Import 1.3.54

CREDITS

@@ -187,6 +187,10 @@ S: 14509 NE 39th Street #1096
 S: Bellevue, Washington 98007
 S: USA
 
+N: Juan Jose Ciarlante
+E: jjciarla@raiz.uncu.edu.ar
+D: Network driver alias support
+
 N: Hamish Coleman
 E: hamish@zot.apana.org.au
 D: SEEQ8005 network driver
@@ -195,11 +199,13 @@ S: East Malvern, Victoria, 3145
 S: Australia
 
 N: Alan Cox
-E: iialan@iifeak.swan.ac.uk
-E: alan@cymru.net (use for commercial traffic)
-E: gw4pts@gw4pts.ampr.org
+E: alan@lxorguk.ukuu.org.uk (linux related - except big patches)
+E: iialan@www.linux.org.uk (linux.org.uk/big patches)
+E: alan@cymru.net (commercial CymruNET stuff)
+E: gw4pts@gw4pts.ampr.org (amateur radio stuff)
 E: GW4PTS@GB7SWN (packet radio)
-S: c/o I^2IT Limited
+E: Please don't use iialan@iifeak.swan.ac.uk for Linux stuff
+S: c/o 3Com/I^2IT Limited
 S: The Innovation Centre
 S: University Of Wales
 S: Swansea, SA2 8PP
@@ -209,6 +215,8 @@ D: Network layer debugging
 D: Initial AX.25 & IPX releases
 D: Original Linux netatalk patches.
 D: Current 3c501 hacker. >>More 3c501 info/tricks wanted<<.
+D: Watchdog timer drivers
+D: Linux/SMP
 
 N: Laurence Culhane
 E: loz@holmes.demon.co.uk
@@ -557,6 +565,7 @@ S: The Netherlands
 N: Volker Lendecke
 E: lendecke@namu01.gwdg.de
 D: Kernel smbfs (to mount WfW, NT and OS/2 network drives.)
+D: NCP filesystem support (to mount NetWare volumes)
 S: Innersteweg 11
 S: 37081 Goettingen
 S: Germany
@@ -799,6 +808,13 @@ N: Greg Page
 E: greg@caldera.com
 D: IPX development and support
 
+N: Avery Pennarun
+E: apenwarr@foxnet.net
+D: ARCnet driver, various small changes
+S: RR #5, 497 Pole Line Road
+S: Thunder Bay, Ontario
+S: CANADA P7C 5M9
+
 N: Kai Petzke
 E: wpp@marie.physik.tu-berlin.de
 D: Driver for Laser Magnetic Storage CD-ROM

Documentation/CodingStyle

@@ -177,9 +177,34 @@ typing - a infinite number of monkeys typing into GNU emacs would never
 make a good program). 
 
 So, you can either get rid of GNU emacs, or change it to use saner
-values.  I did the first, so don't ask me how to do the latter.  But
-even if you fail in getting emacs to do sane formatting, not everything
-is lost: use "indent". 
+values.  To do the latter, you can stick the following in your .emacs file:
+
+(defun linux-c-mode ()
+  "C mode with adjusted defaults for use with the Linux kernel."
+  (interactive)
+  (c-mode)
+  (setq c-indent-level 8)
+  (setq c-brace-imaginary-offset 0)
+  (setq c-brace-offset -8)
+  (setq c-argdecl-indent 8)
+  (setq c-label-offset -8)
+  (setq c-continued-statement-offset 8)
+  (setq indent-tabs-mode nil)
+  (setq tab-width 8))
+
+This will define the M-x linux-c-mode command.  When hacking on a
+module, if you put the string -*- linux-c -*- somewhere on the first
+two lines, this mode will be automatically invoked. Also, you may want
+to add
+
+(setq auto-mode-alist (cons '("/usr/src/linux.*/.*\\.[ch]$" . linux-c-mode)
+                       auto-mode-alist))
+
+to your .emacs file if you want to have linux-c-mode switched on
+automagically when you edit source files under /usr/src/linux.
+
+But even if you fail in getting emacs to do sane formatting, not
+everything is lost: use "indent".
 
 Now, again, GNU indent has the same brain dead settings that GNU emacs
 has, which is why you need to give it a few command line options. 

Documentation/Configure.help

@@ -177,6 +177,17 @@ CONFIG_NET
   consider updating your networking tools too; read net/README for
   details. 
 
+Network aliasing
+CONFIG_NET_ALIAS
+  This is for setting several network addresses on the same low-level network
+  device driver. Typically used for services that act different based
+  on the address they listen (eg. Apache httpd) or for connecting to
+  different logical networks through the same physical interface.
+  This is the generic part, later when configuring network protocol
+  options you will be asked for protocol-specific aliasing support.
+  See net/core/README.alias for more info.
+  If you need this features (for any protocol, like IP) say Y.  
+
 Sun floppy controller support
 CONFIG_BLK_DEV_SUNFD
   This is support for floppy drives on Sun Sparc workstations. Say Y
@@ -492,6 +503,12 @@ CONFIG_IP_MASQUERADE
   has nothing to do with the computer architecture of the same
   name. If you want this, say Y.
 
+IP: aliasing support
+CONFIG_IP_ALIAS
+  You need this to enable IP layer network aliasing. This will also
+  enable ARP resolution for alias devices. If you don't need several
+  IP addresses per interface, answer N.
+
 IP: multicast routing(in progress)
 CONFIG_IP_MROUTE
   This is used if you want your machine to act as a router for IP
@@ -787,9 +804,10 @@ CONFIG_SCSI_BUSLOGIC
   the documentation in drivers/scsi/README.BusLogic for more information.
   BusLogic FlashPoint SCSI Host Adapters are not supported by this driver.
   If this driver does not work correctly without modification, please
-  consult the author. This driver is not currently available as a module.
-  You might want to read the SCSI-HOWTO, available via ftp (user: 
-  anonymous) at sunsite.unc.edu:/pub/Linux/docs/HOWTO.
+  consult the author.  This driver may also be built as a module, but
+  only a single instance may be loaded.  You might also want to read
+  the SCSI-HOWTO, available via anonymous ftp from
+  sunsite.unc.edu:/pub/Linux/docs/HOWTO.
  
