Tons of stuff all over the map again. Lots of merging with various people.
The most noticeable fix (for me personally) was Trond fixing a nasty RPC
problem that caused the NFS client to return bogus dentry pointers that
hung the VFS layer hard on SMP machines.
But as you can see from the "short" changelog version (the full one is
63kB), there's a lot of other stuff there.
Linus
----
Summary of changes from v2.5.29 to v2.5.30
============================================
<[email protected]>:
o designated initializer patches for
<[email protected]>:
o Adaptec Starfire config fix
<[email protected]>:
o Make s390 and s390x print the right freed init memory size
Martin Dalecki <[email protected]>:
o 2.5.29 IDE 108 - 111
<[email protected]>:
o PPC32: clean up the initial mapping of RAM, allow for large-page
mappings
<[email protected]>:
o [PATCH[ 8139cp comment fix
<[email protected]>:
o [ARM PATCH] 1179/1: ldm/stm alignement fixups: treat 920T the same
as 922T The check in do_alignment_ldmstm for addr and eaddr being
the same is #ifdef'd
<[email protected]>:
o kconfig choice defaults 2 (3_3)
o 2.5: kconfig EXPERIMENTAL variant form
o 2.5: kconfig spurious EXPERIMENTAL 3 (2_2)
o 2.5: kconfig spurious bool default value (1_3)
o 2.5: kconfig spurious EXPERIMENTAL 3 (1_2)
o 2.5: kconfig missing EXPERIMENTAL 4 (1_4)
<[email protected]>:
o implement kmem_cache_size()
o update dqblk_xfs.h inclusion guards
o reparent scsi new-EH threads to init
o explicit signed char cast in i386 spin_is_locked
o silence APIC errors a bit
o documentation typos in
<[email protected]>:
o Fix typo in mm_slab.c
o Fix typo in net_ipv4_ipconfig.c
<[email protected]>:
o Typo in linux_net_sched_sch_ingress.c
o Typos in linux_arch_i386_kernel_io_apic.c
<[email protected]>:
o Small fix to assign continuous values to KEY_*
<[email protected]>:
o using ptr after kfree()
<[email protected]>:
o Bluetooth Subsystem PC Card drivers update
<[email protected]>:
o LSM file locking patch is bogus
<[email protected]>:
o sound/oss/trident.c [1/2] merge driver from 2.4-ac
o sound/oss/trident.c [2/2] remove cli/sti calls
<[email protected]>:
o fix Thread-Local Storage GDT access
<[email protected]>:
o 2.5 Trivial patch - 1400x1050 video mode added twice in 2.5.28
<[email protected]>:
o kbuild: Add new define do_cmd do_cmd is a nice shorthand when
creating rules that in one line shall adhere to KBUILD_VERBOSE and
make -s.
o docbook: parportbook dependencies and do_cmd
o kbuild: fix warnings in fixdep when compiled with -pedantic
<[email protected]>:
o Lindent drivers/char/nvram.c in anticipation of more patching
o clean up 'return (x);' style stuff into 'return x' in nvram.c
<[email protected]>:
o fix UP links - current bk tree
<[email protected]>:
o fs/locks.c: eliminate locks_unlock_delete
<[email protected]>:
o PAE compile fix
o remove declaration of __free_pte()
<[email protected]>:
o parport_serial/serial init dependencies
Adam J. Richter <[email protected]>:
o fix do_open() interaction with rd.c
Alan Cox <[email protected]>:
o 2.5.29 Fix pnpbios
o 2.5.29 Fix cmd640 config locking
o cs5530 IDE driver cli/sti fixes and cleanups
Alexander Viro <[email protected]>:
o Get rid of per-partition blk_size[]
o split "gendisk" to be per-disk, parts 1-2
Anders Gustafsson <[email protected]>:
o Add argument to synchronize_irq in cs46xx
Andrew Morton <[email protected]>:
o misc fixes
o use a slab cache for pte_chains
o show_free_areas() cleanup
o speed up pte_chain locking on uniprocessors
o optimise struct page layout
o for_each_pgdat macro
o for_each_zone macro
o strict overcommit
o use c99 initialisers in ext3
o direct IO updates
o permit modular build of raw driver
o put_page() uses audited
o restore lru_cache_del() in truncate_complete_page
o Re: BUG at rmap.c:212
Andries E. Brouwer <[email protected]>:
o partition fix
Anton Blanchard <[email protected]>:
o make cpu_relax a barrier on all architectures
o fix compile warning
o fix warning
Brad Hards <[email protected]>:
o Change the EVIOC?ABS ioctls to use structs rather than arrays of
ints
Christoph Hellwig <[email protected]>:
o VFS: implement sendfile file operation
o Remove kHTTPD
o Implement down_read_trylock() and down_write_trylock() and add a
generic spinlock implementation for downgrade_write().
