]> err.no Git - linux-2.6/log
linux-2.6
17 years ago[PATCH] fault-injection capability for kmalloc
Akinobu Mita [Fri, 8 Dec 2006 10:39:44 +0000 (02:39 -0800)]
[PATCH] fault-injection capability for kmalloc

This patch provides fault-injection capability for kmalloc.

Boot option:

failslab=<interval>,<probability>,<space>,<times>

<interval> -- specifies the interval of failures.

<probability> -- specifies how often it should fail in percent.

<space> -- specifies the size of free space where memory can be
   allocated safely in bytes.

<times> -- specifies how many times failures may happen at most.

Debugfs:

/debug/failslab/interval
/debug/failslab/probability
/debug/failslab/specifies
/debug/failslab/times
/debug/failslab/ignore-gfp-highmem
/debug/failslab/ignore-gfp-wait

Example:

failslab=10,100,0,-1

slab allocation (kmalloc(), kmem_cache_alloc(),..) fails once per 10 times.

Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] fault-injection capabilities infrastructure
Akinobu Mita [Fri, 8 Dec 2006 10:39:43 +0000 (02:39 -0800)]
[PATCH] fault-injection capabilities infrastructure

This patch provides base functions implement to fault-injection
capabilities.

- The function should_fail() is taken from failmalloc-1.0
  (http://www.nongnu.org/failmalloc/)

[akpm@osdl.org: cleanups, comments, add __init]
Cc: <okuji@enbug.org>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Don Mullis <dwm@meer.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] fault injection: documentation and scripts
Akinobu Mita [Fri, 8 Dec 2006 10:39:42 +0000 (02:39 -0800)]
[PATCH] fault injection: documentation and scripts

This patch set provides some fault-injection capabilities.

- kmalloc() failures

- alloc_pages() failures

- disk IO errors

We can see what really happens if those failures happen.

In order to enable these fault-injection capabilities:

1. Enable relevant config options (CONFIG_FAILSLAB, CONFIG_PAGE_ALLOC,
   CONFIG_MAKE_REQUEST) and if you want to configure them via debugfs,
   enable CONFIG_FAULT_INJECTION_DEBUG_FS.

2. Build and boot with this kernel

3. Configure fault-injection capabilities behavior by boot option or debugfs

   - Boot option

     failslab=
     fail_page_alloc=
     fail_make_request=

   - Debugfs

     /debug/failslab/*
     /debug/fail_page_alloc/*
     /debug/fail_make_request/*

   Please refer to the Documentation/fault-injection/fault-injection.txt
   for details.

4. See what really happens.

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Don Mullis <dwm@meer.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] nfsd: replace kmalloc+memset with kcalloc + simplify NULL check
Yan Burman [Fri, 8 Dec 2006 10:39:41 +0000 (02:39 -0800)]
[PATCH] nfsd: replace kmalloc+memset with kcalloc + simplify NULL check

Replace kmalloc+memset with kcalloc and simplify

Signed-off-by: Yan Burman <burman.yan@gmail.com>
Cc: Neil Brown <neilb@cse.unsw.edu.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] NFS3: Calculate 'w' a bit later in nfs3svc_encode_getaclres()
Jesper Juhl [Fri, 8 Dec 2006 10:39:40 +0000 (02:39 -0800)]
[PATCH] NFS3: Calculate 'w' a bit later in nfs3svc_encode_getaclres()

NFS3: Calculate 'w' a bit later in nfs3svc_encode_getaclres()
      This is a small performance optimization since we can return before
      needing 'w'. It also saves a few bytes of .text :
      Before:
           text    data     bss     dec     hex filename
           1632     140       0    1772     6ec fs/nfsd/nfs3acl.o
      After:
           text    data     bss     dec     hex filename
           1624     140       0    1764     6e4 fs/nfsd/nfs3acl.o

Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Cc: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] NFS2: Calculate 'w' a bit later in nfsaclsvc_encode_getaclres()
Jesper Juhl [Fri, 8 Dec 2006 10:39:39 +0000 (02:39 -0800)]
[PATCH] NFS2: Calculate 'w' a bit later in nfsaclsvc_encode_getaclres()

NFS2: Calculate 'w' a bit later in nfsaclsvc_encode_getaclres()
      This is a small performance optimization since we can return before
      needing 'w'. It also saves a few bytes of .text :
      Before:
           text    data     bss     dec     hex filename
           2406     212       0    2618     a3a fs/nfsd/nfs2acl.o
      After:
           text    data     bss     dec     hex filename
           2400     212       0    2612     a34 fs/nfsd/nfs2acl.o

Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Cc: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] lockdep: annotate nfsd4 recover code
Peter Zijlstra [Fri, 8 Dec 2006 10:39:38 +0000 (02:39 -0800)]
[PATCH] lockdep: annotate nfsd4 recover code

> =============================================
> [ INFO: possible recursive locking detected ]
> 2.6.18-1.2724.lockdepPAE #1
> ---------------------------------------------
> nfsd/6884 is trying to acquire lock:
>  (&inode->i_mutex){--..}, at: [<c04811e5>] vfs_rmdir+0x73/0xf4
>
> but task is already holding lock:
>  (&inode->i_mutex){--..}, at: [<f8dfa621>]
> nfsd4_clear_clid_dir+0x1f/0x3d [nfsd]
>
> other info that might help us debug this:
> 3 locks held by nfsd/6884:
>  #0:  (hash_sem){----}, at: [<f8de05eb>] nfsd+0x181/0x2ea [nfsd]
>  #1:  (client_mutex){--..}, at: [<f8df6d19>]
> nfsd4_setclientid_confirm+0x3b/0x2cf [nfsd]
>  #2:  (&inode->i_mutex){--..}, at: [<f8dfa621>]
> nfsd4_clear_clid_dir+0x1f/0x3d [nfsd]
>
> stack backtrace:
>  [<c040524d>] dump_trace+0x69/0x1af
>  [<c04053ab>] show_trace_log_lvl+0x18/0x2c
>  [<c040595f>] show_trace+0xf/0x11
>  [<c0405a53>] dump_stack+0x15/0x17
>  [<c043ca7a>] __lock_acquire+0x110/0x9b6
>  [<c043d91e>] lock_acquire+0x5c/0x7a
>  [<c061a41b>] __mutex_lock_slowpath+0xde/0x234
>  [<c04811e5>] vfs_rmdir+0x73/0xf4
>  [<f8dfa62b>] nfsd4_clear_clid_dir+0x29/0x3d [nfsd]
>  [<f8dfa733>] nfsd4_remove_clid_dir+0xb8/0xf8 [nfsd]
>  [<f8df6e90>] nfsd4_setclientid_confirm+0x1b2/0x2cf [nfsd]
>  [<f8def19a>] nfsd4_proc_compound+0x137a/0x166c [nfsd]
>  [<f8de00d5>] nfsd_dispatch+0xc5/0x180 [nfsd]
>  [<f8d09d83>] svc_process+0x3bd/0x631 [sunrpc]
>  [<f8de0604>] nfsd+0x19a/0x2ea [nfsd]
>  [<c0404e27>] kernel_thread_helper+0x7/0x10
> DWARF2 unwinder stuck at kernel_thread_helper+0x7/0x10
> Leftover inexact backtrace:
>  =======================

Some nesting annotation to the nfsd4 recovery code.
The vfs operations called will take dentry->d_inode->i_mutex.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] ISDN: fix warnings
Jeff Garzik [Fri, 8 Dec 2006 10:39:37 +0000 (02:39 -0800)]
[PATCH] ISDN: fix warnings

* diva, sedlbauer: the 'ready' label is only used in certain configurations
* hfc_pci:
- cast 'arg' to proper size for testing and printing
- print out 'void __iomem *' variables with %p,
  rather than using incorrect casts that throw warnings

Signed-off-by: Jeff Garzik <jeff@garzik.org>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] i4l: remove the broken HISAX_AMD7930 option
Adrian Bunk [Fri, 8 Dec 2006 10:39:36 +0000 (02:39 -0800)]
[PATCH] i4l: remove the broken HISAX_AMD7930 option

HISAX_AMD7930 was never anywhere near to being working, and this doesn't
seem to change in the forseeable future.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] isdn: replace kmalloc+memset with kzalloc
Burman Yan [Fri, 8 Dec 2006 10:39:35 +0000 (02:39 -0800)]
[PATCH] isdn: replace kmalloc+memset with kzalloc

Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] drivers/isdn/*: trivial vsnprintf() conversion
Alexey Dobriyan [Fri, 8 Dec 2006 10:39:34 +0000 (02:39 -0800)]
[PATCH] drivers/isdn/*: trivial vsnprintf() conversion

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Acked-by: Alan Cox <alan@redhat.com>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] ISDN: Avoid a potential NULL ptr deref in ippp
Jesper Juhl [Fri, 8 Dec 2006 10:39:34 +0000 (02:39 -0800)]
[PATCH] ISDN: Avoid a potential NULL ptr deref in ippp

There's a potential problem in isdn_ppp.c::isdn_ppp_decompress().
dev_alloc_skb() may fail and return NULL. If it does we will be passing a
NULL skb_out to ipc->decompress() and may also end up
dereferencing a NULL pointer at

    *proto = isdn_ppp_strip_proto(skb_out);

Correct this by testing 'skb_out' against NULL early and bail out.

Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] isdn: fix missing unregister_capi_driver
Akinobu Mita [Fri, 8 Dec 2006 10:39:33 +0000 (02:39 -0800)]
[PATCH] isdn: fix missing unregister_capi_driver

unregister_capi_driver() needs to be called in module cleanup.
(It fixes data corruption by reloading t1isa driver)

Cc: Kai Germaschewski <kai.germaschewski@gmx.de>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] drivers/isdn: Handcrafted MIN/MAX Macro removal
Amol Lad [Fri, 8 Dec 2006 10:39:32 +0000 (02:39 -0800)]
[PATCH] drivers/isdn: Handcrafted MIN/MAX Macro removal

Cleanups done to use min/max macros from kernel.h.  Handcrafted MIN/MAX
macros are changed to use macros in kernel.h

[akpm@osdl.org: warning fix]
Signed-off-by: Amol Lad <amol@verismonetworks.com>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: cyclades, fix warnings
Jiri Slaby [Fri, 8 Dec 2006 10:39:31 +0000 (02:39 -0800)]
[PATCH] Char: cyclades, fix warnings

fix gcc signed/unsigned warnings

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: cyclades, cleanup
Jiri Slaby [Fri, 8 Dec 2006 10:39:30 +0000 (02:39 -0800)]
[PATCH] Char: cyclades, cleanup

- remove cvs rcsid and alter code that uses it.
- allow a semicolon after use of macro to not confuse parsers (e.g. indent)
  by do {} while (0)
- JIFFIES_DIFF is simple subtraction, subtract directly
- returns cleanup -- do not put values in parenthesis and do not return nothing
  at the end of void functions
- comments are /* */ in C (not //)

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: cyclades, Lindent the code
Jiri Slaby [Fri, 8 Dec 2006 10:39:28 +0000 (02:39 -0800)]
[PATCH] Char: cyclades, Lindent the code

make the code indented by tabs.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: cyclades, save indent levels
Jiri Slaby [Fri, 8 Dec 2006 10:39:28 +0000 (02:39 -0800)]
[PATCH] Char: cyclades, save indent levels

Save 3 indent levels in interrupt routine by moving the code to a separate
function.  This needs to be done to allow Lindent do its work, since only 4
columns are used for indenting now and hence Lindent makes a big mess in the
code, when moves it 4*5 columns to the right.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, use mod_timer
Jiri Slaby [Fri, 8 Dec 2006 10:39:27 +0000 (02:39 -0800)]
[PATCH] Char: istallion, use mod_timer

Do not set expires by hand, use kernel helper, which also calls add_timer.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, dynamic tty device
Jiri Slaby [Fri, 8 Dec 2006 10:39:26 +0000 (02:39 -0800)]
[PATCH] Char: istallion, dynamic tty device

register tty device dynamically according to the count of board ports.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, change init sequence
Jiri Slaby [Fri, 8 Dec 2006 10:39:25 +0000 (02:39 -0800)]
[PATCH] Char: istallion, change init sequence

Reorganizate module init and exit and implement logic, when something fails in
these functions.  The former is needed for proper handling dynamic
tty_register_device.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, move init and exit code
Jiri Slaby [Fri, 8 Dec 2006 10:39:24 +0000 (02:39 -0800)]
[PATCH] Char: istallion, move init and exit code

Let's have these function at the end of the driver and expand stli_init
directly into module_init fucntion, since there is nothing other to have
there.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, fix enabling
Jiri Slaby [Fri, 8 Dec 2006 10:39:24 +0000 (02:39 -0800)]
[PATCH] Char: istallion, fix enabling

Enable ISA cards before pci_register_driver and then, enable each PCI card in
probe function.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, correct fail paths
Jiri Slaby [Fri, 8 Dec 2006 10:39:23 +0000 (02:39 -0800)]
[PATCH] Char: istallion, correct fail paths

Check more retvals and react somehow.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, free only isa
Jiri Slaby [Fri, 8 Dec 2006 10:39:22 +0000 (02:39 -0800)]
[PATCH] Char: istallion, free only isa

Only ISA cards should be freed in module exit.  Pci probed are freed in
pci_remove.  Define a flag, where we store this info a what to check against.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, brdnr locking
Jiri Slaby [Fri, 8 Dec 2006 10:39:21 +0000 (02:39 -0800)]
[PATCH] Char: istallion, brdnr locking

Kill possible race when getting brdnr by locking.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, ifdef eisa code
Jiri Slaby [Fri, 8 Dec 2006 10:39:20 +0000 (02:39 -0800)]
[PATCH] Char: istallion, ifdef eisa code