 EATA-DMA (DPT,NEC&ATT for ISA,EISA,PCI) support
 CONFIG_SCSI_EATA_DMA

Documentation/networking/alias.txt

+NET_ALIAS device aliasing v0.4x
+===============================
+  The main step taken in versions 0.40+ is the implementation of a
+  device aliasing mechanism that creates *actual* devices.
+  This development includes NET_ALIAS (generic aliasing) plus IP_ALIAS
+  (specific IP) support.
+
+Features
+--------
+o  ACTUAL alias devices created & inserted in dev chain
+o  AF_ independent: net_alias_type objects. Generic aliasing engine.
+o  AF_INET optimized
+o  hashed alias address lookup
+o  net_alias_type objs registration/unreg., module-ables.
+o  /proc/net/aliases & /proc/net/alias_types entries
+
+o IP alias implementation: static or runtime module.
+
+Usage (IP aliasing)
+-------------------
+  A very first step to test if you are running a net_alias-ed kernel
+  is to check /proc/net/aliases & /proc/net/alias_types entries:
+     # cat /proc/net/alias*
+
+  For IP aliasing you must have IP_ALIAS support included by
+  static linking ('y' to 2nd question above), or runtime module
+  insertion ('m' to 2nd q. above):
+     # insmod /usr/src/linux/modules/ip_alias.o (1.3.xx)
+     # insmod /usr/src/ip_alias/ip_alias.o      (1.2.xx) see above.
+
+o Alias creation.
+  Alias creation is done by 'magic' iface naming: eg. to create a
+  200.1.1.1 alias for eth0 ...
+  
+    # ifconfig eth0:0 200.1.1.1  etc,etc....
+                   ~~ -> request alias #0 creation (if it not exists) for eth0
+    and routing stuff also ...
+    # route add -host 200.1.1.1 dev eth0:0  (if same IP network as
+					    main device)
+   
+    # route add -net 200.1.1.0 dev eth0:0   (if completely new network wanted
+					    for eth0:0)
+
+o Alias deletion.
+  Also done by magic naming, eg:
+
+    # ifconfig eth0:0-  0  (maybe any address)
+                   ~~~ -> will delete alias (note '-' after dev name)
+  alias device is closed before deletion, so all network stuff that
+  points to it (routes, arp entries, ...) will be released.
+  		   		   
+Alias (re-)configuring
+  Aliases *are* devices, so you configure and refer to them as usual (ifconfig,
+  route, etc).
+  
+o Procfs entries
+  2 entries are added to help fetching alias runtime configuration:
+  a) /proc/net/alias_types
+     Will show you alias_types registered (ie. address families that
+     can be aliased).
+     eg. for IP aliasing with 1 alias configured:
+
+     # cat /proc/net/alias_types
+     type    name            n_attach
+     2       ip              1      
+     
+  b) /proc/net/aliases
+     Will show aliased devices info, eg (same as above):
+
+     # cat /proc/net/aliases
+     device           family address                                
+     eth0:0           2      200.1.1.1
+
+Relationship with main device
+-----------------------------
+  - On main device closing, all aliases will be closed and freed.
+  - Each new alias created is inserted in dev_chain just before next
+    main device (aliases get 'stacked' after main_dev), eg:
+      lo->eth0->eth0:0->eth0:2->eth1->0
+    If eth0 is unregistered, all it aliases will also be:
+      lo->eth1->0     
+
+Contact
+-------
+Please finger or e-mail me:
+   Juan Jose Ciarlante <jjciarla@raiz.uncu.edu.ar>
+   
+		   		
+; local variables:
+; mode: indented-text
+; mode: auto-fill
+; end:

Documentation/ramdisk.txt

+
+Using the RAM disk block device with Linux
+------------------------------------------
+
+Contents:
+
+	1) Overview
+	2) Kernel Command Line Parameters
+	3) Using "rdev -r" With New Kernels
+	4) An Example of Creating a Compressed ramdisk 
+
+
+1) Overview
+-----------
+
+As of kernel v1.3.48, the ramdisk driver was substantially changed.
+
+The older versions would grab a chunk of memory off the top before
+handing the remainder to the kernel at boot time. Thus a size parameter
+had to be specified via "ramdisk=1440" or "rdev -r /dev/fd0 1440" so
+that the driver knew how much memory to grab.
+
+Now the ramdisk dynamically grows as more space is required. It does
+this by using RAM from the buffer cache. The driver marks the buffers
+it is using with a new "BH_Protected" flag so that the kernel does 
+not try to reuse them later. This means that the old size parameter
+is no longer used, new command line parameters exist, and the behavior
+of the "rdev -r" or "ramsize" (usually a symbolic link to "rdev")
+command has changed.
+
+The old "ramdisk=<ram_size>" is now obsolete. The kernel will ignore
+such old commands, and thus they will be passed on through to the init
+program, which will then complain. You should remove any of these old
+style commands from config files such as "/etc/lilo.config".
+
+The new ramdisk also has the ability to load compressed ramdisk images,
+allowing one to squeeze more programs onto an average installation or 
+rescue floppy disk.
+
+Notes: You may have "dev/ram" or "/dev/ramdisk" or both. They are
+equivalent from the standpoint of this document. Also, the new ramdisk
+is a config option. When running "make config", make sure you enable
+ramdisk support for the kernel you intend to use the ramdisk with.
+
+
+2) Kernel Command Line Parameters
+---------------------------------
+
+	ramdisk_start=NNN
+	=================
+
+To allow a kernel image to reside on a floppy disk along with a compressed
+ramdisk image, the "ramdisk_start=<offset>" command was added. The kernel
+can't be included into the compressed ramdisk filesystem image, because
+it needs to be stored starting at block zero so that the BIOS can load the 
+bootsector and then the kernel can bootstrap itself to get going.
+
+Note: If you are using an uncompressed ramdisk image, then the kernel can
+be a part of the filesystem image that is being loaded into the ramdisk,
+and the floppy can be booted with LILO, or the two can be separate as
+is done for the compressed images.
+
+If you are using a two-disk boot/root setup (kernel on #1, ramdisk image
+on #2) then the ramdisk would start at block zero, and an offset of
+zero would be used. Since this is the default value, you would not need
+to actually use the command at all.
+
+If instead, you have a "zImage" of about 350k, and a "fs_image.gz" of
+say about 1MB, and you want them both on the same disk, then you
+would use an offset. If you stored the "fs_image.gz" onto the floppy
+starting at an offset of 400kB, you would use "ramdisk_start=400".
+
+
+	load_ramdisk=N
+	==============
+
+This parameter tells the kernel whether it is to try and load a
+ramdisk image or not. Specifying "load_ramdisk=1" will tell the
+kernel to load a floppy into the ramdisk. The default value is
+zero, meaning that the kernel should not try to load a ramdisk.
+
+
+	prompt_ramdisk=N
+	================
+
+This parameter tells the kernel whether or not to give you a prompt
+asking you to insert the floppy containing the ramdisk image. In
+a single floppy configuration the ramdisk image is on the same floppy
+as the kernel that just finished loading/booting and so a prompt
+is not needed. In this case one can use "prompt_ramdisk=0". In a
+two floppy configuration, you will need the chance to switch disks,
+and thus "prompt_ramdisk=1" can be used. Since this is the default 
+value, it doesn't really need to be specified.