o VM: remove unused /proc/sys/vm/kswapd and swapctl.h
Christopher Hoover <[email protected]>:
o [ARM PATCH] 1199/1: ISA Macros expected for platforms This makes
IRDA happy.
Dave Hansen <[email protected]>:
o fix e1000 after irq craziness
Dave Kleikamp <[email protected]>:
o Remove d_delete call from jfs_rmdir and jfs_unlink
David Brownell <[email protected]>:
o ohci, control/enumeration fix
David Howells <[email protected]>:
o export rwsem downgrade function
David Woodhouse <[email protected]>:
o Revert some bogus changes from mainline -- an over-excessive
search/replace and some BKL additions from someone who evidently
didn't bother to look at either the code they were modifying or the
documentation in README.Locking which explicitly states the BKL is
not needed.
o Remove licensing noise from individual JFFS2 files to refer to a
single 'LICENCE' file, which contains the new licence info. JFFS2
is no longer dual-licensed under RHEPL and GPL; eCos is now GPL'd
(with an exception for linking application) and hence so is JFFS2.
o JFFS2 update
o JFFS2 compile fixes for latest kernel
o Switch JFFS2 to C99 named initialisers
o Prevent crash after mounting JFFS2 images built with incorrect
erase size
o Remove cli() from R3964 line discipline
Eric Sandeen <[email protected]>:
o wan/sdla_chdlc.c oops fix
[email protected] <[email protected]>:
o Hi Vojtech, some superflous keyboard stuff came back with the last
PPC merge to Linus. This patch fixes that. Please apply.
Greg Kroah-Hartman <[email protected]>:
o Removed devfs_register_chrdev and devfs_unregister_chrdev
o Removed devfs_register_blkdev and devfs_unregister_blkdev
o Remove the devfs_should* functions I added, and replace them with
one devfs_only() call
Hugh Dickins <[email protected]>:
o shmem_file_write rounding VM_ACCT
o SHMEM_MAX_BYTES overflow checking
o mremap MAP_NORESERVE not in flags
o mmap MAP_NORESERVE not in vm_flags
o remove unhelpful vm_unacct_vma
o fix shared and private accounting
o update overcommit doc and comment
o shmem_file_setup when MAP_NORESERVE
o remove acct arg from do_munmap
o C99 initializers for mm
Ingo Molnar <[email protected]>:
o fix synchronize_irq() bug
o APM fixes, 2.5.29
o Re: Limit in set_thread_area
o sched-2.5.29-B1
o sanitize TLS API
Jeff Garzik <[email protected]>:
o alpha pid-reporting POSIX comformance bug fix
Jens Axboe <[email protected]>:
o fix REQ_QUEUED clearing in blk_insert_request()
o misc elevator/block updates
Kai Germaschewski <[email protected]>:
o kbuild: Fix compiling/installing as different users
o kbuild: Fix net/8022 selections
o kbuild: Make scripts/compile.h when sh != bash
o kbuild: Fix "export-objs"
o kbuild: Fix double vmlinux link
Linus Torvalds <[email protected]>:
o Move cmd640_lock outside the CONFIG_BLK_DEV_CMD640_ENHANCED test,
since it is needed regardless.
o Since "access_process_vm()" releases pages that can be in the page
cache, it needs to use page_cache_release() instead of plain
"put_page()".
o Make "cpu_relax()" imply a barrier, since that's how it is used.
o Cset exclude: [email protected]|ChangeSet|20020728030719|07783
o Leftover from trident cli/sti removal
o parport: fix warning - "flags" is unused after cli/sti removal
o Do x86 "kernel_thread()" explicitly by hand, instead of using a
system call from kernel space.
o We'd better BUG out inside the NFS code rather than return a bogus
dentry pointer to the VFS layer (which would oops on it inside the
dcache lock).