Disable compiling eisa stuff if STLI_EISAPROBE == 0.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, variables cleanup
Jiri Slaby [Fri, 8 Dec 2006 10:39:19 +0000 (02:39 -0800)]
[PATCH] Char: istallion, variables cleanup

- wipe gcc -W warnings by int -> uint conversion
- move 2 global variables into their local place

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, eliminate typedefs
Jiri Slaby [Fri, 8 Dec 2006 10:39:18 +0000 (02:39 -0800)]
[PATCH] Char: istallion, eliminate typedefs

Use only struct <name> instead of defining a new type <name_t>.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, remove the mess
Jiri Slaby [Fri, 8 Dec 2006 10:39:18 +0000 (02:39 -0800)]
[PATCH] Char: istallion, remove the mess

- remove unneeded license text
- remove functions, that are implemented in kernel -- call them (strtoul,
  min, tolower)
- do not cast NULL
- there is no static table, throw away code, which takes care of it --
  find module param cards in that place instead.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: istallion, convert to pci probing
Jiri Slaby [Fri, 8 Dec 2006 10:39:17 +0000 (02:39 -0800)]
[PATCH] Char: istallion, convert to pci probing

Use probing for pci devices.  Change some __inits to __devinits to use these
functions in probe function.  Create stli_cleanup_ports and move there cleanup
code from module_exit() code to not have duplicite cleanup code.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, use dynamic dev
Jiri Slaby [Fri, 8 Dec 2006 10:39:16 +0000 (02:39 -0800)]
[PATCH] Char: stallion, use dynamic dev

Use dynamic tty device registering depending on board's port count.  (i ->
retval change is relevant, since gcc complains about signedness of i)

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, variables cleanup
Jiri Slaby [Fri, 8 Dec 2006 10:39:15 +0000 (02:39 -0800)]
[PATCH] Char: stallion, variables cleanup

- fix `gcc -W' un/signed warnings by converting some ints -> uints.
- move 3 global variables into functions, where are they used.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, remove syntactic sugar
Jiri Slaby [Fri, 8 Dec 2006 10:39:14 +0000 (02:39 -0800)]
[PATCH] Char: stallion, remove syntactic sugar

Remove useless parenthesis and brackets.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, brd struct locking
Jiri Slaby [Fri, 8 Dec 2006 10:39:14 +0000 (02:39 -0800)]
[PATCH] Char: stallion, brd struct locking

Since assigning of stl_brds[brdnr] is racy, add locking to this critical
section.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, fix fail paths
Jiri Slaby [Fri, 8 Dec 2006 10:39:13 +0000 (02:39 -0800)]
[PATCH] Char: stallion, fix fail paths

Release everything what was allocated and check return value of isa probing.
Release only ISA boards in module exit, since pci have their own
pci-probing-remove.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, functions cleanup
Jiri Slaby [Fri, 8 Dec 2006 10:39:12 +0000 (02:39 -0800)]
[PATCH] Char: stallion, functions cleanup

Delete macros and functions, that are implemented in kernel yet (strtoul, min,
tolower).  Expand one function body in place, where it is called from.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, correct __init macros
Jiri Slaby [Fri, 8 Dec 2006 10:39:11 +0000 (02:39 -0800)]
[PATCH] Char: stallion, correct __init macros

Some functions are now called from pci probing functiuon which is __devinit,
not __init, correct this to not free functions after init if hotplug enabled.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, implement fail paths
Jiri Slaby [Fri, 8 Dec 2006 10:39:10 +0000 (02:39 -0800)]
[PATCH] Char: stallion, implement fail paths

This driver expect everything to work.  Implement fail paths logic to release
regions, irq hangler, memory...  if something is in bad state.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, prints cleanup
Jiri Slaby [Fri, 8 Dec 2006 10:39:09 +0000 (02:39 -0800)]
[PATCH] Char: stallion, prints cleanup

Too many information is printed out (they may be easily obtained through
sysfs), wipe them out in probe function.  Convert rest of them to dev_
variants.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, convert to pci probing
Jiri Slaby [Fri, 8 Dec 2006 10:39:09 +0000 (02:39 -0800)]
[PATCH] Char: stallion, convert to pci probing

Convert stallion driver to pci probing instead of pci_dev_get iteration.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, request regions
Jiri Slaby [Fri, 8 Dec 2006 10:39:08 +0000 (02:39 -0800)]
[PATCH] Char: sx, request regions

Check regions if they are free before we touch them.  Release them in
failpaths.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, use pci_iomap
Jiri Slaby [Fri, 8 Dec 2006 10:39:07 +0000 (02:39 -0800)]
[PATCH] Char: sx, use pci_iomap