+
+
+3) Using "rdev -r" With New Kernels
+-----------------------------------
+
+The usage of the word (two bytes) that "rdev -r" sets in the kernel image
+has changed. The low 11 bits (0 -> 10) specify an offset (in 1k blocks) 
+of up to 2MB (2^11) of where to find the ramdisk (this used to be the 
+size). Bit 14 indicates that a ramdisk is to be loaded, and bit 15
+indicates whether a prompt/wait sequence is to be given before trying
+to read the ramdisk. Since the ramdisk dynamically grows as data is
+being written into it, a size field is no longer required. Bits 11
+to 13 are not presently used and may as well be zero. These numbers
+are no magical secrets, as seen below:
+
+./arch/i386/kernel/setup.c:#define RAMDISK_IMAGE_START_MASK     0x07FF
+./arch/i386/kernel/setup.c:#define RAMDISK_PROMPT_FLAG          0x8000
+./arch/i386/kernel/setup.c:#define RAMDISK_LOAD_FLAG            0x4000
+
+Consider a typical two floppy disk setup, where you will have the 
+kernel on disk one, and have already put a ramdisk image onto disk #2.
+
+Hence you want to set bits 0 to 13 as zero, meaning that your ramdisk
+starts at an offset of zero kB from the beginning of the floppy.
+The command line equivalent is: "ramdisk_start=0"
+
+You want bit 14 as one, indicating that a ramdisk is to be loaded.
+The command line equivalent is: "load_ramdisk=1"
+
+You want bit 15 as one, indicating that you want a prompt/keypress
+sequence so that you have a chance to switch floppy disks.
+The command line equivalent is: "prompt_ramdisk=1"
+
+Putting that together gives 2^15 + 2^14 + 0 = 49152 for an rdev word.
+So to create disk one of the set, you would do:
+
+	/usr/src/linux# cat arch/i386/boot/zImage > /dev/fd0
+	/usr/src/linux# rdev /dev/fd0 /dev/fd0
+	/usr/src/linux# rdev -r /dev/fd0 49152
+
+If you make a boot disk that has LILO, then for the above, you would use:
+	append = "ramdisk_start=0 load_ramdisk=1 prompt_ramdisk=1"
+Since the default start = 0 and the default prompt = 1, you could use:
+	append = "load_ramdisk=1"
+
+
+4) An Example of Creating a Compressed ramdisk 
+----------------------------------------------
+
+To create a ramdisk image, you will need a spare block device to
+construct it on. This can be the ramdisk device itself, or an
+unused disk partition (such as an unmounted swap partition). For this 
+example, we will use the ramdisk device, "/dev/ram".
+
+Note: This technique should not be done on a machine with less than 8MB
+of RAM. If using a spare disk partition instead of /dev/ram, then this
+restriction does not apply.
+
+a) Decide on the ramdisk size that you want. Say 2MB for this example.
+   Create it by writing to the ramdisk device. (This step is not presently
+   required, but may be in the future.) It is wise to zero out the
+   area (esp. for disks) so that maximal compression is achieved for
+   the unused blocks of the image that you are about to create.
+
+	dd if=/dev/zero of=/dev/ram bs=1k count=2048
+
+b) Make a filesystem on it. Say ext2fs for this example.
+
+	mke2fs -vm0 /dev/ram 2048
+
+c) Mount it, copy the files you want to it (eg: /etc/* /dev/* ...)
+   and unmount it again.
+
+d) Compress the contents of the ramdisk. The level of compression
+   will be approximately 50% of the space used by the files. Unused
+   space on the ramdisk will compress to almost nothing.
+
+	dd if=/dev/ram bs=1k count=2048 | gzip -v9 > /tmp/ram_image.gz
+
+e) Put the kernel onto the floppy
+
+	cat zImage > /dev/fd0
+
+f) Put the ramdisk image onto the floppy, after the kernel. Use an offset
+   that is slightly larger than the kernel, so that you can put another
+   (possibly larger) kernel onto the same floppy later without overlapping
+   the ramdisk image. An offset of 400kB for kernels about 350kB in
+   size would be reasonable. Make sure offset+size of ram_image.gz is
+   not larger than the total space on your floppy (usually 1440kB).
+
+	dd if=/tmp/ram_image.gz of=/dev/fd0 bs=1k skip=400
+
+g) Use "rdev" to set the boot device, ramdisk offset, prompt flag, etc.
+   For ramdisk_start=400, load_ramdisk=1, ramdisk_start=400, one would
+   have 2^15 + 2^14 + 400 = 49552.
+
+	rdev /dev/fd0 /dev/fd0
+	rdev -r /dev/fd0 49552
+
+That is it. You now have your boot/root compressed ramdisk floppy. Some
+users may wish to combine steps (d) and (f) by using a pipe.
+
+--------------------------------------------------------------------------
+						Paul Gortmaker 12/95	

Makefile

 VERSION = 1
 PATCHLEVEL = 3
-SUBLEVEL = 53
+SUBLEVEL = 54
 
 ARCH = i386
 
@@ -249,7 +249,7 @@ MODFLAGS += -DMODVERSIONS -include $(HPATH)/linux/modversions.h
 endif
 
 modules: include/linux/version.h
-	@set -e; for i in $(SUBDIRS); do $(MAKE) -C $$i CFLAGS="-dc $(CFLAGS) $(MODFLAGS)" modules; done
+	@set -e; for i in $(SUBDIRS); do $(MAKE) -C $$i CFLAGS="$(CFLAGS) $(MODFLAGS)" modules; done
 
 modules_install:
 	@( \

README

@@ -71,6 +71,18 @@ INSTALLING the kernel:
    failed patches (xxx# or xxx.rej). If there are, either you or me has
    made a mistake.
 
+   Alternatively, the script patch-kernel can be used to automate this
+   process.  It determines the current kernel version and applies any
+   patches found.
+
+		cd /usr/src
+		linux/scripts/patch-kernel
+
+   The default directory for the kernel source is /usr/src/linux, but
+   can be specified as the first argument.  Patches are applied from
+   the current directory, but an alternative directory can be specified
+   as the second argument.
+
  - make sure your /usr/include/linux and /usr/include/asm directories
    are just symlinks to the kernel sources:
 

arch/alpha/defconfig

@@ -153,10 +153,10 @@ CONFIG_MSDOS_FS=y
 # CONFIG_UMSDOS_FS is not set
 CONFIG_PROC_FS=y
 # CONFIG_NFS_FS is not set
+# CONFIG_SMB_FS is not set
 CONFIG_ISO9660_FS=y
 # CONFIG_HPFS_FS is not set
 # CONFIG_SYSV_FS is not set
-# CONFIG_SMB_FS is not set
 
 #
 # character devices

arch/alpha/kernel/bios32.c

@@ -643,10 +643,13 @@ static inline void avanti_and_noname_fixup(void)
 					  PCI_INTERRUPT_LINE, dev->irq);
 #endif
 	}
-	/* now, set any level-triggered IRQs */
-	if (level_bits)
-		outw(level_bits, 0x4d0);
-
+	/*
+	 * Now, make all PCI interrupts level sensitive.  Notice:
+	 * these registers must be accessed byte-wise.  outw() doesn't
+	 * work, for some reason.
+	 */
+	outb((level_bits >> 0) & 0xff, 0x4d0);
+	outb((level_bits >> 8) & 0xff, 0x4d1);
 
 #if PCI_MODIFY
 	{

arch/alpha/mm/init.c

@@ -21,7 +21,6 @@
 #include <asm/pgtable.h>
 #include <asm/hwrpb.h>
 
-extern void scsi_mem_init(unsigned long);
 extern void die_if_kernel(char *,struct pt_regs *,long);
 extern void show_net_buffers(void);
 
@@ -155,11 +154,6 @@ void mem_init(unsigned long start_mem, unsigned long end_mem)
 		tmp += PAGE_SIZE;
 	}
 
-
-#ifdef CONFIG_SCSI
-	scsi_mem_init(high_memory);
-#endif
-
 	for (tmp = PAGE_OFFSET ; tmp < high_memory ; tmp += PAGE_SIZE) {
 		if (mem_map[MAP_NR(tmp)].reserved)
 			continue;

arch/i386/kernel/smp.c

@@ -14,6 +14,7 @@
  *		Felix Koop	:	NR_CPUS used properly
  *		Jose Renau	:	Handle single CPU case.
  *		Alan Cox	:	By repeated request 8) - Total BogoMIP report.
+ *		Greg Wright	:	Fix for kernel stacks panic.
  *
  */
 
@@ -42,7 +43,7 @@ int smp_threads_ready=0;				/* Set when the idlers are all forked 			*/
