_DVB_NXT200X),)
EXTRA_CFLAGS += -DHAVE_NXT200X=1
endif
On Wednesday 23 November 2005 12:42, Adrian Bunk wrote:
>On Tue, Nov 22, 2005 at 11:36:48PM -0500, Gene Heskett wrote:
>>...
>> Well, I just went thru it again, and turned off everything but the
>> cx8800 and ORv51132 stuffs, and now I get this at the and of the
>> 'makeit' script I use here:
>>
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol mt352_attach
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol nxt200x_attach
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol mt352_write
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol lgdt330x_attach
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol cx22702_attach
>>...
>
>Nice catch and thanks for your report.
You're welcome.
>The bug is obvious. A possible patch is below (and at least
>drivers/media/video/saa7134/Makefile contains the same bug),
>but I'd really prfer getting rid of the -DHAVE_* stuff in the
>Makefiles and using Kconfig variables instead.
>
>Would such a patch be accepted?
I'm trying this to build with this patch right now. No errors,
although I did see it building a cx88-blackbird.o, which I've ndi what
that is. Now the acid test, does tvtime work again, or do I have to do
a powerdown reset and reboot to 2.6.14.2 to restore it. Reboot to
2.6.15-rc2-patched comeing up.
Rebooted to this patch. Tvtime now has very faint sound, like the
station is 200 miles away, and no video. Blue screen even if signal
detection is turned off. Its like the antenna input has been
disconnected. The dish receiver in use here is banging it with nearly
65,000 u-volts of signal, so some is leaking thru.
Now, to see if a simple reboot to 2.6.14.2 fixes it, or if I have to do
the full powerdown again. brb.
>> Cheers, Gene
>
>cu
>Adrian
>
>BTW: Please don't strip the Cc whenreplying to linux-kernel.
>
>
>
>--- linux-2.6.15-rc2/drivers/media/video/cx88/Makefile.old 2005-11-23
> 18:34:07.000000000 +0100 +++
> linux-2.6.15-rc2/drivers/media/video/cx88/Makefile 2005-11-23
> 18:34:18.000000000 +0100 @@ -9,21 +9,21 @@
> EXTRA_CFLAGS += -I$(src)/..
> EXTRA_CFLAGS += -I$(srctree)/drivers/media/dvb/dvb-core
> EXTRA_CFLAGS += -I$(srctree)/drivers/media/dvb/frontends
>-ifneq ($(CONFIG_VIDEO_BUF_DVB),n)
>+ifneq ($(CONFIG_VIDEO_BUF_DVB),)
> EXTRA_CFLAGS += -DHAVE_VIDEO_BUF_DVB=1
> endif
>-ifneq ($(CONFIG_DVB_CX22702),n)
>+ifneq ($(CONFIG_DVB_CX22702),)
> EXTRA_CFLAGS += -DHAVE_CX22702=1
> endif
>-ifneq ($(CONFIG_DVB_OR51132),n)
>+ifneq ($(CONFIG_DVB_OR51132),)
> EXTRA_CFLAGS += -DHAVE_OR51132=1
> endif
>-ifneq ($(CONFIG_DVB_LGDT330X),n)
>+ifneq ($(CONFIG_DVB_LGDT330X),)
> EXTRA_CFLAGS += -DHAVE_LGDT330X=1
> endif
>-ifneq ($(CONFIG_DVB_MT352),n)
>+ifneq ($(CONFIG_DVB_MT352),)
> EXTRA_CFLAGS += -DHAVE_MT352=1
> endif
>-ifneq ($(CONFIG_DVB_NXT200X),n)
>+ifneq ($(CONFIG_DVB_NXT200X),)
> EXTRA_CFLAGS += -DHAVE_NXT200X=1
> endif
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Wed, Nov 23, 2005 at 06:42:37PM +0100, Adrian Bunk wrote:
> EXTRA_CFLAGS += -I$(src)/..
Wonder if this compiles with O=...
> -ifneq ($(CONFIG_VIDEO_BUF_DVB),n)
> +ifneq ($(CONFIG_VIDEO_BUF_DVB),)
> EXTRA_CFLAGS += -DHAVE_VIDEO_BUF_DVB=1
> endif
> -ifneq ($(CONFIG_DVB_CX22702),n)
> +ifneq ($(CONFIG_DVB_CX22702),)
> EXTRA_CFLAGS += -DHAVE_CX22702=1
> endif
> -ifneq ($(CONFIG_DVB_OR51132),n)
> +ifneq ($(CONFIG_DVB_OR51132),)
> EXTRA_CFLAGS += -DHAVE_OR51132=1
> endif
> -ifneq ($(CONFIG_DVB_LGDT330X),n)
> +ifneq ($(CONFIG_DVB_LGDT330X),)
> EXTRA_CFLAGS += -DHAVE_LGDT330X=1
> endif
> -ifneq ($(CONFIG_DVB_MT352),n)
> +ifneq ($(CONFIG_DVB_MT352),)
> EXTRA_CFLAGS += -DHAVE_MT352=1
> endif
> -ifneq ($(CONFIG_DVB_NXT200X),n)
> +ifneq ($(CONFIG_DVB_NXT200X),)
> EXTRA_CFLAGS += -DHAVE_NXT200X=1
> endif
> -
If we stick with HAVE_XXX then please use following style:
extra-cflags-$(CONFIG_VIDEO_BUF_DVB) += -DHAVE_VIDEO_BUF_DVB=1
extra-cflags-$(CONFIG_DVB_CX22702) += -DHAVE_CX22702=1
extra-cflags-$(CONFIG_DVB_OR51132) += -DHAVE_OR51132=1
extra-cflags-$(CONFIG_DVB_LGDT330X) += -DHAVE_LGDT330X=1
extra-cflags-$(CONFIG_DVB_MT352) += -DHAVE_MT352=1
extra-cflags-$(CONFIG_DVB_NXT200X) += -DHAVE_NXT200X=1
EXTRA_CFLAGS += $(extra-cflags-y) $(extra-cflags-m)
Sam
On Wednesday 23 November 2005 12:42, Adrian Bunk wrote:
>On Tue, Nov 22, 2005 at 11:36:48PM -0500, Gene Heskett wrote:
>>...
>> Well, I just went thru it again, and turned off everything but the
>> cx8800 and ORv51132 stuffs, and now I get this at the and of the
>> 'makeit' script I use here:
>>
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol mt352_attach
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol nxt200x_attach
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol mt352_write
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol lgdt330x_attach
>> WARNING:
>> /lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>> needs unknown symbol cx22702_attach
>>...
>
>Nice catch and thanks for your report.
>
>The bug is obvious. A possible patch is below (and at least
>drivers/media/video/saa7134/Makefile contains the same bug),
>but I'd really prfer getting rid of the -DHAVE_* stuff in the
>Makefiles and using Kconfig variables instead.
Damn, I sent the last message too quick. So this is a continuation of
that message.