Use pci_ friends for memory remapping of pci devices.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, fix return in module init
Jiri Slaby [Fri, 8 Dec 2006 10:39:06 +0000 (02:39 -0800)]
[PATCH] Char: sx, fix return in module init

If pci_register_driver fails, but eisa_driver_register doesn't, we don't call
misc_deregister, but returns error.  Return OK in such cases.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, simplify timer logic
Jiri Slaby [Fri, 8 Dec 2006 10:39:05 +0000 (02:39 -0800)]
[PATCH] Char: sx, simplify timer logic

Use kernel helpers for changing timer internals.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, whitespace cleanup
Jiri Slaby [Fri, 8 Dec 2006 10:39:04 +0000 (02:39 -0800)]
[PATCH] Char: sx, whitespace cleanup

Use Lindent to cleanup whitespace.  Wrap long lines by hand.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, remove duplicite code
Jiri Slaby [Fri, 8 Dec 2006 10:39:03 +0000 (02:39 -0800)]
[PATCH] Char: sx, remove duplicite code

sx_remove_code contents were used twice.  Call this function instead.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, lock boards struct
Jiri Slaby [Fri, 8 Dec 2006 10:39:03 +0000 (02:39 -0800)]
[PATCH] Char: sx, lock boards struct

Fix race condition which may occurs when multiple cards are probed at the same
time.  Add mutex to critical sections to avoid this situation.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, ifdef ISA code
Jiri Slaby [Fri, 8 Dec 2006 10:39:02 +0000 (02:39 -0800)]
[PATCH] Char: sx, ifdef ISA code

Most users won't need old probe code -- make it depndent on CONFIG_(E)ISA.

Cc: <R.E.Wolff@BitWizard.nl>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, use eisa probing
Jiri Slaby [Fri, 8 Dec 2006 10:39:01 +0000 (02:39 -0800)]
[PATCH] Char: sx, use eisa probing

Instead of finding eisa directly, use eisa probing.

Cc: <R.E.Wolff@BitWizard.nl>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, mark functions as devinit
Jiri Slaby [Fri, 8 Dec 2006 10:39:00 +0000 (02:39 -0800)]
[PATCH] Char: sx, mark functions as devinit

Mark as much as possible functions as __devinit to free them after driver
initialization (if no hotplug).

Cc: <R.E.Wolff@BitWizard.nl>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, use kcalloc
Jiri Slaby [Fri, 8 Dec 2006 10:38:59 +0000 (02:38 -0800)]
[PATCH] Char: sx, use kcalloc

Convert self-implemented kzalloc to kernel kcalloc.

Cc: <R.E.Wolff@BitWizard.nl>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: sx, convert to pci probing
Jiri Slaby [Fri, 8 Dec 2006 10:38:58 +0000 (02:38 -0800)]
[PATCH] Char: sx, convert to pci probing

Convert old pci code to pci probing.

Cc: <R.E.Wolff@BitWizard.nl>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, fix tty index check
Jiri Slaby [Fri, 8 Dec 2006 10:38:57 +0000 (02:38 -0800)]
[PATCH] Char: isicom, fix tty index check

Since tty->index is signed and may be < 0, we should assign this to int not
uint.  There is already a check to ensure if it is not negative, but gcc
complains with -W flag enabled and it is perfectly correct:
drivers/char/isicom.c:953: warning: comparison of unsigned expression < 0
is always false

Fix this issue by converting `line' variable from uint to int.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, remove cvs stuff
Jiri Slaby [Fri, 8 Dec 2006 10:38:56 +0000 (02:38 -0800)]
[PATCH] Char: isicom, remove cvs stuff

We don't need RCS_ID & co. Extract them from the code and define only one
macro -- SX_VERSION.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, simplify timer
Jiri Slaby [Fri, 8 Dec 2006 10:38:55 +0000 (02:38 -0800)]
[PATCH] Char: isicom, simplify timer

Don't init timer in such complicated way. Use DEFINE_TIMER and then only
mod_timer to reset the expiration.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, use completion
Jiri Slaby [Fri, 8 Dec 2006 10:38:55 +0000 (02:38 -0800)]
[PATCH] Char: isicom, use completion

Use wait_for_completion+complete instead of variables+msleep hack.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, check kmalloc retval
Jiri Slaby [Fri, 8 Dec 2006 10:38:54 +0000 (02:38 -0800)]
[PATCH] Char: isicom, check kmalloc retval

Value returned from kamlloc may be NULL, we should check if ENOMEM occured.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, use pci_request_region
Jiri Slaby [Fri, 8 Dec 2006 10:38:53 +0000 (02:38 -0800)]
[PATCH] Char: isicom, use pci_request_region