 volatile unsigned long cpu_callin_map[NR_CPUS] = {0,};	/* We always use 0 the rest is ready for parallel delivery */
 volatile unsigned long smp_invalidate_needed;		/* Used for the invalidate map thats also checked in the spinlock */
 struct cpuinfo_x86 cpu_data[NR_CPUS];			/* Per cpu bogomips and other parameters 		*/
-static unsigned int num_processors = 0;			/* Internal processor count				*/
+static unsigned int num_processors = 1;			/* Internal processor count				*/
 static unsigned long io_apic_addr = 0;			/* Address of the I/O apic (not yet used) 		*/
 unsigned char boot_cpu_id = 0;				/* Processor that is doing the boot up 			*/
 static unsigned char *kstack_base,*kstack_end;		/* Kernel stack list pointers 				*/
@@ -249,7 +250,6 @@ void smp_scan_config(unsigned long base, unsigned long length)
 {
 	unsigned long *bp=(unsigned long *)base;
 	struct intel_mp_floating *mpf;
-	num_processors = 1;		/* The boot processor */
 	
 /*	printk("Scan SMP from %p for %ld bytes.\n",
 		bp,length);*/

arch/i386/mm/init.c

@@ -28,7 +28,6 @@
 #undef USE_PENTIUM_MM
 #endif
 
-extern void scsi_mem_init(unsigned long);
 extern void die_if_kernel(char *,struct pt_regs *,long);
 extern void show_net_buffers(void);
 
@@ -232,9 +231,6 @@ void mem_init(unsigned long start_mem, unsigned long end_mem)
 		mem_map[MAP_NR(start_mem)].reserved = 0;
 		start_mem += PAGE_SIZE;
 	}
-#ifdef CONFIG_SCSI
-	scsi_mem_init(high_memory);
-#endif
 	for (tmp = 0 ; tmp < high_memory ; tmp += PAGE_SIZE) {
 		if (mem_map[MAP_NR(tmp)].reserved) {
 			if (tmp >= 0xA0000 && tmp < 0x100000)

arch/mips/mm/init.c

@@ -24,7 +24,6 @@
 #include <asm/pgtable.h>
 
 extern void deskstation_tyne_dma_init(void);
-extern void scsi_mem_init(unsigned long);
 extern void sound_mem_init(void);
 extern void die_if_kernel(char *,struct pt_regs *,long);
 extern void show_net_buffers(void);
@@ -258,9 +257,6 @@ void mem_init(unsigned long start_mem, unsigned long end_mem)
 #ifdef CONFIG_DESKSTATION_TYNE
 	deskstation_tyne_dma_init();
 #endif
-#ifdef CONFIG_SCSI
-	scsi_mem_init(high_memory);
-#endif
 #ifdef CONFIG_SOUND
 	sound_mem_init();
 #endif

arch/ppc/mm/init.c

@@ -45,9 +45,6 @@ extern unsigned long pg0[1024];		/* page table for 0-4MB for everybody */
 #ifdef CONFIG_DESKSTATION_TYNE
 extern void deskstation_tyne_dma_init(void);
 #endif
-#ifdef CONFIG_SCSI
-extern void scsi_mem_init(unsigned long);
-#endif
 #ifdef CONFIG_SOUND
 extern void sound_mem_init(void);
 #endif
@@ -225,9 +222,6 @@ printk("Mem init - Start: %x, End: %x\n", start_mem, high_memory);
 #ifdef CONFIG_DESKSTATION_TYNE
 	deskstation_tyne_dma_init();
 #endif
-#ifdef CONFIG_SCSI
-	scsi_mem_init(high_memory);
-#endif
 #ifdef CONFIG_SOUND
 	sound_mem_init();
 #endif

drivers/block/README.ide

@@ -391,6 +391,9 @@ by doing the following after installing slackware (or whatever):
 	program is run (like Norton "speeddisk"), it may move the Linux boot
 	files around, confusing LILO and making the (Linux) system unbootable.
 	Be sure to keep a kernel "boot floppy" at hand for such circumstances.
+	A possible workaround is to mark the Linux files as S+H+R (System,
+	Hidden, Readonly), to prevent most defragmentation programs from
+	moving the files around.
 
 If you "don't do DOS", then partition as you please, but remember to create
 a small partition to hold the /boot directory (and vmlinuz) as described above

drivers/block/floppy.c

@@ -272,7 +272,7 @@ static inline int DRIVE(kdev_t x) {
 /*
  * globals used by 'result()'
  */
-#define MAX_REPLIES 10
+#define MAX_REPLIES 17
 static unsigned char reply_buffer[MAX_REPLIES];
 static int inr; /* size of reply buffer, when called from interrupt */
 #define ST0 (reply_buffer[0])
@@ -3656,8 +3656,38 @@ static char get_fdc_version(void)
 		printk("FDC init: UNLOCK: unexpected return of %d bytes.\n", r);
 		return FDC_UNKNOWN;
 	}
-	printk("FDC %d is a post-1991 82077\n",fdc);
-	return FDC_82077;	/* Revised 82077AA passes all the tests */
+	output_byte(FD_PARTID);
+	r = result();
+	if (r != 1) {
+		printk("FDC init: PARTID: unexpected return of %d bytes.\n",r);
+		return FDC_UNKNOWN;
+	}
+	if (reply_buffer[0] == 0x80) {
+		printk("FDC %d is a post-1991 82077\n",fdc);
+		return FDC_82077;	/* Revised 82077AA passes all the tests */
+	}
+	switch (reply_buffer[0] >> 5) {
+	case 0x0:
+		output_byte(FD_SAVE);
+		r = result();
+		if (r != 17) {
+			printk("FDC init: SAVE: unexpected return of %d bytes.\n",r);
+			return FDC_UNKNOWN;
+		}
+		if (!(reply_buffer[0] & 0x40)) {
+			printk("FDC %d is a 3Volt 82078SL.\n",fdc);
+			return FDC_82078;
+		}
+		/* Either a 82078-1 or a 82078SL running at 5Volt */
+		printk("FDC %d is a 82078-1.\n",fdc);
+		return FDC_82078_1;
+	case 0x1:
+		printk("FDC %d is a 44pin 82078\n",fdc);
+		return FDC_82078;
+	default:
+		printk("FDC %d init: PARTID returned an unknown ID: %d.\n", fdc, reply_buffer[0] >> 5);
+		return FDC_UNKNOWN;
+	}
 } /* get_fdc_version */
 
 /* lilo configuration */

drivers/block/ide-tape.h

 /*
- * linux/drivers/block/ide-tape.h	Version 1.1 - ALPHA	Dec  14, 1995
+ * linux/drivers/block/ide-tape.h	Version 1.2 - ALPHA	Jan   1, 1996
  *
- * Copyright (C) 1995 Gadi Oxman <tgud@tochnapc2.technion.ac.il>
+ * Copyright (C) 1995, 1996 Gadi Oxman <tgud@tochnapc2.technion.ac.il>
  */
 
 /*
@@ -23,8 +23,7 @@
  *
  *	Pipelined operation mode has the potential to maximize the
  *	performance of the driver and thus to saturate the throughput
- *	to the maximum value supported by the tape. Currently, pipelined
- *	mode is supported only on writes.
+ *	to the maximum value supported by the tape.
  *
  *	In pipelined mode we are servicing requests without blocking the
  *	user backup program. For example, on a write request, we will add it
@@ -32,9 +31,8 @@
  *	user program will then have enough time to prepare the next blocks
  *	while the tape is still busy working on the previous requests.
  *
- *	Pipelined (write) operation mode is enabled by default, but since
- *	it has a few downfalls as well (Use of additional memory and deferred
- *      error code to the application), you may wish to disable it.
+ *	Pipelined operation mode is enabled by default, but since it has a
+ *	few downfalls as well, you may wish to disable it.
  *	Further explanation of pipelined mode is available in ide-tape.c .
  */
 
@@ -57,6 +55,29 @@
 #define	IDETAPE_MAX_PIPELINE_STAGES		200
 #define	IDETAPE_INCREASE_STAGES_RATE		0.2
 
+/*
+ *	Assuming the tape shares an interface with another device, the default
+ *	behavior is to service our pending pipeline requests as soon as
+ *	possible, but to gracefully postpone them in favor of the other device
+ *	when the tape is busy. This has the potential to maximize our
+ *	throughput and in the same time, to make efficient use of the IDE bus.
+ *
+ *	Note that when we transfer data to / from the tape, we co-operate with
+ *	the relatively fast tape buffers and the tape will perform the
+ *	actual media access in the background, without blocking the IDE
+ *	bus. This means that as long as the maximum IDE bus throughput is much
+ *	higher than the sum of our maximum throughput and the maximum
+ *	throughput of the other device, we should probably leave the default
+ *	behavior.
+ *
+ *	However, if it is still desired to give the other device a share even