o Rename "sys_pread/pwrite" to "sys_pread64/pwrite64" to match the
actual implementation and avoid confusion.
o Update file lock security check to match the API change
o Chris Wright points out that this was also missed in the file
locking LSM update
o Split up "do_fork()" into "copy_process()" and "do_fork()"
o Undo d_drop removal at Trond's request Cset exclude:
[email protected]|ChangeSet|20020801011106|51286
o Fix missing semicolon from gendisk work
Matthew Dharm <[email protected]>:
o SCSI INQUIRY transfer length fix
o SCSI MODE_SENSE transfer length fix
Mikael Pettersson <[email protected]>:
o typo in fs_ufs_super.c:ufs_fill_super()
o drm_mga bitops -> long fix
o ftape bitops -> long fix
Patrick Mochel <[email protected]>:
o driverfs: use dentry->d_subdirs list instead of our own list when
removing directory
o driverfs: Do hashed lookup of dentry's when deleting a driverfs
file (instead of searching the list we keep)
o driverfs: don't do addition/deletion of driver_file_entry's into
local lists, as we don't use the lists anymore
o driverfs: Remove struct driver_dir_entry::files, since it's not
being used anymore
o driverfs: Remove references to struct driver_file_entry::dentry, as
we don't use it for anything useful anymore.
o driverfs: Don't put the driver_file_entry in struct
inode::u.generic_ip or struct file::private_data (since it's
already in struct dentry::d_fsdata and we always get to that)
o driverfs: use the parent directory's struct driver_dir_entry (in
struct dentry::d_fsdata) to access the struct device, rather than
via struct driver_file_entry::parent pointer.
o driverfs: Remove struct driver_file_entry::parent, as we can get to
it by the dentry
o driverfs: don't dynamically allocate and duplicate struct
driver_file_entry's any more Now that all unique information about
struct driver_file_entry's are gone (the dentry and parent
pointers), the data in them is shared among all users of the
entry. So, we don't have any reason to dynamically allocate and
duplicate the data anymore.
o driverfs: consolidate all the hashed lookups into a static helper:
get_dentry()
o device symlinks: just pass name, not struct driver_file_entry to
driverfs_create_symlink symlinks now only use the name field of the
struct driver_file_entry, so instead of allocating a new one each
time we want to create one, this changes the API to only accept the
name (since the driverfs core will never use the other fields
either)
o driverfs: Change the name of struct driver_file_entry to struct
device_attribute
o Fixup users of driverfs
o driverfs: add device_remove_symlink wrapper for removing symlinks
(since driverfs API is just about to change...)
o driverfs
o fixup users of device_remove_file to pass a struct
device_attribute, instead of a char *
o driverfs: Declare DEVICE_ATTR macro for initializing device
attributes (hide internal format of the structure)
o Convert users of struct device_attribute to initialize the structs
using DEVICE_ATTR macro.
o driverfs: Add struct attribute driverfs can only handle passing
struct device to read/write functions. In order to free it of this
limitation, we need a common data structure for driverfs to pass
around.
o driverfs: define struct driverfs_ops and remove struct device
dependencies
Paul Mackerras <[email protected]>:
o PPC32: Update to match the removal of global cli/sti etc
o PPC32: update hard/soft IRQ stuff along the lines of the i386 code
o PPC32: Simple mmu_gather implementation for now
o PPC32: remove unused code in ppc_htab.c
o PPC32: use VM_FAULT_* names instead of numbers
o PPC32: implement __downgrade_write for rwsems
o PPC32: add a comment explaining leap year calculation
o fix include/linux/timer.h compile
o page table page->index
Pavel Machek <[email protected]>:
o Move sleep_on() above refrigerator so that the kseriod thread in
serio.c doesn't exit on suspend because of a pending signal.
o swsusp: Vojtech pointed error in usb/hub.c
o swsusp: comment updates and warning fixes
Peter Osterlund <[email protected]>:
o Fix "make xconfig"
Petr Vandrovec <[email protected]>:
o Convert ncpfs to C99 initializers. By Rusty Trivial Russell
Richard Gooch <[email protected]>:
o do_mounts.c, block_dev.c, hiddev.c, md.c
o Removed deprecated devfs_find_handle()
Russell King <[email protected]>:
o [ARM] ARM keyboard code fixups for the input layer
o [SERIAL] Locking fixup (part 1) After the last few days of
debugging, we've ended up with the caller
o [SERIAL] Locking fixup (part 2) After the last few days of
debugging, we've ended up with the caller
o [ARM] Cleanup fallout from global IRQ changes
o [ARM] Remove {prepare,finish}_arch_{schedule,switch} These macros
are no longer required; the generic versions defined in sched.c are
sufficient for ARM.