Anyway, a hot reboot to 2.6.14.2 did not fix it fully although I
noticed that with signal detection turned off, I now had digital noise
bars instead of a blue screen. Its now working normally again after a
full powerdown reboot to 2.6.14.2. So while that patch would make it
build ok, it doesn't make it work...
>Would such a patch be accepted?
>
>> Cheers, Gene
>
>cu
>Adrian
>
>BTW: Please don't strip the Cc whenreplying to linux-kernel.
Oopps, sorry. I guess I need to make a reply2all button for kmail...
>
>--- linux-2.6.15-rc2/drivers/media/video/cx88/Makefile.old 2005-11-23
> 18:34:07.000000000 +0100 +++
> linux-2.6.15-rc2/drivers/media/video/cx88/Makefile 2005-11-23
> 18:34:18.000000000 +0100 @@ -9,21 +9,21 @@
> EXTRA_CFLAGS += -I$(src)/..
> EXTRA_CFLAGS += -I$(srctree)/drivers/media/dvb/dvb-core
> EXTRA_CFLAGS += -I$(srctree)/drivers/media/dvb/frontends
>-ifneq ($(CONFIG_VIDEO_BUF_DVB),n)
>+ifneq ($(CONFIG_VIDEO_BUF_DVB),)
> EXTRA_CFLAGS += -DHAVE_VIDEO_BUF_DVB=1
> endif
>-ifneq ($(CONFIG_DVB_CX22702),n)
>+ifneq ($(CONFIG_DVB_CX22702),)
> EXTRA_CFLAGS += -DHAVE_CX22702=1
> endif
>-ifneq ($(CONFIG_DVB_OR51132),n)
>+ifneq ($(CONFIG_DVB_OR51132),)
> EXTRA_CFLAGS += -DHAVE_OR51132=1
> endif
>-ifneq ($(CONFIG_DVB_LGDT330X),n)
>+ifneq ($(CONFIG_DVB_LGDT330X),)
> EXTRA_CFLAGS += -DHAVE_LGDT330X=1
> endif
>-ifneq ($(CONFIG_DVB_MT352),n)
>+ifneq ($(CONFIG_DVB_MT352),)
> EXTRA_CFLAGS += -DHAVE_MT352=1
> endif
>-ifneq ($(CONFIG_DVB_NXT200X),n)
>+ifneq ($(CONFIG_DVB_NXT200X),)
> EXTRA_CFLAGS += -DHAVE_NXT200X=1
> endif
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Wed, Nov 23, 2005 at 07:26:09PM +0100, Sam Ravnborg wrote:
> On Wed, Nov 23, 2005 at 06:42:37PM +0100, Adrian Bunk wrote:
> > EXTRA_CFLAGS += -I$(src)/..
> Wonder if this compiles with O=...
>
> > -ifneq ($(CONFIG_VIDEO_BUF_DVB),n)
> > +ifneq ($(CONFIG_VIDEO_BUF_DVB),)
> > EXTRA_CFLAGS += -DHAVE_VIDEO_BUF_DVB=1
> > endif
> > -ifneq ($(CONFIG_DVB_CX22702),n)
> > +ifneq ($(CONFIG_DVB_CX22702),)
> > EXTRA_CFLAGS += -DHAVE_CX22702=1
> > endif
> > -ifneq ($(CONFIG_DVB_OR51132),n)
> > +ifneq ($(CONFIG_DVB_OR51132),)
> > EXTRA_CFLAGS += -DHAVE_OR51132=1
> > endif
> > -ifneq ($(CONFIG_DVB_LGDT330X),n)
> > +ifneq ($(CONFIG_DVB_LGDT330X),)
> > EXTRA_CFLAGS += -DHAVE_LGDT330X=1
> > endif
> > -ifneq ($(CONFIG_DVB_MT352),n)
> > +ifneq ($(CONFIG_DVB_MT352),)
> > EXTRA_CFLAGS += -DHAVE_MT352=1
> > endif
> > -ifneq ($(CONFIG_DVB_NXT200X),n)
> > +ifneq ($(CONFIG_DVB_NXT200X),)
> > EXTRA_CFLAGS += -DHAVE_NXT200X=1
> > endif
> > -
>
> If we stick with HAVE_XXX then please use following style:
>
> extra-cflags-$(CONFIG_VIDEO_BUF_DVB) += -DHAVE_VIDEO_BUF_DVB=1
> extra-cflags-$(CONFIG_DVB_CX22702) += -DHAVE_CX22702=1
> extra-cflags-$(CONFIG_DVB_OR51132) += -DHAVE_OR51132=1
> extra-cflags-$(CONFIG_DVB_LGDT330X) += -DHAVE_LGDT330X=1
> extra-cflags-$(CONFIG_DVB_MT352) += -DHAVE_MT352=1
> extra-cflags-$(CONFIG_DVB_NXT200X) += -DHAVE_NXT200X=1
>
> EXTRA_CFLAGS += $(extra-cflags-y) $(extra-cflags-m)
And this does still not solve all problems, the whole approach is a
mess.
E.g. CONFIG_VIDEO_CX88_DVB=y, CONFIG_DVB_CX22702=m is one of the many
cases that still won't compile.
I'll try to find a way for doing this all properly.
> Sam
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
Adrian Bunk wrote:
>On Tue, Nov 22, 2005 at 11:36:48PM -0500, Gene Heskett wrote:
>
>
>>...
>>Well, I just went thru it again, and turned off everything but the
>>cx8800 and ORv51132 stuffs, and now I get this at the and of the
>>'makeit' script I use here:
>>
>>WARNING:
>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>needs unknown symbol mt352_attach
>>WARNING:
>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>needs unknown symbol nxt200x_attach
>>WARNING:
>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>needs unknown symbol mt352_write
>>WARNING:
>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>needs unknown symbol lgdt330x_attach
>>WARNING:
>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>needs unknown symbol cx22702_attach
>>...
>>
>>
>Nice catch and thanks for your report.
>
>The bug is obvious. A possible patch is below (and at least
>drivers/media/video/saa7134/Makefile contains the same bug),
>but I'd really prfer getting rid of the -DHAVE_* stuff in the
>Makefiles and using Kconfig variables instead.
>
>
We need to keep the -DHAVE_FOO stuff there, in order to satisfy the
following requirements:
1) To allow the option of only selecting those frontends required by
specific dvb hardware, without forcing all modules to be loaded... This
feature is optional, and I implemented it in response to the demand from
some hybrid v4l/dvb device users, (and myself) Why force a driver to
load every frontend module if it isnt required by the hardware? --
apparantly the implementation was less than perfect. I had originally
intended for this to live in -mm for a bit, but when the merge window
came around, Mauro had sent it upstream before I had the chance to
create alternate patches for linus' tree.
2) (more importantly) To allow v4l-kernel cvs to retain backwards
compatability with older kernels..
I had originally tried to rename these to use the Kconfig variables, but
LKML people asked for it to be changed back.
Please do not remove this feature -- if it is broken, then we should try
to fix it, rather than remove it. If the specific frontend selection
isn't working, then I guess we can revert back to the old behavior where
every frontend is forced, but I would rather not.