+ *	in our own (small) bus bandwidth, you can set IDETAPE_LOW_TAPE_PRIORITY
+ *	to 1. This will let the other device finish *all* its pending requests
+ *	before we even check if we can service our next pending request.
+ */
+		 
+#define IDETAPE_LOW_TAPE_PRIORITY		0
+
 /*
  *	It seems that dynamically allocating buffers of about 32KB
  *	each is doomed to fail, unless we are in or very near the
@@ -76,21 +97,25 @@
 #endif
 
 /*
- *	Setting IDETAPE_DEBUG to 1 will:
- *
- *		1.	Generally log all driver actions.
- *		2.	Enable self-sanity checks in some places.
+ *	The following are used to debug the driver:
  *
- *	Use IDETAPE_DEBUG when encountering a problem with the driver.
+ *	Setting IDETAPE_DEBUG_LOG to 1 will log driver flow control.
+ *	Setting IDETAPE_DEBUG_BUGS to 1 will enable self-sanity checks in
+ *	some places.
  *
- *	Setting IDETAPE_DEBUG to 0 will restore normal operation mode:
+ *	Setting them to 0 will restore normal operation mode:
  *
  *		1.	Disable logging normal successful operations.
  *		2.	Disable self-sanity checks.
  *		3.	Errors will still be logged, of course.
+ *
+ *	All the #if DEBUG code will be removed some day, when the driver
+ *	is verified to be stable enough. This will make it much more
+ *	esthetic.
  */
  
-#define	IDETAPE_DEBUG		0
+#define	IDETAPE_DEBUG_LOG		0
+#define	IDETAPE_DEBUG_BUGS		1
 
 /*
  *	After each failed packet command we issue a request sense command
@@ -147,9 +172,12 @@
  *		request queue, so that ide.c will service requests from
  *		the other device on the same interface meanwhile.
  *
- *		The polling frequency is 1/IDETAPE_DSC_READ_WRITE_FREQUENCY,
- *		and it should be relatively fast. The default is a period
- *		of 50 msec.
+ *		We can now automatically select the "best" polling frequency.
+ *		Have a look at IDETAPE_ANTICIPATE_READ_WRITE_DSC below.
+ *
+ *		In case you don't want to use the automatic selection,
+ *		choose it to be relatively fast. The default fallback
+ *		frequency is 1/50 msec.
  *
  *	2.	After the successful initialization of a "media access
  *		packet command", which is a command which can take a long
@@ -173,7 +201,30 @@
  *
  */
 
-#define	IDETAPE_DSC_READ_WRITE_FREQUENCY	5*HZ/100	/* 50 msec */
+/*
+ *	Setting IDETAPE_ANTICIPATE_READ_WRITE_DSC to 1 will allow ide-tape
+ *	to cleverly select the lowest possible frequency which will
+ *	not affect performance, based on the tape parameters and our operation
+ *	mode. This has potential to dramatically decrease our polling load
+ *	on Linux.
+ *
+ *	However, for the cases in which our calculation fails, setting
+ *	the following option to 0 will force the use of the "fallback"
+ *	polling period defined below (defaults to 50 msec).
+ *
+ *	In any case, the frequency will be between the "lowest" value
+ *	to the "fallback" value, to ensure that our selected "best" frequency
+ *	is reasonable.
+ */
+
+#define IDETAPE_ANTICIPATE_READ_WRITE_DSC	1
+
+/*
+ *	DSC timings.
+ */
+ 
+#define	IDETAPE_DSC_READ_WRITE_FALLBACK_FREQUENCY   5*HZ/100	/* 50 msec */
+#define IDETAPE_DSC_READ_WRITE_LOWEST_FREQUENCY	30*HZ/100	/* 300 msec */
 #define	IDETAPE_DSC_FAST_MEDIA_ACCESS_FREQUENCY	1*HZ		/* 1 second */
 #define	IDETAPE_FAST_SLOW_THRESHOLD		5*60*HZ		/* 5 minutes */
 #define IDETAPE_DSC_SLOW_MEDIA_ACCESS_FREQUENCY	60*HZ		/* 1 minute */
@@ -184,14 +235,13 @@
 /*
  *	Definitions which are already needed in ide.h
  */
- 
+
 /*
- *	The following is currently not used.
+ *	Current character device data transfer direction.
  */
+  
+typedef enum {idetape_direction_none,idetape_direction_read,idetape_direction_write} chrdev_direction_t;
 
-typedef enum {no_excess_data,excess_data_read,excess_data_write} excess_data_status_t;
-
- 
 struct ide_drive_s;				/* Forward declaration - Will be defined later in ide.h */
 typedef void (idetape_pc_completed_t)(struct ide_drive_s *);
 
@@ -203,10 +253,9 @@ typedef struct idetape_packet_command_s {
 	byte c [12];				/* Actual packet bytes */
 	
 	byte retries;				/* On each retry, we increment retries */
-	byte error;				/* Set when an error occured */
-	byte active;				/* Set when a packet command is in progress */
-	byte wait_for_dsc;			/* 1 When polling for DSC */
-	byte dsc_count;		
+	byte error;				/* Error code */
+	byte abort;				/* Set when an error is considered normal - We won't retry */
+	byte wait_for_dsc;			/* 1 When polling for DSC on a media access command */
 	unsigned long request_transfer;		/* Bytes to transfer */
 	unsigned long actually_transferred; 	/* Bytes actually transferred */
 	unsigned long buffer_size;		/* Size of our data buffer */
@@ -353,12 +402,20 @@ typedef struct {
 	byte dsc_count;				/* We received DSC dsc_count times in a row */
 	unsigned long dsc_polling_start;	/* The time in which we started polling for DSC */
 	struct timer_list dsc_timer;		/* Timer used to poll for dsc */
+
+	/*
+	 *	We can now be much more clever in our selection of the
+	 *	read/write polling frequency. This is used along with
+	 *	the compile time option IDETAPE_ANTICIPATE_DSC.
+	 */
+ 
+	unsigned long best_dsc_rw_frequency;	/* Read/Write dsc polling frequency */
+
 	unsigned long dsc_polling_frequency;	/* The current polling frequency */
 	unsigned long dsc_timeout;		/* Maximum waiting time */
 	byte dsc_received;			/* Set when we receive DSC */
 
 	byte request_status;
-	byte request_dsc_callback;
 	byte last_status;			/* Contents of the tape status register */
 						/* before the current request (saved for us */
 						/* by ide.c) */
@@ -374,7 +431,7 @@ typedef struct {
 
 	/* Character device operation */
 
-	unsigned char last_dt_was_write;	/* Last character device data transfer was a write */
+	chrdev_direction_t chrdev_direction;	/* Current character device data transfer direction */
 	byte busy;				/* Device already opened */
 
 	/* Device information */
@@ -410,14 +467,19 @@ typedef struct {
 		
 	int current_number_of_stages;		/* Number of currently used stages */
 	int max_number_of_stages;		/* We will not allocate more than this number of stages */
-	idetape_pipeline_stage_t *first_stage;	/* Will be serviced after the currently active request */
-	idetape_pipeline_stage_t *last_stage;	/* New write requests will be added to the pipeline here */
-	int pipeline_was_full_once;		/* Set at the first time we fill the pipeline since the tape was opened */
+	idetape_pipeline_stage_t *first_stage;	/* The first stage which will be removed from the pipeline */
+	idetape_pipeline_stage_t *active_stage;	/* The currently active stage */
+	idetape_pipeline_stage_t *next_stage;	/* Will be serviced after the currently active request */
+	idetape_pipeline_stage_t *last_stage;	/* New requests will be added to the pipeline here */
 	int error_in_pipeline_stage;		/* Set when an error was detected in one of the pipeline stages */	
-	int pipeline_locked;			/* Against race conditions ... */
 	
 } idetape_tape_t;
 
+/*
+ *	The following is used to have a quick look at the tape's status
+ *	register between requests of the other device.
+ */
+ 
 #define POLL_HWIF_TAPE_DRIVE							\
 	if (hwif->tape_drive != NULL) {						\
 		if (hwif->tape_drive->tape.request_status) {			\
@@ -425,10 +487,6 @@ typedef struct {
 			hwif->tape_drive->tape.last_status=GET_STAT();		\
 			hwif->tape_drive->tape.request_status=0;		\
 		}								\