Use pci_request_region in pci probing function instead of request_region.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, move to tty_register_device
Jiri Slaby [Fri, 8 Dec 2006 10:38:52 +0000 (02:38 -0800)]
[PATCH] Char: isicom, move to tty_register_device

Instead of registering all devices in register_tty_driver, register devices in
probe function and register only port_count devices.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, remove unneeded memset
Jiri Slaby [Fri, 8 Dec 2006 10:38:51 +0000 (02:38 -0800)]
[PATCH] Char: isicom, remove unneeded memset

Memsetting of global static variables is not needed in the init code.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, remove isa code
Jiri Slaby [Fri, 8 Dec 2006 10:38:51 +0000 (02:38 -0800)]
[PATCH] Char: isicom, remove isa code

ISA is not supported by this driver, remove parts, that take care of this.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, rename init function
Jiri Slaby [Fri, 8 Dec 2006 10:38:50 +0000 (02:38 -0800)]
[PATCH] Char: isicom, rename init function

Rename init function from setup to init and mark it as __init, not __devinit.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: isicom, expand function
Jiri Slaby [Fri, 8 Dec 2006 10:38:49 +0000 (02:38 -0800)]
[PATCH] Char: isicom, expand function

Simple functions (those, that calls reuest_irq, request_region et al) may be
expanded directly in code.  It makes probe function more readable and
transparent.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] termios: Enable new style termios ioctls on x86-64
Alan Cox [Fri, 8 Dec 2006 10:38:48 +0000 (02:38 -0800)]
[PATCH] termios: Enable new style termios ioctls on x86-64

This turns on the split input/output speed features and arbitary baud rate
handling for the x86-64 platform.  Nothing should break if you use existing
standard speeds.  If you use the new speed stuff then you may see some
drivers failing to report the speed changes properly in error cases.  This
will be worked on further.  For the working cases this all seems happy.
I'll post a test suite used to test the basic stuff as well.

Patches for i386 will follow when I get a moment but are basically the
same.  If people could patch/test-suite other architectures and submit them
that would be great.

Signed-off-by: Alan Cox <alan@redhat.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] tty_ioctl: use termios for the old structure and termios2 for the new
Alan Cox [Fri, 8 Dec 2006 10:38:47 +0000 (02:38 -0800)]
[PATCH] tty_ioctl: use termios for the old structure and termios2 for the new

Having split out the user and kernel structures it turns out that some non
glibc C libraries pull their termios struct from the kernel headers
directly or indirectly.  This means we must keep "struct termios" as the
library sees it correct for the old ioctls.  Not a big problem just shuffle
the names and ifdef around a bit

[akpm@osdl.org: build fix]
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] tty: switch to ktermios
Alan Cox [Fri, 8 Dec 2006 10:38:45 +0000 (02:38 -0800)]
[PATCH] tty: switch to ktermios

This is the grungy swap all the occurrences in the right places patch that
goes with the updates.  At this point we have the same functionality as
before (except that sgttyb() returns speeds not zero) and are ready to
begin turning new stuff on providing nobody reports lots of bugs

If you are a tty driver author converting an out of tree driver the only
impact should be termios->ktermios name changes for the speed/property
setting functions from your upper layers.

If you are implementing your own TCGETS function before then your driver
was broken already and its about to get a whole lot more painful for you so
please fix it 8)

Also fill in c_ispeed/ospeed on init for most devices, although the current
code will do this for you anyway but I'd like eventually to lose that extra
paranoia

[akpm@osdl.org: bluetooth fix]
[mp3@de.ibm.com: sclp fix]
[mp3@de.ibm.com: warning fix for tty3270]
[hugh@veritas.com: fix tty_ioctl powerpc build]
[jdike@addtoit.com: uml: fix ->set_termios declaration]
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Martin Peschke <mp3@de.ibm.com>
Acked-by: Peter Oberparleiter <oberpar@de.ibm.com>
Cc: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] tty: switch to ktermios and new framework
Alan Cox [Fri, 8 Dec 2006 10:38:44 +0000 (02:38 -0800)]
[PATCH] tty: switch to ktermios and new framework

This is the core of the switch to the new framework.  I've split it from the
driver patches which are mostly search/replace and would encourage people to
give this one a good hard stare.

The references to BOTHER and ISHIFT are the termios values that must be
defined by a platform once it wants to turn on "new style" ioctl support.  The
code patches here ensure that providing

1. The termios overlays the ktermios in memory
2. The only new kernel only fields are c_ispeed/c_ospeed (or none)

the existing behaviour is retained.  This is true for the patches at this
point in time.