>Would such a patch be accepted?
>
>>Cheers, Gene
>>
>>
>cu
>Adrian
>
>
If it fixes Gene's problem (a quick glance at his emails suggests that
it does) then:
Acked-by: Michael Krufky <[email protected]>
although Sam Ravnborg's suggestion looks better to me. Unfortunately, I
will be unable to test this out on my system until After Thanksgiving
(on Friday) ...
>BTW: Please don't strip the Cc whenreplying to linux-kernel.
>
>
>--- linux-2.6.15-rc2/drivers/media/video/cx88/Makefile.old 2005-11-23 18:34:07.000000000 +0100
>+++ linux-2.6.15-rc2/drivers/media/video/cx88/Makefile 2005-11-23 18:34:18.000000000 +0100
>@@ -9,21 +9,21 @@
> EXTRA_CFLAGS += -I$(src)/..
> EXTRA_CFLAGS += -I$(srctree)/drivers/media/dvb/dvb-core
> EXTRA_CFLAGS += -I$(srctree)/drivers/media/dvb/frontends
>-ifneq ($(CONFIG_VIDEO_BUF_DVB),n)
>+ifneq ($(CONFIG_VIDEO_BUF_DVB),)
> EXTRA_CFLAGS += -DHAVE_VIDEO_BUF_DVB=1
> endif
>-ifneq ($(CONFIG_DVB_CX22702),n)
>+ifneq ($(CONFIG_DVB_CX22702),)
> EXTRA_CFLAGS += -DHAVE_CX22702=1
> endif
>-ifneq ($(CONFIG_DVB_OR51132),n)
>+ifneq ($(CONFIG_DVB_OR51132),)
> EXTRA_CFLAGS += -DHAVE_OR51132=1
> endif
>-ifneq ($(CONFIG_DVB_LGDT330X),n)
>+ifneq ($(CONFIG_DVB_LGDT330X),)
> EXTRA_CFLAGS += -DHAVE_LGDT330X=1
> endif
>-ifneq ($(CONFIG_DVB_MT352),n)
>+ifneq ($(CONFIG_DVB_MT352),)
> EXTRA_CFLAGS += -DHAVE_MT352=1
> endif
>-ifneq ($(CONFIG_DVB_NXT200X),n)
>+ifneq ($(CONFIG_DVB_NXT200X),)
> EXTRA_CFLAGS += -DHAVE_NXT200X=1
> endif
>-
>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/
>
>
--
Michael Krufky
On Wednesday 23 November 2005 14:17, Michael Krufky wrote:
[...]
>If it fixes Gene's problem (a quick glance at his emails suggests that
>it does) then:
Read further Michael, it still takes a _cold_ reboot to 2.6.14.2 to fix
it.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
Gene Heskett wrote:
>On Wednesday 23 November 2005 14:17, Michael Krufky wrote:
>
>[...]
>
>>f it fixes Gene's problem (a quick glance at his emails suggests that
>>it does) then:
>>
>>
>Read further Michael, it still takes a _cold_ reboot to 2.6.14.2 to fix
>it.
>
I'm sorry -- I should have been clearer... It fixes the following error
message, correct?
Gene Heskett wrote:
> WARNING:/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb
> .ko needs unknown symbol nxt200x_attach.
About the cold reboot needed for 2.6.14.2, well, that is completely
unrelated...
First, does the patch fix the unknown symbol error? If so, then the
patch is correct.
Moving on........
Kirk Lapray wrote both OR51132 and NXT200X frontend modules (cc added) ...
First off, Gene, I am still under the impression that both v4l and dvb
subsystems are broken under 2.6.15 due to the memory bugs... I don't
know if Hugh Dickins fixed those yet or not.
Please try to build merged v4l+dvb cvs trees against your 2.6.14.2
kernel, and tell me if you are having the same problems. If you are
indeed having the same problem, then it confirms that something in the
nxt200x module is causing problems in the OR51132 module.
Kirk, are you able to use both modules together using both pcHDTV and
ATI HDTV Wonder PCI cx88 boards simultaneously without causing any
conflicts?
Once again, Gene, please follow the tree-merge instructions located at:
http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS
Please let me know if the problem persists. If the problem is gone,
then nxt200x is a red herring.
Regards,
Michael Krufky
On Wednesday 23 November 2005 14:54, Michael Krufky wrote:
>Gene Heskett wrote:
>>On Wednesday 23 November 2005 14:17, Michael Krufky wrote:
>>
>>[...]
>>
>>>f it fixes Gene's problem (a quick glance at his emails suggests that
>>>it does) then:
>>
>>Read further Michael, it still takes a _cold_ reboot to 2.6.14.2 to
>> fix it.
>
>I'm sorry -- I should have been clearer... It fixes the following error
>message, correct?
>
>Gene Heskett wrote:
>> WARNING:/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-
>>dvb .ko needs unknown symbol nxt200x_attach.
>
>About the cold reboot needed for 2.6.14.2, well, that is completely
>unrelated...
>
>First, does the patch fix the unknown symbol error? If so, then the
>patch is correct.
Yes, it fixes that just fine.
>
>Moving on........
>
>Kirk Lapray wrote both OR51132 and NXT200X frontend modules (cc added)
> ...
>
>First off, Gene, I am still under the impression that both v4l and dvb
>subsystems are broken under 2.6.15 due to the memory bugs... I don't
>know if Hugh Dickins fixed those yet or not.
Neither do I. But as a tv engineer with 50+ years of experience, the
general appearance is if the antenna cable has been disconnected and
held about 2" away from the f-59 connector when a hot reboot is done.
The audio in both cases sounds like its a station 300 miles away when
the atmospherics are behaving themselves.
>
>Please try to build merged v4l+dvb cvs trees against your 2.6.14.2
>kernel, and tell me if you are having the same problems. If you are
>indeed having the same problem, then it confirms that something in the
>nxt200x module is causing problems in the OR51132 module.
And how & where do I obtain that?
>Kirk, are you able to use both modules together using both pcHDTV and
>ATI HDTV Wonder PCI cx88 boards simultaneously without causing any
>conflicts?
>
>Once again, Gene, please follow the tree-merge instructions located at:
>
>http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS
I'll give this a shot and advise on the results.
>Please let me know if the problem persists. If the problem is gone,
>then nxt200x is a red herring.
>
>Regards,
>
>Michael Krufky
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
Gene Heskett wrote:
>On Wednesday 23 November 2005 14:54, Michael Krufky wrote:
>
>
>>Gene Heskett wrote:
>>
>>
>>>On Wednesday 23 November 2005 14:17, Michael Krufky wrote:
>>>
>>>[...]
>>>
>>>>f it fixes Gene's problem (a quick glance at his emails suggests that
>>>>it does) then:
>>>>
>>>>
>>>Read further Michael, it still takes a _cold_ reboot to 2.6.14.2 to
>>>fix it.
>>>
>>>
>>I'm sorry -- I should have been clearer... It fixes the following error
>>message, correct?