o [ARM] Update ARM IRQ code for 2.5.28 global IRQ changes
irq_enter/irq_exit no longer take arguments. We also use the x86
methods for hardirq.h and softirq.h; they're sufficient for ARM.
o [ARM] Add security framework hooks (ptrace and syscall)
o [ARM] Optimise kernel->userspace exit code a little
o [ARM] Make ARM work with rmap ARM can use the generic rmap.h. It
doesn't need its own version.
o [ARM] Remove mach-types.h header file from asm/arch/hardware.h
Someone included mach-types.h into asm/arch/hardware.h, which
causes the whole kernel to needlessly rebuild when the machine type
file is updated. Move mach-types.h include into the files that
really need it.
o [ARM] Update SA11x0 cpufreq code to reflect CVS version
o [ARM] Prevent oops in free_pages() when freeing a pgd free_pages()
oopses if page->mapping is non-NULL. Ensure that any rmap
datastructures for the page are freed before freeing the concerned
page.
o [ARM] Two small changes Remove a couple of needless includes from
system3.c, and update mach-types file.
o [SERIAL] Remove some old compatibility cruft from 8250_pci.c
8250_pci.c contains some old compatibility cruft for when __devexit
wasn't defined by the generic kernel. It is now, so it's gone.
o [SERIAL] Add pci_disable_device() to initialisation failure paths
o [SERIAL] Add HP Diva PCI serial port support
o [SERIAL] Cleanup includes
Rusty Russell <[email protected]>:
o Fix ksoftirqd and migration threads initcalls
o Hot-plug CPU notifier warning fix
o CPU#1 not working with CONFIG_SMP=y, 2.5.28 OK
o Mark sparc32 unmaintained in 2.5
Tom Rini <[email protected]>:
o PPC32: Allow Motorola machines with PPCBUG to get their boot args
from NVRAM
o PPC32: Update the CPC700 code to be able to use all 32 IRQs
o PPC32: Change some instances of CONFIG_SERIAL_CONSOLE to
CONFIG_SERIAL_8250_CONSOLE
o PPC32: Update the Motorola LoPEC platform to work with CONFIG_VT
o Export synchronize_irq on CONFIG_SMP=y
Trond Myklebust <[email protected]>:
o Set PG_uptodate in nfs_writepage_sync()
o Support for cached lookups via readdirplus [1-6]
o NFS lookup() BKL imbalance
o Re: better getattr caching
o Fix brown paper bag race in RPC receive code
Vojtech Pavlik <[email protected]>:
o This simplifies the software autorepeat code in input/input.c, also
killing a race which could be the cause of autorepeat not stopping
after a key was released.
o Use stdint.h types instead of __u16 et al in input.h, to make life
easier for userspace people, as Brad Hards has suggested.