Future patches will define BOTHER, ISHIFT and enable newer termios structures
for each architecture, and once they are all done some of the ifdefs also
vanish.

[akpm@osdl.org: warning fix]
[akpm@osdl.org: IRDA fix]
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] tty: preparatory structures for termios revamp
Alan Cox [Fri, 8 Dec 2006 10:38:43 +0000 (02:38 -0800)]
[PATCH] tty: preparatory structures for termios revamp

In order to sort out our struct termios and add proper speed control we need
to separate the kernel and user termios structures.  Glibc is fine but the
other libraries rely on the kernel exported struct termios and we need to
extend this without breaking the ABI/API

To do so we add a struct ktermios which is the kernel view of a termios
structure and overlaps the struct termios with extra fields on the end for
now.  (That limitation will go away in later patches).  Some platforms (eg
alpha) planned ahead and thus use the same struct for both, others did not.

This just adds the structures but does not use them, it seems a sensible
splitting point for bisect if there are compile failures (not that I expect
them)

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, remove many prototypes
Jiri Slaby [Fri, 8 Dec 2006 10:38:42 +0000 (02:38 -0800)]
[PATCH] Char: stallion, remove many prototypes

Many prototypes are useless, since functions are declared before they are
called.  Also some code was easy to move to resolve this dependency and delete
prototypes.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, mark functions as init
Jiri Slaby [Fri, 8 Dec 2006 10:38:41 +0000 (02:38 -0800)]
[PATCH] Char: stallion, mark functions as init

Use __init macro for functions, that may be freed after initialization.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, uninline functions
Jiri Slaby [Fri, 8 Dec 2006 10:38:40 +0000 (02:38 -0800)]
[PATCH] Char: stallion, uninline functions

- Do not inline such long functions, it won't speed up anything.
- Remove prototypes of these functions.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, move init/deinit
Jiri Slaby [Fri, 8 Dec 2006 10:38:40 +0000 (02:38 -0800)]
[PATCH] Char: stallion, move init/deinit

- Move code from stl_init into module init function, because calling it was the
  only one thing, that it did.
- Move this code to the end of the driver (usual place for this) to resolve
  dependencies simply -- without prototypes.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, kill typedefs
Jiri Slaby [Fri, 8 Dec 2006 10:38:39 +0000 (02:38 -0800)]
[PATCH] Char: stallion, kill typedefs

Typedefs are considered ugly in the kernel. Eliminate them.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, remove unneeded casts
Jiri Slaby [Fri, 8 Dec 2006 10:38:38 +0000 (02:38 -0800)]
[PATCH] Char: stallion, remove unneeded casts

casts of NULL are unnecessary. And so casts to (void *) are.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: stallion, use pr_debug macro
Jiri Slaby [Fri, 8 Dec 2006 10:38:37 +0000 (02:38 -0800)]
[PATCH] Char: stallion, use pr_debug macro

Use pr_debug kernel macro instead of #ifdef DEBUG | printk() | #endif

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] mxser-workqueue-fixes
Andrew Morton [Fri, 8 Dec 2006 10:38:36 +0000 (02:38 -0800)]
[PATCH] mxser-workqueue-fixes

Cc: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] mxser_new: correct tty driver name
Jiri Slaby [Fri, 8 Dec 2006 10:38:35 +0000 (02:38 -0800)]
[PATCH] mxser_new: correct tty driver name

Mxser tty driver name should be ttyMI, not ttyM.  Correct this in mxser_new to
avoid conflicts with isicom driver, which is ttyM.

Acked-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] isicom, mxser MAINTAINERS update
Jiri Slaby [Fri, 8 Dec 2006 10:38:35 +0000 (02:38 -0800)]
[PATCH] isicom, mxser MAINTAINERS update

I can maintain moxa and isicom char drivers, because I've rewritten them to
the new API.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, clean macros
Jiri Slaby [Fri, 8 Dec 2006 10:38:34 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, clean macros

Celan redundant macros.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, pci probing
Jiri Slaby [Fri, 8 Dec 2006 10:38:33 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, pci probing

Finally, the intention of the patch serie: PCI probing for this driver.  add
pci_driver structure, (un)register it and add some stuff, which this needs.
Remove pdev pointer from board structure, because it's no longer needed.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, delete ttys and termios
Jiri Slaby [Fri, 8 Dec 2006 10:38:32 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, delete ttys and termios

- Driver uses global tty_struct array, which tries to assign to the
  tty_driver's ttys.
- the very same thing for termios
- the same for termios_locked

Kill such constructions.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, correct intr handler proto
Jiri Slaby [Fri, 8 Dec 2006 10:38:31 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, correct intr handler proto