>>
>>Gene Heskett wrote:
>>
>>
>>>WARNING:/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-
>>>dvb .ko needs unknown symbol nxt200x_attach.
>>>
>>>
>>About the cold reboot needed for 2.6.14.2, well, that is completely
>>unrelated...
>>
>>First, does the patch fix the unknown symbol error? If so, then the
>>patch is correct.
>>
>>
>Yes, it fixes that just fine.
>
>
GREAT! ... So then Adrian's patch with my ACK should be applied, in
addition to a similar patch for saa7134-dvb ... I'd like to try Sam's
method on my machine... if it works, then I will send in a new patch on
Friday.
>>Moving on........
>>
>>Kirk Lapray wrote both OR51132 and NXT200X frontend modules (cc added)
>>...
>>
>>First off, Gene, I am still under the impression that both v4l and dvb
>>subsystems are broken under 2.6.15 due to the memory bugs... I don't
>>know if Hugh Dickins fixed those yet or not.
>>
>>
>Neither do I. But as a tv engineer with 50+ years of experience, the
>general appearance is if the antenna cable has been disconnected and
>held about 2" away from the f-59 connector when a hot reboot is done.
>The audio in both cases sounds like its a station 300 miles away when
>the atmospherics are behaving themselves.
>
>
More than likely, I would assume that the issue has something to do with
some code inside nxt200x.c that makes some additional devices visible on
the i2c bus... This should only affect devices that use nxt200x module,
but perhaps there is something going on that is causing interference on
the i2c bus of your card? This is just a guess.... Results from your
cvs test will show us some better information.
If you are successfully loading cx88-dvb WITHOUT nxt200x module (before
the cold boot) then the above is not the case.
Meanwhile, I repeat, it is well-known that the media tree is broken in
2.6.15 ... I think we'll get better results after Hugh Dickin's sends in
another patch. In the meantime, you can test the v4l + dvb code using
the merge-trees build method on the cvs repositories. (see below)
>>Please try to build merged v4l+dvb cvs trees against your 2.6.14.2
>>kernel, and tell me if you are having the same problems. If you are
>>indeed having the same problem, then it confirms that something in the
>>nxt200x module is causing problems in the OR51132 module.
>>
>>
>And how & where do I obtain that?
>
Link to wiki-howto on linuxtv.org provided below.
>>Kirk, are you able to use both modules together using both pcHDTV and
>>ATI HDTV Wonder PCI cx88 boards simultaneously without causing any
>>conflicts?
>>
>>Once again, Gene, please follow the tree-merge instructions located at:
>>
>>http://linuxtv.org/v4lwiki/index.php/How_to_build_from_CVS
>>
>>
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>I'll give this a shot and advise on the results.
>
>>Please let me know if the problem persists. If the problem is gone,
>>then nxt200x is a red herring.
>>
>>Regards,
>>
>>Michael Krufky
>>
>>
I'll keep my email open...
Thanks,
Michael Krufky
On Wed, Nov 23, 2005 at 02:17:45PM -0500, Michael Krufky wrote:
> Adrian Bunk wrote:
>
> >On Tue, Nov 22, 2005 at 11:36:48PM -0500, Gene Heskett wrote:
> >
> >
> >>...
> >>Well, I just went thru it again, and turned off everything but the
> >>cx8800 and ORv51132 stuffs, and now I get this at the and of the
> >>'makeit' script I use here:
> >>
> >>WARNING:
> >>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> >>needs unknown symbol mt352_attach
> >>WARNING:
> >>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> >>needs unknown symbol nxt200x_attach
> >>WARNING:
> >>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> >>needs unknown symbol mt352_write
> >>WARNING:
> >>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> >>needs unknown symbol lgdt330x_attach
> >>WARNING:
> >>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
> >>needs unknown symbol cx22702_attach
> >>...
> >>
> >>
> >Nice catch and thanks for your report.
> >
> >The bug is obvious. A possible patch is below (and at least
> >drivers/media/video/saa7134/Makefile contains the same bug),
> >but I'd really prfer getting rid of the -DHAVE_* stuff in the
> >Makefiles and using Kconfig variables instead.
> >
> >
> We need to keep the -DHAVE_FOO stuff there, in order to satisfy the
> following requirements:
>
> 1) To allow the option of only selecting those frontends required by
> specific dvb hardware, without forcing all modules to be loaded... This
> feature is optional, and I implemented it in response to the demand from
> some hybrid v4l/dvb device users, (and myself) Why force a driver to
> load every frontend module if it isnt required by the hardware? --
> apparantly the implementation was less than perfect. I had originally
> intended for this to live in -mm for a bit, but when the merge window
> came around, Mauro had sent it upstream before I had the chance to
> create alternate patches for linus' tree.
>
> 2) (more importantly) To allow v4l-kernel cvs to retain backwards
> compatability with older kernels..
>
> I had originally tried to rename these to use the Kconfig variables, but
> LKML people asked for it to be changed back.
>
> Please do not remove this feature -- if it is broken, then we should try
> to fix it, rather than remove it. If the specific frontend selection
> isn't working, then I guess we can revert back to the old behavior where
> every frontend is forced, but I would rather not.
I do not yet know how to fix it, but configurations like
CONFIG_VIDEO_CX88_DVB=y, CONFIG_DVB_CX22702=m are currently compile
errors.
> Michael Krufky
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
Adrian Bunk wrote:
>On Wed, Nov 23, 2005 at 02:17:45PM -0500, Michael Krufky wrote:
>
>
>>Adrian Bunk wrote:
>>
>>>On Tue, Nov 22, 2005 at 11:36:48PM -0500, Gene Heskett wrote:
>>>
>>>>...
>>>>Well, I just went thru it again, and turned off everything but the
>>>>cx8800 and ORv51132 stuffs, and now I get this at the and of the
>>>>'makeit' script I use here:
>>>>
>>>>WARNING:
>>>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>>>needs unknown symbol mt352_attach
>>>>WARNING:
>>>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>>>needs unknown symbol nxt200x_attach
>>>>WARNING:
>>>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>>>needs unknown symbol mt352_write
>>>>WARNING:
>>>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>>>needs unknown symbol lgdt330x_attach
>>>>WARNING:
>>>>/lib/modules/2.6.15-rc2/kernel/drivers/media/video/cx88/cx88-dvb.ko
>>>>needs unknown symbol cx22702_attach
>>>>...
>>>>
>>>Nice catch and thanks for your report.
>>>
>>>The bug is obvious. A possible patch is below (and at least
>>>drivers/media/video/saa7134/Makefile contains the same bug),
>>>but I'd really prfer getting rid of the -DHAVE_* stuff in the
>>>Makefiles and using Kconfig variables instead.
>>>
>>We need to keep the -DHAVE_FOO stuff there, in order to satisfy the
>>following requirements:
>>
>>1) To allow the option of only selecting those frontends required by
>>specific dvb hardware, without forcing all modules to be loaded... This
>>feature is optional, and I implemented it in response to the demand from
>>some hybrid v4l/dvb device users, (and myself) Why force a driver to
>>load every frontend module if it isnt required by the hardware? --
>>apparantly the implementation was less than perfect. I had originally
>>intended for this to live in -mm for a bit, but when the merge window
>>came around, Mauro had sent it upstream before I had the chance to
>>create alternate patches for linus' tree.