o Revert input.h back to kernel types
William Stinson <[email protected]>:
o region changes for rocket
o small region change for baycom_ser_hdx.c
o small region change for boardergo.c
There is a complie error:
gcc -Wp,-MD,./.nls_utf8.o.d -D__KERNEL__ -I/usr/src/linux-2.5.24/include
-Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer
-fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2
-march=i686 -nostdinc -iwithprefix include -DKBUILD_BASENAME=nls_utf8
-c -o nls_utf8.o nls_utf8.c
ld -m elf_i386 -r -o built-in.o nls_base.o nls_cp437.o nls_cp850.o
nls_cp852.o nls_iso8859-1.o nls_iso8859-2.o nls_utf8.o
make[2]: Leaving directory `/usr/src/linux-2.5.24/fs/nls'
make[2]: Entering directory `/usr/src/linux-2.5.24/fs/partitions'
gcc -Wp,-MD,./.check.o.d -D__KERNEL__
-I/usr/src/linux-2.5.24/include -Wall -Wstrict-prototypes
-Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing
-fno-common -pipe -mpreferred-stack-boundary=2 -march=i686
-nostdinc -iwithprefix include -DKBUILD_BASENAME=check
-DEXPORT_SYMTAB -c -o check.o check.c
check.c: In function `devfs_register_partitions':
check.c:470: array subscript is not an integer
make[2]: *** [check.o] Error 1
make[2]: Leaving directory `/usr/src/linux-2.5.24/fs/partitions'
make[1]: *** [partitions] Error 2
make[1]: Leaving directory `/usr/src/linux-2.5.24/fs'
make: *** [fs] Error 2
Greatings,
--
Banai
On Fri, 2 Aug 2002, Banai Zoltan wrote:
> There is a complie error:
> make[2]: Entering directory `/usr/src/linux-2.5.24/fs/partitions'
> gcc -Wp,-MD,./.check.o.d -D__KERNEL__
> -I/usr/src/linux-2.5.24/include -Wall -Wstrict-prototypes
> -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing
> -fno-common -pipe -mpreferred-stack-boundary=2 -march=i686
> -nostdinc -iwithprefix include -DKBUILD_BASENAME=check
> -DEXPORT_SYMTAB -c -o check.o check.c
> check.c: In function `devfs_register_partitions':
> check.c:470: array subscript is not an integer
Argh. My fault - it's devfs-only code and it didn't get tested ;-/
Fix: replace line 470 with
p[part].de = NULL;
On Thu, Aug 01, 2002 at 08:12:34PM -0400, Alexander Viro wrote:
>
> Argh. My fault - it's devfs-only code and it didn't get tested ;-/
>
> Fix: replace line 470 with
> p[part].de = NULL;
>
Thanks, that help!
But it does not boot,( nor does 2.5.24)
with 2.5.30 it panics at PNP BIOS initalisation,
without PNPBIOS it freezes after loop device init(no network card)
after network card init if configured (Intel e100).
No SysRq helps.:(
--
Banai Zoltan
Hi!
building 2.5.30 fails at serial driver 8250 in file 8250.c compiled as
module.
This is how it looks like (full log gzipped and attached):
make[2]: Entering directory /usr/src/linux-2.5.30/drivers/serial'
gcc -Wp,-MD,./.8250.o.d -D__KERNEL__ -I/usr/src/linux-2.5.30/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer
-fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2
-march=i686 -nostdinc -iwithprefix
include -DMODULE -include /usr/src/linux-2.5.30/include/linux/modversions.h
-DKBUILD_BASENAME=8250 -DEXPORT_SYMTAB -c -o 8250.o 8250.c
In file included from 8250.c:34:
/usr/src/linux-2.5.30/include/linux/serialP.h:50: field count' has
incomplete
type
8250.c:106: `ASYNC_BOOT_AUTOCONF' undeclared here (not in a function)
8250.c:106: `ASYNC_SKIP_TEST' undeclared here (not in a function)
8250.c:106: initializer element is not constant
8250.c:106: (near initialization for d_serial_port[0].flags')
8250.c:106: initializer element is not constant
8250.c:106: (near initialization for d_serial_port[0]')
8250.c:106: `ASYNC_BOOT_AUTOCONF' undeclared here (not in a function)
8250.c:106: `ASYNC_SKIP_TEST' undeclared here (not in a function)
8250.c:106: initializer element is not constant
8250.c:106: (near initialization for d_serial_port[1].flags')
8250.c:106: initializer element is not constant
8250.c:106: (near initialization for d_serial_port[1]')
8250.c:106: `ASYNC_BOOT_AUTOCONF' undeclared here (not in a function)
8250.c:106: `ASYNC_SKIP_TEST' undeclared here (not in a function)
8250.c:106: initializer element is not constant
8250.c:106: (near initialization for d_serial_port[2].flags')
8250.c:106: initializer element is not constant
...
and so on.
Regards,
Axel Siebenwirth
Banai Zoltan <[email protected]> wrote:
On Thu, Aug 01, 2002 at 08:12:34PM -0400, Alexander Viro wrote:
> >
> > Argh. My fault - it's devfs-only code and it didn't get tested ;-/
> >
> > Fix: replace line 470 with
> > p[part].de = NULL;
> >
> Thanks, that help!
>
> But it does not boot,( nor does 2.5.24)
> with 2.5.30 it panics at PNP BIOS initalisation,
> without PNPBIOS it freezes after loop device init(no network card)
> after network card init if configured (Intel e100).