-		if (hwif->tape_drive->tape.request_dsc_callback) {		\
-			hwif->tape_drive->tape.request_dsc_callback=0;		\
-			idetape_put_back_postponed_request(hwif->tape_drive);	\
-		}								\
 	}
 
 #endif /* IDETAPE_H */

drivers/block/ide.c

 /*
- *  linux/drivers/block/ide.c	Version 5.23  Dec 15, 1995
+ *  linux/drivers/block/ide.c	Version 5.24  Jan 4, 1996
  *
- *  Copyright (C) 1994, 1995  Linus Torvalds & authors (see below)
+ *  Copyright (C) 1994-1996  Linus Torvalds & authors (see below)
  */
 
 /*
@@ -179,6 +179,7 @@
  *  Version 5.21	fix nasty cdrom/tape bug (ide_preempt was messed up)
  *  Version 5.22	fix ide_xlate_1024() to work with/without drive->id
  *  Version 5.23	miscellaneous touch-ups
+ *  Version 5.24	fix #if's for SUPPORT_CMD640
  *
  *  Driver compile-time options are in ide.h
  *
@@ -221,7 +222,7 @@
 
 #include "ide.h"
 
-#ifdef SUPPORT_CMD640
+#if SUPPORT_CMD640
 void cmd640_tune_drive(ide_drive_t *);
 static int cmd640_vlb = 0;
 #endif
@@ -2209,7 +2210,7 @@ static inline void do_identify (ide_drive_t *drive, byte cmd)
 			printk(", DMA");
 	}
 	printk("\n");
-#ifdef SUPPORT_CMD640
+#if SUPPORT_CMD640
 	cmd640_tune_drive(drive);	/* but can we tune a fish? */
 #endif
 }
@@ -2526,7 +2527,7 @@ static void init_qd6580 (void)
 #include "umc8672.c"	/* until we tidy up the interface some more */
 #endif
 
-#ifdef SUPPORT_CMD640
+#if SUPPORT_CMD640
 #include "cmd640.c"	/* until we tidy up the interface some more */
 #endif
 
@@ -3037,7 +3038,7 @@ int ide_init (void)
 	if (pcibios_present())
 		ide_init_pci ();
 #endif /* CONFIG_PCI */
-#ifdef SUPPORT_CMD640
+#if SUPPORT_CMD640
 	ide_probe_for_cmd640x();
 #endif
 

drivers/block/ide.h

@@ -598,13 +598,6 @@ void idetape_blkdev_release (struct inode *inode, struct file *filp, ide_drive_t
  
 void idetape_register_chrdev (void);
 
-/*
- *	The following function is called to re-insert a postponed tape
- *	request into the request queue.
- */
- 
-void idetape_put_back_postponed_request (ide_drive_t *drive);
-
 #endif /* CONFIG_BLK_DEV_IDETAPE */
 
 #ifdef CONFIG_BLK_DEV_TRITON

drivers/block/rd.c

@@ -25,6 +25,11 @@
  * been completely removed.
  *
  * Loadable module support added by Tom Dyas.
+ *
+ * Further cleanups by Chad Page (page0588@sundance.sjsu.edu):
+ *	Cosmetic changes in #ifdef MODULE, code movement, etc...
+ * 	When the ramdisk is rmmod'ed, free the protected buffers
+ * 	Default ramdisk size changed to 2.88MB
  */
 
 #include <linux/sched.h>
@@ -52,10 +57,15 @@ extern void wait_for_keypress(void);
 #define MAJOR_NR RAMDISK_MAJOR
 #include <linux/blk.h>
 
-#define BUILD_CRAMDISK
+/* These *should* be defined as parameters */
 #define NUM_RAMDISKS 8
+#define RD_DEFAULTSIZE	2880	/* 2.88 MB */
 
 #ifndef MODULE
+/* We don't have to load ramdisks or gunzip them in a module... */
+#define RD_LOADER
+#define BUILD_CRAMDISK
+
 void rd_load(void);
 static int crd_load(struct file *fp, struct file *outfp);
 #endif
@@ -106,10 +116,17 @@ static void rd_request(void)
 		goto repeat;
 	}
 
-	if (CURRENT->cmd == READ) {	
+	/*
+	 * If we're reading, fill the buffer with 0's.  This is okay since
+         * we're using protected buffers which should never get freed...
+	 *
+	 * If we're writing, we protect the buffer.
+  	 */
+
+	if (CURRENT->cmd == READ) 
 		memset(CURRENT->buffer, 0, len); 
-	}
-	set_bit(BH_Protected, &CURRENT->bh->b_state);
+	else	
+		set_bit(BH_Protected, &CURRENT->bh->b_state);
 
 	end_request(1);
 	goto repeat;
@@ -146,7 +163,6 @@ static int rd_ioctl(struct inode *inode, struct file *file, unsigned int cmd, un
 
 static int rd_open(struct inode * inode, struct file * filp)
 {
-
 	if (DEVICE_NR(inode->i_rdev) >= NUM_RAMDISKS)
 		return -ENODEV;
 
@@ -192,7 +208,7 @@ int rd_init(void)
 	blk_dev[MAJOR_NR].request_fn = &rd_request;
 
 	for (i = 0; i < NUM_RAMDISKS; i++) {
-		rd_length[i] = (16384 * 1024);
+		rd_length[i] = (RD_DEFAULTSIZE * 1024);
 		rd_blocksizes[i] = 1024;
 	}
 
@@ -201,7 +217,35 @@ int rd_init(void)
 	return 0;
 }
 
-#ifndef MODULE
+/* loadable module support */
+
+#ifdef MODULE
+
+int init_module(void)
+{
+	int error = rd_init();
+	if (!error)
+		printk(KERN_INFO "RAMDISK: Loaded as module.\n");
+	return error;
+}
+
+/* Before freeing the module, invalidate all of the protected buffers! */
+void cleanup_module(void)
+{
+	int i;
+
+	for (i = 0 ; i < NUM_RAMDISKS; i++)
+		invalidate_buffers(MKDEV(MAJOR_NR, i));
+
+	unregister_blkdev( MAJOR_NR, "ramdisk" );
+	blk_dev[MAJOR_NR].request_fn = 0;
+}
+
+#endif  /* MODULE */
+
+/* End of non-loading portions of the ramdisk driver */
+
+#ifdef RD_LOADER 
 /*
  * This routine tries to a ramdisk image to load, and returns the
  * number of blocks to read for a non-compressed image, 0 if the image
@@ -403,6 +447,7 @@ void rd_load()
 		infile.f_op->release(&inode, &infile);
 	set_fs(fs);
 }
+#endif /* RD_LOADER */
 
 #ifdef BUILD_CRAMDISK
 
@@ -546,25 +591,3 @@ crd_load(struct file * fp, struct file *outfp)
 
 #endif  /* BUILD_CRAMDISK */
 
-#endif  /* MODULE */
-
-
-/* loadable module support */
-
-#ifdef MODULE
-
-int init_module(void)
-{
-	int error = rd_init();
-	if (!error)
-		printk(KERN_INFO "RAMDISK: Loaded as module.\n");
-	return error;
-}
-
-void cleanup_module(void)
-{
-	unregister_blkdev( MAJOR_NR, "ramdisk" );
-	blk_dev[MAJOR_NR].request_fn = 0;
-}
-
-#endif  /* MODULE */

drivers/char/random.c

 /*
  * random.c -- A strong random number generator
  *
- * Version 0.95, last modified 4-Nov-95
+ * Version 0.96, last modified 29-Dec-95
  * 
  * Copyright Theodore Ts'o, 1994, 1995.  All rights reserved.
  *
@@ -276,7 +276,7 @@ void rand_initialize_irq(int irq)
 	}
 }
 
-void rand_initialize_blkdev(int major)
+void rand_initialize_blkdev(int major, int mode)
 {
 	struct timer_rand_state *state;
 	
@@ -287,7 +287,7 @@ void rand_initialize_blkdev(int major)
 	 * If kamlloc returns null, we just won't use that entropy
 	 * source.
 	 */
-	state = kmalloc(sizeof(struct timer_rand_state), GFP_KERNEL);
+	state = kmalloc(sizeof(struct timer_rand_state), mode);
 	if (state) {
 		blkdev_timer_state[major] = state;
 		memset(state, 0, sizeof(struct timer_rand_state));
@@ -437,9 +437,15 @@ void add_interrupt_randomness(int irq)
 
 void add_blkdev_randomness(int major)
 {
-	if (major >= MAX_BLKDEV || blkdev_timer_state[major] == 0)
+	if (major >= MAX_BLKDEV)
 		return;
 
+	if (blkdev_timer_state[major] == 0) {
+		rand_initialize_blkdev(major, GFP_ATOMIC);
+		if (blkdev_timer_state[major] == 0)
+			return;
+	}
+		
 	add_timer_randomness(&random_state, blkdev_timer_state[major],
 			     0x200+major);
 }

drivers/net/lance.c

@@ -976,36 +976,47 @@ lance_rx(struct device *dev)
 			if (status & 0x08) lp->stats.rx_crc_errors++;
 			if (status & 0x04) lp->stats.rx_fifo_errors++;
 			lp->rx_ring[entry].base &= 0x03ffffff;
-		} else {
+		}
+		else 
+		{
 			/* Malloc up new buffer, compatible with net-2e. */
 			short pkt_len = (lp->rx_ring[entry].msg_length & 0xfff)-4;
 			struct sk_buff *skb;
-
-			skb = dev_alloc_skb(pkt_len+2);
-			if (skb == NULL) {
-				printk("%s: Memory squeeze, deferring packet.\n", dev->name);
-				for (i=0; i < RX_RING_SIZE; i++)
-				  if (lp->rx_ring[(entry+i) & RX_RING_MOD_MASK].base < 0)
+			
+			if(pkt_len<60)
+			{
+				printk("%s: Runt packet!\n",dev->name);
+				lp->stats.rx_errors++;
+			}
+			else
+			{
+				skb = dev_alloc_skb(pkt_len+2);