>>
>>2) (more importantly) To allow v4l-kernel cvs to retain backwards
>>compatability with older kernels..
>>
>>I had originally tried to rename these to use the Kconfig variables, but
>>LKML people asked for it to be changed back.
>>
>>Please do not remove this feature -- if it is broken, then we should try
>>to fix it, rather than remove it. If the specific frontend selection
>>isn't working, then I guess we can revert back to the old behavior where
>>every frontend is forced, but I would rather not.
>>
>>
>I do not yet know how to fix it, but configurations like
>CONFIG_VIDEO_CX88_DVB=y, CONFIG_DVB_CX22702=m are currently compile
>errors.
>
>
AHA! I have not tested this with cx88-dvb compiled into the kernel (y)
-- I have only tested as a module (m) ..... Looks like I have a lot of
testing to do before the end of this week.
Adrian, does it work if you select CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS
?? Selecting this option is, in effect, exactly equal to the old
behavior of forcing support for every single frontend supported by
cx88-dvb to be built.
Looks like the problem is the following:
If cx88-dvb is selected (y), then then the frontends should also be
selected (y) ... but instead, they are being selected (m)
Meanwhile, if cx88-dvb is selected (m) then everything is fine, since
the frontends are also selected(m) ...
Is my assessment correct?
-Michael Krufky
On Wed, Nov 23, 2005 at 03:49:31PM -0500, Michael Krufky wrote:
> Adrian Bunk wrote:
>...
> >I do not yet know how to fix it, but configurations like
> >CONFIG_VIDEO_CX88_DVB=y, CONFIG_DVB_CX22702=m are currently compile
> >errors.
> >
> >
> AHA! I have not tested this with cx88-dvb compiled into the kernel (y)
> -- I have only tested as a module (m) ..... Looks like I have a lot of
> testing to do before the end of this week.
>
> Adrian, does it work if you select CONFIG_VIDEO_CX88_DVB_ALL_FRONTENDS
> ?? Selecting this option is, in effect, exactly equal to the old
> behavior of forcing support for every single frontend supported by
> cx88-dvb to be built.
Yes, this should fix these problems.
> Looks like the problem is the following:
> If cx88-dvb is selected (y), then then the frontends should also be
> selected (y) ... but instead, they are being selected (m)
s/are being/can be/
> Meanwhile, if cx88-dvb is selected (m) then everything is fine, since
> the frontends are also selected(m) ...
I haven't yet found any problem with it.
> Is my assessment correct?
>
> -Michael Krufky
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 Wed, 23 Nov 2005, Michael Krufky wrote:
> > >
> > > First off, Gene, I am still under the impression that both v4l and dvb
> > > subsystems are broken under 2.6.15 due to the memory bugs... I don't
> > > know if Hugh Dickins fixed those yet or not.
They should be fixed in today's 2.6.15-rc2-git3
(aside from a couple of patches to drivers/char/drm coming later).
If you still have problems you think I'm responsible for, let me know.
Hugh
On Wednesday 23 November 2005 16:26, Hugh Dickins wrote:
>On Wed, 23 Nov 2005, Michael Krufky wrote:
>> > > First off, Gene, I am still under the impression that both v4l
>> > > and dvb subsystems are broken under 2.6.15 due to the memory
>> > > bugs... I don't know if Hugh Dickins fixed those yet or not.
>
>They should be fixed in today's 2.6.15-rc2-git3
>(aside from a couple of patches to drivers/char/drm coming later).
>If you still have problems you think I'm responsible for, let me know.
>
>Hugh
I'm not familiar enough with git yet to try that without some hand
holding :(
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Wed, 23 Nov 2005, Gene Heskett wrote:
> On Wednesday 23 November 2005 16:26, Hugh Dickins wrote:
> >
> >They should be fixed in today's 2.6.15-rc2-git3
> >(aside from a couple of patches to drivers/char/drm coming later).
> >If you still have problems you think I'm responsible for, let me know.
>
> I'm not familiar enough with git yet to try that without some hand
> holding :(
No git familiarity needed:
http://ftp.kernel.org/pub/linux/kernel/v2.6/snapshots/
contains the daily patches against recent -rcs
Hugh
On Wednesday 23 November 2005 18:40, Hugh Dickins wrote:
>On Wed, 23 Nov 2005, Gene Heskett wrote:
>> On Wednesday 23 November 2005 16:26, Hugh Dickins wrote:
>> >They should be fixed in today's 2.6.15-rc2-git3
>> >(aside from a couple of patches to drivers/char/drm coming later).
>> >If you still have problems you think I'm responsible for, let me
>> > know.
>>
>> I'm not familiar enough with git yet to try that without some hand
>> holding :(
>
>No git familiarity needed:
>http://ftp.kernel.org/pub/linux/kernel/v2.6/snapshots/
>contains the daily patches against recent -rcs
Got it, now I just need some more giddyup. This cold is dragging me
down.
Thanks Hugh.
>Hugh
>-
>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/
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
On Wednesday 23 November 2005 18:40, Hugh Dickins wrote:
>On Wed, 23 Nov 2005, Gene Heskett wrote:
>> On Wednesday 23 November 2005 16:26, Hugh Dickins wrote:
>> >They should be fixed in today's 2.6.15-rc2-git3
>> >(aside from a couple of patches to drivers/char/drm coming later).
>> >If you still have problems you think I'm responsible for, let me
>> > know.
>>
>> I'm not familiar enough with git yet to try that without some hand
>> holding :(
>
>No git familiarity needed:
>http://ftp.kernel.org/pub/linux/kernel/v2.6/snapshots/
>contains the daily patches against recent -rcs
>
>Hugh
Unforch, using a 2.6.14 base, applying 2.6.15-rc2 followed by
2.6.15-rc2-git3 blows up about 24 seconds into my makeit script:
CC arch/i386/kernel/cpu/mtrr/main.o
arch/i386/kernel/cpu/mtrr/main.c: In function `set_mtrr':
arch/i386/kernel/cpu/mtrr/main.c:225: error: `ipi_handler' undeclared
(first use in this function)
arch/i386/kernel/cpu/mtrr/main.c:225: error: (Each undeclared identifier
is reported only once
arch/i386/kernel/cpu/mtrr/main.c:225: error: for each function it
appears in.)
make[3]: *** [arch/i386/kernel/cpu/mtrr/main.o] Error 1
make[2]: *** [arch/i386/kernel/cpu/mtrr] Error 2
make[1]: *** [arch/i386/kernel/cpu] Error 2
make: *** [arch/i386/kernel] Error 2
??