> No SysRq helps.:(
Banai,
Yep, the anti-devfs regime broke it a few versions ago
with console/tty (serial) driver changes. It sort of defeats
the purpose, but I can boot with "devfs=nomount" as a
kernel boot up option. [This worked in lk 2.5.29]
Resistance may be futile, but it is the reason some of
us use linux.
Doug Gilbert
On Fri, 2 Aug 2002, Banai Zoltan wrote:
>
> But it does not boot,( nor does 2.5.24)
> with 2.5.30 it panics at PNP BIOS initalisation,
If the PNP BIOS panic is new (ie it didn't happen in 2.5.24), can you
write down the whole panic (and look up the symbols) and send that one to
Ingo Molnar <[email protected]>?
That would most likely be due to some of the GDT reorganizations that
happened for 2.5.30 due to the thread-local-storage patches.
Linus
The following trivial patches are still missing from the kernel:
- devfs patch to fix the problem with missing virtual consoles - only 0 in
/dev/vc: drivers/char/console.c
- patch to drivers/acpi/system.c, adding a missing include...
I believe it was interrupt.h
> The following trivial patches are still missing from the kernel:
>
> - devfs patch to fix the problem with missing virtual consoles - only 0 in
> /dev/vc: drivers/char/console.c
Try out my console patch at
http://www.transvirtual.com/~jsimmons/console.diff.gz
Please tell me if it solves the problem. I waiting for someone to say it
works then I will push my BK stuff to Linus.
> On Thu, Aug 01, 2002 at 08:12:34PM -0400, Alexander Viro wrote:
> > >
> > > Argh. My fault - it's devfs-only code and it didn't get tested ;-/
> > >
> > > Fix: replace line 470 with
> > > p[part].de = NULL;
> > >
> > Thanks, that help!
> >
> > But it does not boot,( nor does 2.5.24)
> > with 2.5.30 it panics at PNP BIOS initalisation,
> > without PNPBIOS it freezes after loop device init(no network card)
> > after network card init if configured (Intel e100).
> > No SysRq helps.:(
>
> Banai,
> Yep, the anti-devfs regime broke it a few versions ago
> with console/tty (serial) driver changes. It sort of defeats
> the purpose, but I can boot with "devfs=nomount" as a
> kernel boot up option. [This worked in lk 2.5.29]
I have nothing against devfs. In fact I spent the morning attempting to
get my machine to use devfs. I discovered that was harding than I thoyght
which is why I asked people to test my patch out. I think I'm going to
just send it into linus soon anyways.
On Fri, 2002-08-02 at 04:17, Linus Torvalds wrote:
> If the PNP BIOS panic is new (ie it didn't happen in 2.5.24), can you
> write down the whole panic (and look up the symbols) and send that one to
> Ingo Molnar <[email protected]>?
>
> That would most likely be due to some of the GDT reorganizations that
> happened for 2.5.30 due to the thread-local-storage patches.
The PnPBIOS gdt setup changes I did are wrong somewhere. If I can get
2.5.30 to actually boot I'll try and track it down. The traces I have
basically go
kernel -> kernel -> kernel -> pnpbios *BANG*
and appear to be jumping to the wrong physical address (ie gdt is
incorrect)
James Simmons wrote:
>
> > The following trivial patches are still missing from the kernel:
> >
> > - devfs patch to fix the problem with missing virtual consoles - only 0 in
> > /dev/vc: drivers/char/console.c
>
> Try out my console patch at
>
> http://www.transvirtual.com/~jsimmons/console.diff.gz
>
> Please tell me if it solves the problem. I waiting for someone to say it
> works then I will push my BK stuff to Linus.
The patch works for me. 2.5.30 with this and the compile-fix for devfs
comes up with /dev/vc/ populated normally.
Helge Hafting
> James Simmons wrote:
> >
> > > The following trivial patches are still missing from the kernel:
> > >
> > > - devfs patch to fix the problem with missing virtual consoles - only 0 in
> > > /dev/vc: drivers/char/console.c
> >
> > Try out my console patch at
> >
> > http://www.transvirtual.com/~jsimmons/console.diff.gz
> >
> > Please tell me if it solves the problem. I waiting for someone to say it
> > works then I will push my BK stuff to Linus.
>
> The patch works for me. 2.5.30 with this and the compile-fix for devfs
> comes up with /dev/vc/ populated normally.