+				if (skb == NULL) 
+				{
+					printk("%s: Memory squeeze, deferring packet.\n", dev->name);
+					for (i=0; i < RX_RING_SIZE; i++)
+						if (lp->rx_ring[(entry+i) & RX_RING_MOD_MASK].base < 0)
+							break;
+
+					if (i > RX_RING_SIZE -2) 
+					{
+						lp->stats.rx_dropped++;
+						lp->rx_ring[entry].base |= 0x80000000;
+						lp->cur_rx++;
+					}
 					break;
-
-				if (i > RX_RING_SIZE -2) {
-				  lp->stats.rx_dropped++;
-				  lp->rx_ring[entry].base |= 0x80000000;
-				  lp->cur_rx++;
 				}
-				break;
+				skb->dev = dev;
+				skb_reserve(skb,2);	/* 16 byte align */
+				skb_put(skb,pkt_len);	/* Make room */
+				eth_copy_and_sum(skb,
+					(unsigned char *)(lp->rx_ring[entry].base & 0x00ffffff),
+					pkt_len,0);
+				skb->protocol=eth_type_trans(skb,dev);
+				netif_rx(skb);
+				lp->stats.rx_packets++;
 			}
-			skb->dev = dev;
-			skb_reserve(skb,2);	/* 16 byte align */
-			skb_put(skb,pkt_len);	/* Make room */
-			eth_copy_and_sum(skb,
-				   (unsigned char *)(lp->rx_ring[entry].base & 0x00ffffff),
-				   pkt_len,0);
-			skb->protocol=eth_type_trans(skb,dev);
-			netif_rx(skb);
-			lp->stats.rx_packets++;
 		}
-
 		/* The docs say that the buffer length isn't touched, but Andrew Boyd
 		   of QNX reports that some revs of the 79C965 clear it. */
 		lp->rx_ring[entry].buf_length = -PKT_BUF_SZ;

drivers/scsi/BusLogic.h

 /*
 
-  Linux Driver for BusLogic SCSI Host Adapters
+  Linux Driver for BusLogic MultiMaster SCSI Host Adapters
 
   Copyright 1995 by Leonard N. Zubkoff <lnz@dandelion.com>
 
@@ -95,13 +95,11 @@ int BusLogic_BIOSDiskParameters(SCSI_Disk_T *, KernelDevice_T, int *);
 
 /*
   Define the number of Incoming and Outgoing Mailboxes used by this driver.
-  Since BusLogic Host Adapters have room to buffer 32 commands internally,
-  there is significant virtue in setting BusLogic_MailboxCount to 32 or above.
   The maximum possible value is 255, since the MailboxCount parameter to the
   Initialize Extended Mailbox command is limited to a single byte.
 */
 
-#define BusLogic_MailboxCount			32
+#define BusLogic_MailboxCount			64
 
 
 /*
@@ -124,11 +122,12 @@ int BusLogic_BIOSDiskParameters(SCSI_Disk_T *, KernelDevice_T, int *);
 
 /*
   Define the default number of Concurrent Commands per Logical Unit to allow
-  for Target Devices on non-ISA and ISA Host Adapters.
+  for Target Devices depending on whether or not ISA bounce buffers are
+  required.
 */
 
 #define BusLogic_Concurrency			7
-#define BusLogic_Concurrency_ISA		1
+#define BusLogic_Concurrency_BB			1
 
 
 /*
@@ -142,7 +141,14 @@ int BusLogic_BIOSDiskParameters(SCSI_Disk_T *, KernelDevice_T, int *);
 
 
 /*
-  Define the possible Tracing Options.
+  Define the possible Local Options.
+*/
+
+#define BusLogic_InhibitTargetInquiry		1
+
+
+/*
+  Define the possible Global Options.
 */
 
 #define BusLogic_TraceProbe			1
@@ -267,7 +273,7 @@ typedef enum
   BusLogic_InitializeExtendedMailbox =		0x81,	/* documented */
   BusLogic_InquireFirmwareVersion3rdDigit =	0x84,	/* undocumented */
   BusLogic_InquireFirmwareVersionLetter =	0x85,	/* undocumented */
-  BusLogic_InquireBoardModelAndRevision =	0x8B,	/* undocumented */
+  BusLogic_InquireBoardModelNumber	 =	0x8B,	/* undocumented */
   BusLogic_InquireSynchronousPeriod =		0x8C,	/* undocumented */
   BusLogic_InquireExtendedSetupInformation =	0x8D,	/* documented */
   BusLogic_EnableStrictRoundRobinMode =		0x8F,	/* documented */
@@ -395,15 +401,10 @@ typedef unsigned char BusLogic_FirmwareVersionLetter_T;
 
 
 /*
-  Define the Inquire Board Model and Revision reply structure.
+  Define the Inquire Board Model Number reply type.
 */
 
-typedef struct BusLogic_ModelAndRevision
-{
-  unsigned char Model[5];
-  unsigned char Revision;
-}
-BusLogic_ModelAndRevision_T;
+typedef unsigned char BusLogic_BoardModelNumber_T[5];
 
 
 /*
@@ -427,7 +428,7 @@ typedef struct BusLogic_ExtendedSetupInformation
   unsigned char MailboxCount;				/* Byte 4 */
   void *BaseMailboxAddress __attribute__ ((packed));	/* Bytes 5-8 */
   struct { unsigned char :6;				/* Byte 9 */
-	   boolean LevelTriggeredInterrupts:1;
+	   boolean LevelSensitiveInterrupts:1;
 	   unsigned char :1; } Misc;
   unsigned char FirmwareRevision[3];			/* Bytes 10-12 */
   boolean HostWideSCSI:1;				/* Byte 13 Bit 0 */
@@ -479,8 +480,8 @@ typedef unsigned char BusLogic_WideModeCCBRequest_T;
 
 /*
   Define the Requested Reply Length type used by the Inquire Setup Information,
-  Inquire Board Model and Revision, Inquire Synchronous Period, and Inquire
-  Extended Setup Information commands.
+  Inquire Board Model Number, Inquire Synchronous Period, and Inquire Extended
+  Setup Information commands.
 */
 
 typedef unsigned char BusLogic_RequestedReplyLength_T;
@@ -676,7 +677,7 @@ typedef struct BusLogic_CCB
   BusLogic_HostAdapterStatus_T HostAdapterStatus:8;	/* Byte 14 */
   BusLogic_TargetDeviceStatus_T TargetDeviceStatus:8;	/* Byte 15 */
   unsigned char TargetID;				/* Byte 16 */
-  unsigned char LogicalUnit:5;				/* Byte 17 Bits 0-2 */
+  unsigned char LogicalUnit:5;				/* Byte 17 Bits 0-4 */
   boolean TagEnable:1;					/* Byte 17 Bit 5 */
   BusLogic_QueueTag_T QueueTag:2;			/* Byte 17 Bits 6-7 */
   SCSI_CDB_T CDB;					/* Bytes 18-29 */
@@ -691,7 +692,8 @@ typedef struct BusLogic_CCB
   SCSI_Command_T *Command;
   enum { BusLogic_CCB_Free =	    0,
 	 BusLogic_CCB_Active =	    1,
-	 BusLogic_CCB_Completed =   2 } Status;
+	 BusLogic_CCB_Completed =   2,
+	 BusLogic_CCB_Reset =	    3 } Status;
   BusLogic_CompletionCode_T MailboxCompletionCode;
   unsigned int SerialNumber;
   struct BusLogic_CCB *Next;
@@ -759,6 +761,7 @@ typedef struct BusLogic_CommandLineEntry
   unsigned short IO_Address;
   unsigned short Concurrency;
   unsigned short BusSettleTime;
+  unsigned short LocalOptions;
   unsigned short TaggedQueuingPermitted;
   unsigned short TaggedQueuingPermittedMask;
   unsigned char ErrorRecoveryOption[BusLogic_MaxTargetIDs];
@@ -774,10 +777,10 @@ typedef struct BusLogic_HostAdapter
 {
   SCSI_Host_T *SCSI_Host;
   unsigned char HostNumber;
-  unsigned char ModelName[6];
+  unsigned char ModelName[9];
   unsigned char FirmwareVersion[6];
-  unsigned char BoardName[15];
-  unsigned char InterruptLabel[63];
+  unsigned char BoardName[18];
+  unsigned char InterruptLabel[62];
   unsigned short IO_Address;
   unsigned char IRQ_Channel;
   unsigned char DMA_Channel;
@@ -788,10 +791,11 @@ typedef struct BusLogic_HostAdapter
   boolean SynchronousInitiation:1;
   boolean ParityChecking:1;
   boolean ExtendedTranslation:1;
-  boolean LevelTriggeredInterrupts:1;
+  boolean LevelSensitiveInterrupts:1;
   boolean HostWideSCSI:1;
   boolean HostDifferentialSCSI:1;
   boolean HostAdapterResetPending:1;
+  boolean BounceBuffersRequired:1;
   volatile boolean HostAdapterCommandCompleted:1;
   unsigned short HostAdapterScatterGatherLimit;
   unsigned short DriverScatterGatherLimit;
@@ -799,6 +803,7 @@ typedef struct BusLogic_HostAdapter
   unsigned short MaxLogicalUnits;
   unsigned short Concurrency;
   unsigned short BusSettleTime;
+  unsigned short LocalOptions;
   unsigned short DisconnectPermitted;
   unsigned short TaggedQueuingPermitted;
   unsigned long BIOS_Address;
@@ -813,6 +818,8 @@ typedef struct BusLogic_HostAdapter
   unsigned char ErrorRecoveryOption[BusLogic_MaxTargetIDs];