Prototype of the driver's interrupt handler is old-fashioned, past changes
tell us, we won't get pt_regs as the 3rd argument.  There are only 2 params.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, comments cleanup
Jiri Slaby [Fri, 8 Dec 2006 10:38:31 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, comments cleanup

mxser_new, comments cleanup

- Remove commented code, since we have version control.
- Remove comments containing "following added by..." and "above added by,..".
  It's useless.
- Align other comments.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, reverse if-else-paths patch
Jiri Slaby [Fri, 8 Dec 2006 10:38:30 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, reverse if-else-paths patch

This patch was intorduced in mxser to 1.9.1 conversion, but causes endless
sleep in tcdrain userspace call.

Thanks to Sergei Organov <osv@javad.com> for testing and pointing this out.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, compact structures, round2
Jiri Slaby [Fri, 8 Dec 2006 10:38:29 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, compact structures, round2

Too many structures are in the driver for each type of card.  Make only one
for all cards and appropriate alter the code.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, register tty devices on the fly
Jiri Slaby [Fri, 8 Dec 2006 10:38:28 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, register tty devices on the fly

Register tty indexes only for real devices, udev then creates nodes for them
(and only for them).  Move tty_register_driver before probing, to be correct
when calling tty_register_device.  Also tell tty layer by tty_driver flags,
that we are registering devices.

Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, compress isa finding
Jiri Slaby [Fri, 8 Dec 2006 10:38:27 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, compress isa finding

ISA cards finding was too complex -- 2 (module params + predefined) absolutely
same routines, join them together with one for loop, one if and one indent
level.

Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, don't check tty_unregister retval
Jiri Slaby [Fri, 8 Dec 2006 10:38:26 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, don't check tty_unregister retval

Like other drivers silently unregister_tty_driver and put_tty_driver.  It
shouldn't be busy when module release function is called, since we are not
bsd, no refs shouldn't be held.

Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, correct fail paths
Jiri Slaby [Fri, 8 Dec 2006 10:38:25 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, correct fail paths

Resources were not released in some fail paths.  Correct this behaviour by
implementing function and calling it when something fails.

Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, testbit for bit testing
Jiri Slaby [Fri, 8 Dec 2006 10:38:25 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, testbit for bit testing

Use testbit like in tty subsystem for TTY_IO_ERROR testing.

Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, eliminate tty ldisc deref
Jiri Slaby [Fri, 8 Dec 2006 10:38:24 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, eliminate tty ldisc deref

Use tty_ldisc_flush and tty_wakeup helpers for accessing ldisc internals.

Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] char: remove unneded termbits redefinitions (mxser_new)
Jiri Slaby [Fri, 8 Dec 2006 10:38:23 +0000 (02:38 -0800)]
[PATCH] char: remove unneded termbits redefinitions (mxser_new)

No need to redefine termbits in char tree.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, CMSPAR is defined
Jiri Slaby [Fri, 8 Dec 2006 10:38:22 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, CMSPAR is defined

There is no need to have another (ifndeffed) definition of CMSPAR.  It's
defined in includes.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, code upside down
Jiri Slaby [Fri, 8 Dec 2006 10:38:21 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, code upside down

Reorder functions upside down not to have too many prototypes of each function
and have some order (similar to other drivers).

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, alter license terms
Jiri Slaby [Fri, 8 Dec 2006 10:38:20 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, alter license terms

- We don't need useless paragraphs in license terms.

- Add me to copyright.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, debug printk dependent on DEBUG
Jiri Slaby [Fri, 8 Dec 2006 10:38:20 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, debug printk dependent on DEBUG

Print some debug info only when DEBUG is enabled (use pr_debug macro).
Eliminate verbose kernel parameter which takes care of this.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, remove request for testers line
Jiri Slaby [Fri, 8 Dec 2006 10:38:19 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, remove request for testers line

Remove printk with info we are looking for a tester.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, revert spin_lock changes
Jiri Slaby [Fri, 8 Dec 2006 10:38:18 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, revert spin_lock changes

Some spinlock changes were introduced in 1.9.1 original moxa driver.  Revert
them, since they cause machine not responding.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
17 years ago[PATCH] Char: mxser_new, kill unneeded memsets
Jiri Slaby [Fri, 8 Dec 2006 10:38:17 +0000 (02:38 -0800)]
[PATCH] Char: mxser_new, kill unneeded memsets

mxser_new, kill unneeded memsets

There is no need to re-zero static global variables' memory, hence memsets
doing this are useless.  alloc_tty_struct also zeroes allocated memory:
another candidate for removing.

This fixes also a bug -- global structures are cleaned up after
initialization of some its parts.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>