Great. I will push my stuff ot Linus now :-)
On 2 Aug 2002, Alan Cox wrote:
> On Fri, 2002-08-02 at 04:17, Linus Torvalds wrote:
> > If the PNP BIOS panic is new (ie it didn't happen in 2.5.24), can you
> > write down the whole panic (and look up the symbols) and send that one to
> > Ingo Molnar <[email protected]>?
> >
> > That would most likely be due to some of the GDT reorganizations that
> > happened for 2.5.30 due to the thread-local-storage patches.
>
> The PnPBIOS gdt setup changes I did are wrong somewhere. If I can get
> 2.5.30 to actually boot I'll try and track it down. The traces I have
> basically go
> kernel -> kernel -> kernel -> pnpbios *BANG*
>
> and appear to be jumping to the wrong physical address (ie gdt is
> incorrect)
I have an open problem report on pnpbios, together with an oops, on
my status report page at http://members.cox.net/kernprobs/status.html
On Fri, Aug 02, 2002 at 03:25:17AM +0200, Axel Siebenwirth wrote:
> building 2.5.30 fails at serial driver 8250 in file 8250.c compiled as
> module.
Oddly, other people appear to be able to build it fine. Could you start
from a brand new 2.5.30 source and try again with the same configuration
please?
Thanks.
--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html
Hi Russell!
On Fri, 02 Aug 2002, Russell King wrote:
> Oddly, other people appear to be able to build it fine. Could you start
> from a brand new 2.5.30 source and try again with the same configuration
> please?
Ok. I'll do so. Maybe it's related to an odd configuration, all the debug
stuff...?
.config.gz attached.
Regards and many thanks,
Axel
On 2 Aug 2002, Alan Cox wrote:
>
> The PnPBIOS gdt setup changes I did are wrong somewhere.
Alan, in your PnP patches you seem to have changed the
"set_limit()"
to a
"_set_limit()"
which looks wrong (or at least doesn't look consistent with the notion of
just doing the same code as before, except on all CPU's).
It _looks_ to me like the QX_SET_SET() macros should be have the "_"
removed from the set_limit part. As it is, _set_limit() gets the address
calculations wrong (because you don't cast it to "char *") and also gets
the limit wrong (because you no longer do the page size adjustment).
Does it work with that small change? I have no idea about the pnpbios
code, I'm just looking at Alan's diff.
Linus
Linus Torvalds wrote:
> On 2 Aug 2002, Alan Cox wrote:
>
>>The PnPBIOS gdt setup changes I did are wrong somewhere.
>
>
> Alan, in your PnP patches you seem to have changed the
>
> "set_limit()"
>
> to a
>
> "_set_limit()"
>
> which looks wrong (or at least doesn't look consistent with the notion of
> just doing the same code as before, except on all CPU's).
>
> It _looks_ to me like the QX_SET_SET() macros should be have the "_"
> removed from the set_limit part. As it is, _set_limit() gets the address
> calculations wrong (because you don't cast it to "char *") and also gets
> the limit wrong (because you no longer do the page size adjustment).
>
> Does it work with that small change? I have no idea about the pnpbios
> code, I'm just looking at Alan's diff.
Applying the following change will make 2.5.30 get past the PNPBIOS load (at least on my box).
However, it still doesn't work (I'll post something later).
That said, here:
--- linux-2.5.30/drivers/pnp/pnpbios_core.c.orig Sat Aug 3 01:13:48 2002
+++ linux-2.5.30/drivers/pnp/pnpbios_core.c Sat Aug 3 01:13:20 2002
@@ -126,11 +126,11 @@
#define Q_SET_SEL(cpu, selname, address, size) \
set_base(cpu_gdt_table[cpu][(selname) >> 3], __va((u32)(address))); \
-_set_limit(&cpu_gdt_table[cpu][(selname) >> 3], size)
+set_limit(cpu_gdt_table[cpu][(selname) >> 3], size)
#define Q2_SET_SEL(cpu, selname, address, size) \
set_base(cpu_gdt_table[cpu][(selname) >> 3], (u32)(address)); \
-_set_limit((char *)&cpu_gdt_table[cpu][(selname) >> 3], size)
+set_limit(cpu_gdt_table[cpu][(selname) >> 3], size)
/*
* At some point we want to use this stack frame pointer to unwind
--Nathan