>-
>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/
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
Gene wrote:
> I'm not familiar enough with git yet to try that without some hand
> holding :(
Git's bisect was announced in Linus's post:
http://lkml.org/lkml/2005/7/30/106
Matt Mackall's mercurial (http://www.selenic.com/mercurial/) has a bisect
as well:
http://www.selenic.com/pipermail/mercurial/2005-September/004697.html
--
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <[email protected]> 1.925.600.0401
On Wed, 23 Nov 2005, Gene Heskett wrote:
> On Wednesday 23 November 2005 18:40, Hugh Dickins wrote:
> >
> >No git familiarity needed:
> >http://ftp.kernel.org/pub/linux/kernel/v2.6/snapshots/
> >contains the daily patches against recent -rcs
>
> Unforch, using a 2.6.14 base, applying 2.6.15-rc2 followed by
> 2.6.15-rc2-git3 blows up about 24 seconds into my makeit script:
>
> CC arch/i386/kernel/cpu/mtrr/main.o
> arch/i386/kernel/cpu/mtrr/main.c: In function `set_mtrr':
> arch/i386/kernel/cpu/mtrr/main.c:225: error: `ipi_handler' undeclared
> (first use in this function)
> arch/i386/kernel/cpu/mtrr/main.c:225: error: (Each undeclared identifier
> is reported only once
> arch/i386/kernel/cpu/mtrr/main.c:225: error: for each function it
> appears in.)
> make[3]: *** [arch/i386/kernel/cpu/mtrr/main.o] Error 1
> make[2]: *** [arch/i386/kernel/cpu/mtrr] Error 2
> make[1]: *** [arch/i386/kernel/cpu] Error 2
> make: *** [arch/i386/kernel] Error 2
That's one of the things fixed by Andrew's patch below
(though Linus fixed it differently in the end).
Or you could just wait for 2.6.15-rc2-git4, should be along soon.
Hugh
diff -puN include/linux/smp.h~smp_call_function-must-be-a-macro include/linux/smp.h
--- devel/include/linux/smp.h~smp_call_function-must-be-a-macro 2005-11-23 00:14:19.000000000 -0800
+++ devel-akpm/include/linux/smp.h 2005-11-23 00:20:54.000000000 -0800
@@ -94,13 +94,7 @@ void smp_prepare_boot_cpu(void);
*/
#define raw_smp_processor_id() 0
#define hard_smp_processor_id() 0
-
-static inline int smp_call_function(void (*func) (void *info), void *info,
- int retry, int wait)
-{
- return 0;
-}
-
+#define smp_call_function(func,info,retry,wait) ({ 0; })
#define on_each_cpu(func,info,retry,wait) ({ func(info); 0; })
static inline void smp_send_reschedule(int cpu) { }
#define num_booting_cpus() 1
diff -puN net/core/flow.c~smp_call_function-must-be-a-macro net/core/flow.c
--- devel/net/core/flow.c~smp_call_function-must-be-a-macro 2005-11-23 00:17:40.000000000 -0800
+++ devel-akpm/net/core/flow.c 2005-11-23 00:17:47.000000000 -0800
@@ -292,7 +292,7 @@ void flow_cache_flush(void)
init_completion(&info.completion);
local_bh_disable();
- smp_call_function(flow_cache_flush_per_cpu, &info, 1, 0);
+ (void)smp_call_function(flow_cache_flush_per_cpu, &info, 1, 0);
flow_cache_flush_tasklet((unsigned long)&info);
local_bh_enable();
On Saturday 19 November 2005 22:40, Linus Torvalds wrote:
> There it is (or will soon be - the tar-ball and patches are still
> uploading, and mirroring can obviously take some time after that).
Something strange here. After a cg-update, I had no tag for rc2. Checking
showed no problems so I used cg-clone to get another copy of the repository.
Still no rc2.
ed@grover:/usr/src/2.6$ cg-version
cogito-0.16rc2 (73874dddeec2d0a8e5cd343eec762d98314def63)
ed@grover:/usr/src/2.6$ git --version
git version 0.99.9.GIT
cg-clone http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git 2.6
It looks to be the tag that is missing, gitk show commits after Nov 19.
Both git and cg were updated just prior to the cg-update (~Nov 22 8pm EST).
What is happening?
TIA
Ed Tomlinson
Ed Tomlinson wrote:
> Something strange here. After a cg-update, I had no tag for rc2. Checking
> showed no problems so I used cg-clone to get another copy of the repository.
> Still no rc2.
>
> ed@grover:/usr/src/2.6$ cg-version
> cogito-0.16rc2 (73874dddeec2d0a8e5cd343eec762d98314def63)
> ed@grover:/usr/src/2.6$ git --version
> git version 0.99.9.GIT
>
> cg-clone http://www.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git 2.6
>
This happened a while ago to someone else too. Apparently the http
transport needs serverside help (git-update-server-info or some such
must be run on the remote side).
Unless you're restricted by firewalls and other you could try
git clone
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git 2.6
which works flawlessly for me although it takes quite some time to
transfer all the data.
Linus, HPA: Are the packs cached on kernel.org? It seems to be at least
a minute before the transfers start.
--
Andreas Ericsson [email protected]
OP5 AB http://www.op5.se
Tel: +46 8-230225 Fax: +46 8-230231
On Thursday 24 November 2005 02:45, Hugh Dickins wrote:
>On Wed, 23 Nov 2005, Gene Heskett wrote:
>> On Wednesday 23 November 2005 18:40, Hugh Dickins wrote:
[snip previouus blowup]
>That's one of the things fixed by Andrew's patch below
>(though Linus fixed it differently in the end).
>Or you could just wait for 2.6.15-rc2-git4, should be along soon.
>
>Hugh
Got it, building src tree now, config'd & building.
And, I unchecked everything but what I need to run this card (I think,
whatdoIknow) and got this at depmod time:
WARNING:
/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol mt352_attach
WARNING:
/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol nxt200x_attach
WARNING:
/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol mt352_write
WARNING:
/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol lgdt330x_attach
WARNING:
/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
needs unknown symbol cx22702_attach
Maybe somebody can take the time to tell me what I do need to run a
pcHDTV-3000 in both ntsc and atsc modes using this newer code?
I was under the impression I needed the cx88 stuffs, ORV51132 (for
atsc) and nxt2002(for ntsc), but now we have lots of other dependencies
out the wazoo. Please clarify.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.36% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.
Gene Heskett wrote:
>Got it, building src tree now, config'd & building.
>
>And, I unchecked everything but what I need to run this card (I think,
>whatdoIknow) and got this at depmod time:
>WARNING:
>/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
>needs unknown symbol mt352_attach
>WARNING:
>/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
>needs unknown symbol nxt200x_attach
>WARNING:
>/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
>needs unknown symbol mt352_write
>WARNING:
>/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
>needs unknown symbol lgdt330x_attach
>WARNING:
>/lib/modules/2.6.15-rc2-git4/kernel/drivers/media/video/cx88/cx88-dvb.ko
>needs unknown symbol cx22702_attach
>
>Maybe somebody can take the time to tell me what I do need to run a
>pcHDTV-3000 in both ntsc and atsc modes using this newer code?
>I was under the impression I needed the cx88 stuffs, ORV51132 (for
>atsc) and nxt2002(for ntsc), but now we have lots of other dependencies
>out the wazoo. Please clarify.
>
>
Gene-
These other dependencies have always been there, except that nxt200x and
lgdt330x are relatively new frontends.
The difference is that a new Kconfig / Makefile feature is allowing us
to only select the specific frontend needed by your hardware...
Previously, all frontend support was forced to be built-in.
You and Adrian have clearly demonstrated that this frontend selection
capability isn't working properly. I think I will send Linus a patch to
restore previous functionality, forcing all frontends to be built...
Then I will resubmit a patch to Andrew that will re-enable this frontend
selection support, and I'll ask him to hold it in -mm until we can work
out the bugs.
The problem is that you are selecting cx88-dvb to be built-in to the
kernel (not as a module) , but the frontends are being built as modules
only. This is a problem.
You can make it all work, if you select the option to build support for
ALL FRONTENDS (this is the same as the older functionality)... This
option was selected by default in menuconfig. ... If you do not want to
have all frontends supported, then you should be fine if you recompile
the kernel again, but be sure to compile cx88 and cx88-dvb as MODULES
(m) ... and not in-kernel (y)
Regards,
Michael Krufky
On Thu, 24 Nov 2005, Ed Tomlinson wrote:
>
> What is happening?
The http transport isn't very good for git, so git adds various special
files to make it work at all. They need to be specially updated, and I
hadn't done that.
Using the native git protocol through git://git.kernel.org/.. gets around
it, as does using rsync.
I just repacked and updated it now, so how http should work too, although
inefficiently (because it will get a whole new pack - just one of the
disadvantages of the non-native protocols).
Linus
On Thu, 24 Nov 2005, Andreas Ericsson wrote:
>
> git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git 2.6
>
> which works flawlessly for me although it takes quite some time to transfer
> all the data.
The initial clone is very expensive for the native git protocol: the
protocol is designed to scale well for incremental updates (ie you have a
_huge_ repository that has changed just a bit, and the protocol should
work well for that), and that makes the initial clone quite expensive as
it marshalls the whole damn repository into this nice packed format.
So it's often nicer (certainly on the remote server) to use "rsync" for
the initial clone, and then only after that start using the git protocol.
(This is in no way really fundamental, and the server could cache the
packs it generates for initial clones, but that isn't implemented yet, and
probably won't be for some times).
Of course, especially if you're mostly bandwidth-constrained and the
server side is not under a big load, using the native git protocol may
actually be faster anyway. Because it's always going to generate the
nicest packing, while rsync:// will just use whatever packing that the
server happens to have at that point (but I do repack every few weeks, so
rsync for the initial clone should never be horribly bad - and since I
just repacked, it should get that "perfect" pack too).
Linus
diff -upr linux-2.6.15-rc2-git4/drivers/media/video/cx88/Kconfig linux/drivers/media/video/cx88/Kconfig
--- linux-2.6.15-rc2-git4/drivers/media/video/cx88/Kconfig 2005-11-24 13:47:02.148734660 -0500
+++ linux/drivers/media/video/cx88/Kconfig 2005-11-24 13:50:29.849190791 -0500
@@ -46,8 +46,8 @@ config VIDEO_CX88_DVB_ALL_FRONTENDS
If you are unsure, choose Y.
config VIDEO_CX88_DVB_MT352
- tristate "Zarlink MT352 DVB-T Support"
- default m
+ bool "Zarlink MT352 DVB-T Support"
+ default y
depends on VIDEO_CX88_DVB && !VIDEO_CX88_DVB_ALL_FRONTENDS
select DVB_MT352
---help---
@@ -55,8 +55,8 @@ config VIDEO_CX88_DVB_MT352
Connexant 2388x chip and the MT352 demodulator.
config VIDEO_CX88_DVB_OR51132
- tristate "OR51132 ATSC Support"
- default m
+ bool "OR51132 ATSC Support"
+ default y
depends on VIDEO_CX88_DVB && !VIDEO_CX88_DVB_ALL_FRONTENDS
select DVB_OR51132
---help---
@@ -64,8 +64,8 @@ config VIDEO_CX88_DVB_OR51132
Connexant 2388x chip and the OR51132 demodulator.
config VIDEO_CX88_DVB_CX22702
- tristate "Conexant CX22702 DVB-T Support"
- default m
+ bool "Conexant CX22702 DVB-T Support"
+ default y
depends on VIDEO_CX88_DVB && !VIDEO_CX88_DVB_ALL_FRONTENDS
select DVB_CX22702
---help---
@@ -73,8 +73,8 @@ config VIDEO_CX88_DVB_CX22702
Connexant 2388x chip and the CX22702 demodulator.
config VIDEO_CX88_DVB_LGDT330X
- tristate "LG Electronics DT3302/DT3303 ATSC Support"
- default m
+ bool "LG Electronics DT3302/DT3303 ATSC Support"
+ default y
depends on VIDEO_CX88_DVB && !VIDEO_CX88_DVB_ALL_FRONTENDS
select DVB_LGDT330X
---help---
@@ -82,8 +82,8 @@ config VIDEO_CX88_DVB_LGDT330X
Connexant 2388x chip and the LGDT3302/LGDT3303 demodulator.
config VIDEO_CX88_DVB_NXT200X
- tristate "NXT2002/NXT2004 ATSC Support"
- default m
+ bool "NXT2002/NXT2004 ATSC Support"
+ default y
depends on VIDEO_CX88_DVB && !VIDEO_CX88_DVB_ALL_FRONTENDS
select DVB_NXT200X
---help---
diff -upr linux-2.6.15-rc2-git4/drivers/media/video/Kconfig linux/drivers/media/video/Kconfig
--- linux-2.6.15-rc2-git4/drivers/media/video/Kconfig 2005-11-24 13:47:02.123743990 -0500
+++ linux/drivers/media/video/Kconfig 2005-11-24 13:52:00.525339321 -0500
@@ -26,7 +26,7 @@ config VIDEO_BT848
module will be called bttv.
config VIDEO_BT848_DVB
- tristate "DVB/ATSC Support for bt878 based TV cards"
+ bool "DVB/ATSC Support for bt878 based TV cards"
depends on VIDEO_BT848 && DVB_CORE
select DVB_BT8XX
---help---
diff -upr linux-2.6.15-rc2-git4/drivers/media/video/saa7134/Kconfig linux/drivers/media/video/saa7134/Kconfig
--- linux-2.6.15-rc2-git4/drivers/media/video/saa7134/Kconfig 2005-11-24 13:47:02.185720851 -0500
+++ linux/drivers/media/video/saa7134/Kconfig 2005-11-24 13:51:24.175908125 -0500
@@ -42,8 +42,8 @@ config VIDEO_SAA7134_DVB_ALL_FRONTENDS
If you are unsure, choose Y.
config VIDEO_SAA7134_DVB_MT352
- tristate "Zarlink MT352 DVB-T Support"
- default m
+ bool "Zarlink MT352 DVB-T Support"
+ default y
depends on VIDEO_SAA7134_DVB && !VIDEO_SAA7134_DVB_ALL_FRONTENDS
select DVB_MT352
---help---
@@ -51,8 +51,8 @@ config VIDEO_SAA7134_DVB_MT352
Philips saa7134 chip and the MT352 demodulator.
config VIDEO_SAA7134_DVB_TDA1004X
- tristate "Phillips TDA10045H/TDA10046H DVB-T Support"
- default m
+ bool "Phillips TDA10045H/TDA10046H DVB-T Support"
+ default y
depends on VIDEO_SAA7134_DVB && !VIDEO_SAA7134_DVB_ALL_FRONTENDS
select DVB_TDA1004X
---help---
@@ -60,8 +60,8 @@ config VIDEO_SAA7134_DVB_TDA1004X
Philips saa7134 chip and the TDA10045H/TDA10046H demodulator.
config VIDEO_SAA7134_DVB_NXT200X
- tristate "NXT2002/NXT2004 ATSC Support"
- default m
+ bool "NXT2002/NXT2004 ATSC Support"
+ default y
depends on VIDEO_SAA7134_DVB && !VIDEO_SAA7134_DVB_ALL_FRONTENDS
select DVB_NXT200X
---help---
Linus Torvalds <[email protected]> writes:
> (This is in no way really fundamental, and the server could cache the
> packs it generates for initial clones, but that isn't implemented yet, and
> probably won't be for some times).
Performance perceived by cloners is helped by
$ mkdir -p .git/pack-cache
$ git-rev-list --objects --all | git-pack-objects .git/pack-cache/pack
on the server side. This exact example of preparing by the
repository maintainer is optimizing for a wrong case, and I do
not think it is worth doing in practice, but this will give you
the lower bound when server side cache is implemented to do it
on demand.
On Thu, Nov 24, 2005 at 10:37:15AM -0800, Linus Torvalds wrote:
> I just repacked and updated it now, so how http should work too, although
> inefficiently (because it will get a whole new pack - just one of the
> disadvantages of the non-native protocols).
There's room to improve on that particular inefficiency. The http
commit walker could use Range: headers to fetch loose objects directly
from inside a pack if it didn't make sense to fetch the entire pack.
For this to work, pack fetches would need to be deferred until the
entire tree had been walked, and the commit walker could decide whether
to fetch the pack or loose objects based on the percentage of packed
objects it needed to fetch. It would also need to fetch all
tag/commit/tree objects using ranges to be able to fully walk the tree.
--
For a successful technology, reality must take precedence over public
relations, for nature cannot be fooled.
On Thu, 24 Nov 2005, Junio C Hamano wrote:
>
> Performance perceived by cloners is helped by
>
> $ mkdir -p .git/pack-cache
> $ git-rev-list --objects --all | git-pack-objects .git/pack-cache/pack
That really doesn't work very well. I push to that tree often several
times a day, and you'd have to re-do the cache each time.
So it would be much better if git-pack-objects would just always cache its
output in .git/pack-cache - along with some logic to just get rid of old
ones regularly.
Since git-pack-objects has to generate the pack _anyway_, it might as well
save it away when it does - so that if you have lots of people doing
clones or pulling, you'd only need to run it once for a particular set of
objects, and you'd not have to do any extra (or unnecessary) maintenance.
Linus
Linus Torvalds <[email protected]> writes:
> Since git-pack-objects has to generate the pack _anyway_, it might as well
> save it away when it does - so that if you have lots of people doing
> clones or pulling, you'd only need to run it once for a particular set of
> objects, and you'd not have to do any extra (or unnecessary) maintenance.
Caching itself is relatively easy (just implement an equivalent
of tee inside pack-objects ourselves). More problematic is
pruning. We could do it from cron based on atime _if_ the
filesystem is not mounted noatime but without arranging a
reasonably way for automated pruning this would become a disk
hog and extra maintenance burden, which is why I did not
implement the dynamic caching part in the initial round.
Since git-daemon would be the primary user of pack-cache/, this
implies a repository writable by git-daemon user on public
machine (not master), which is an extra thing to note.
On Thursday 24 November 2005 14:52, Nick Hengeveld wrote:
> On Thu, Nov 24, 2005 at 10:37:15AM -0800, Linus Torvalds wrote:
>
> > I just repacked and updated it now, so how http should work too, although
> > inefficiently (because it will get a whole new pack - just one of the
> > disadvantages of the non-native protocols).
>
> There's room to improve on that particular inefficiency. The http
> commit walker could use Range: headers to fetch loose objects directly
> from inside a pack if it didn't make sense to fetch the entire pack.
> For this to work, pack fetches would need to be deferred until the
> entire tree had been walked, and the commit walker could decide whether
> to fetch the pack or loose objects based on the percentage of packed
> objects it needed to fetch. It would also need to fetch all
> tag/commit/tree objects using ranges to be able to fully walk the tree.
Alternately, when creating a new archive the client could ask the server
what protocols are active. It could then use the best one for the clone and
update the .git/origin files with the optimal one for incremental pulls.
Thoughts?
Ed Tomlinson
Ed Tomlinson wrote:
> On Thursday 24 November 2005 14:52, Nick Hengeveld wrote:
>
>>On Thu, Nov 24, 2005 at 10:37:15AM -0800, Linus Torvalds wrote:
>>
>>
>>>I just repacked and updated it now, so how http should work too, although
>>>inefficiently (because it will get a whole new pack - just one of the
>>>disadvantages of the non-native protocols).
>>
>>There's room to improve on that particular inefficiency. The http
>>commit walker could use Range: headers to fetch loose objects directly
>>from inside a pack if it didn't make sense to fetch the entire pack.
>>For this to work, pack fetches would need to be deferred until the
>>entire tree had been walked, and the commit walker could decide whether
>>to fetch the pack or loose objects based on the percentage of packed
>>objects it needed to fetch. It would also need to fetch all
>>tag/commit/tree objects using ranges to be able to fully walk the tree.
>
>
> Alternately, when creating a new archive the client could ask the server
> what protocols are active. It could then use the best one for the clone and
> update the .git/origin files with the optimal one for incremental pulls.
>
This would only work with the git protocol, and since that's the fastest
protocol (theoretically that is, Pasky seems to have gotten other
figures but I'm not sure I believe those) it should really only ever
return itself which wouldn't make much sense.
--
Andreas Ericsson [email protected]
OP5 AB http://www.op5.se
Tel: +46 8-230225 Fax: +46 8-230231
Linus-
This patch corrects the build problems for cx88 and saa7134 hybrid
v4l/dvb drivers, described in the "Re: Linux 2.6.15-rc2" thread on LKML,
by Adrian and Gene. Please apply this to your -git tree.
Adrian Bunk wrote:
>configurations like CONFIG_VIDEO_CX88_DVB=y, CONFIG_DVB_CX22702=m are currently compile
>errors.
>
>
Gene Heskett wrote:
>*** Warning: "nxt200x_attach" [drivers/media/video/cx88/cx88-dvb.ko]
>undefined!
>
>
Thanks to Sam Ravnborg for pointing out a much needed correction in the
Makefile.