   unsigned char CommandSuccessfulFlag[BusLogic_MaxTargetIDs];
   unsigned long ReadWriteOperationCount[BusLogic_MaxTargetIDs];
+  unsigned char QueuedOperationCount[BusLogic_MaxTargetIDs];
+  unsigned long LastSequencePoint[BusLogic_MaxTargetIDs];
   BusLogic_OutgoingMailbox_T *FirstOutgoingMailbox;
   BusLogic_OutgoingMailbox_T *LastOutgoingMailbox;
   BusLogic_OutgoingMailbox_T *NextOutgoingMailbox;
@@ -943,14 +950,17 @@ void BusLogic_StartMailboxScan(BusLogic_HostAdapter_T *HostAdapter)
 
 
 /*
-  BusLogic_Delay waits for Seconds to elapse.  It must be called with
-  interrupts enabled so that jiffies is updated.
+  BusLogic_Delay waits for Seconds to elapse.
 */
 
 static inline void BusLogic_Delay(int Seconds)
 {
   unsigned long TimeoutJiffies = jiffies + Seconds * HZ;
+  unsigned long ProcessorFlags;
+  save_flags(ProcessorFlags);
+  sti();
   while (jiffies < TimeoutJiffies) ;
+  restore_flags(ProcessorFlags);
 }
 
 

drivers/scsi/ChangeLog

-Fri Nov 10 15:29:49 1995  Leonard N. Zubkoff (lnz@dandelion.com)
+Sun Dec 31 23:26:00 1995  Leonard N. Zubkoff <lnz@dandelion.com>
+
+	* BusLogic Driver Version 1.3.1 Released.
+
+Fri Nov 10 15:29:49 1995  Leonard N. Zubkoff <lnz@dandelion.com>
 
 	* Released new BusLogic driver.
 

drivers/scsi/Config.in

@@ -18,7 +18,7 @@ dep_tristate 'Adaptec AHA152X support' CONFIG_SCSI_AHA152X $CONFIG_SCSI
 dep_tristate 'Adaptec AHA1542 support' CONFIG_SCSI_AHA1542 $CONFIG_SCSI
 dep_tristate 'Adaptec AHA1740 support' CONFIG_SCSI_AHA1740 $CONFIG_SCSI
 dep_tristate 'Adaptec AHA274X/284X/294X support' CONFIG_SCSI_AIC7XXX $CONFIG_SCSI
-bool 'BusLogic SCSI support' CONFIG_SCSI_BUSLOGIC
+dep_tristate 'BusLogic SCSI support' CONFIG_SCSI_BUSLOGIC
 dep_tristate 'EATA-DMA (DPT, NEC, ATT, Olivetti) support' CONFIG_SCSI_EATA_DMA $CONFIG_SCSI
 dep_tristate 'EATA-PIO (old DPT PM2001, PM2012A) support' CONFIG_SCSI_EATA_PIO $CONFIG_SCSI
 dep_tristate 'UltraStor 14F/34F support' CONFIG_SCSI_U14_34F $CONFIG_SCSI

drivers/scsi/README.BusLogic

-		     BusLogic SCSI Driver for Linux 1.3.41
-		       Version 1.3.0 ~ 13 November 1995
+	       BusLogic MultiMaster SCSI Driver for Linux 1.3.51
+		       Version 1.3.1 ~ 31 December 1995
 
 			       Leonard N. Zubkoff
 			       Dandelion Digital
@@ -15,7 +15,7 @@ adapters which share a common programming interface across a diverse collection
 of bus architectures by virtue of their MultiMaster ASIC technology.  This
 driver supports all present BusLogic MultiMaster Host Adapters, and should
 support any future MultiMaster designs with little or no modification.  Host
-adapters based on the new FlashPoint technology are not supported by this
+adapters based on the new FlashPoint architecture are not supported by this
 driver.
 
 My primary goals in writing this completely new BusLogic driver for Linux are
@@ -86,9 +86,10 @@ o Performance Features
   number of concurrent commands per logical unit is available from the kernel
   command line.  In addition, tagged queuing is automatically disabled whenever
   the host adapter firmware version is known not to implement it correctly, or
-  whenever a concurrency value of 1 is selected.  In performance testing,
-  sustained disk writes of 7.3MB per second have been observed to a /dev/sd
-  device.
+  whenever a concurrency value of 1 is selected.  Tagged queuing is also
+  disabled for individual target devices if disconnect/reconnect is disabled
+  for that device.  In performance testing, sustained disk writes of 7.3MB per
+  second have been observed to a /dev/sd device.
 
 o Robustness Features
 
@@ -106,7 +107,9 @@ o Robustness Features
   correct operation, the next command that is reset will force a full host
   adapter hard reset and SCSI bus reset.  SCSI bus resets caused by other
   devices and detected by the host adapter are also handled by issuing a hard
-  reset to the host adapter and full reinitialization.  These error recovery
+  reset to the host adapter and full reinitialization.  Finally, if a command
+  using tagged queuing causes a bus device reset or SCSI bus reset, then tagged
+  queuing will be disabled for that target device.  These error recovery
   options should improve overall system robustness by preventing individual
   errant devices from causing the system as a whole to lock up or crash, and
   thereby allowing a clean shutdown and restart after the offending component
@@ -126,7 +129,9 @@ o PCI Configuration Support
   On PCI systems running kernels compiled with PCI BIOS support enabled, this
   driver will interrogate the PCI configuration space and use the I/O port
   addresses assigned by the system BIOS, rather than the ISA compatible I/O
-  port addresses.
+  port addresses.  The ISA compatible I/O port address is then disabled by the
+  driver.  On PCI systems it is also recommended that the AutoSCSI utility be
+  used to disable the ISA compatible I/O port entirely as it is not necessary.
 
 o Shared Interrupts Support
 
@@ -156,34 +161,34 @@ that it is or will be supported.
 
 "C" Series Host Adapters:
 
-946C	PCI	Fast Single-ended SCSI-2
-956C	PCI	Fast/Wide Single-ended SCSI-2
-956CD	PCI	Fast/Wide Differential SCSI-2
-445C	VLB	Fast Single-ended SCSI-2
-747C	EISA	Fast Single-ended SCSI-2
-757C	EISA	Fast/Wide Single-ended SCSI-2
-757CD	EISA	Fast/Wide Differential SCSI-2
-545C	ISA	Fast Single-ended SCSI-2
-540CF	ISA	Fast Single-ended SCSI-2
+BT-946C	    PCI	    Fast Single-ended SCSI-2
+BT-956C	    PCI	    Fast/Wide Single-ended SCSI-2
+BT-956CD    PCI	    Fast/Wide Differential SCSI-2
+BT-445C	    VLB	    Fast Single-ended SCSI-2
+BT-747C	    EISA    Fast Single-ended SCSI-2
+BT-757C	    EISA    Fast/Wide Single-ended SCSI-2
+BT-757CD    EISA    Fast/Wide Differential SCSI-2
+BT-545C	    ISA	    Fast Single-ended SCSI-2
+BT-540CF    ISA	    Fast Single-ended SCSI-2
 
 "S" Series Host Adapters:
 
-445S	VLB	Fast Single-ended SCSI-2
-747S	EISA	Fast Single-ended SCSI-2
-747D	EISA	Fast Differential SCSI-2
-757S	EISA	Fast/Wide Single-ended SCSI-2
-757D	EISA	Fast/Wide Differential SCSI-2
-545S	ISA	Fast Single-ended SCSI-2
-542D	ISA	Fast Differential SCSI-2
-742A 	EISA	Single-ended SCSI-2 (742A revision H)
-542B	ISA	Single-ended SCSI-2 (542B revision H)
+BT-445S	    VLB	    Fast Single-ended SCSI-2
+BT-747S	    EISA    Fast Single-ended SCSI-2
+BT-747D	    EISA    Fast Differential SCSI-2
+BT-757S	    EISA    Fast/Wide Single-ended SCSI-2
+BT-757D	    EISA    Fast/Wide Differential SCSI-2
+BT-545S	    ISA	    Fast Single-ended SCSI-2
+BT-542D	    ISA	    Fast Differential SCSI-2
+BT-742A	    EISA    Single-ended SCSI-2 (742A revision H)
+BT-542B	    ISA	    Single-ended SCSI-2 (542B revision H)
 
 "A" Series Host Adapters:
 
-742A 	EISA	Single-ended SCSI-2 (742A revisions A - G)
-542B	ISA	Single-ended SCSI-2 (542B revisions A - G)
+BT-742A	    EISA    Single-ended SCSI-2 (742A revisions A - G)
+BT-542B	    ISA	    Single-ended SCSI-2 (542B revisions A - G)
 
-The FlashPoint LT, also known as the 930 Ultra, implements a different host
+The FlashPoint LT, also known as the BT-930 Ultra, implements a different host
 interface and is not supported by this driver.
 
 AMI FastDisk Host Adapters are true BusLogic clones and are supported by this
@@ -235,18 +240,16 @@ substantially impact performance.
 
 				 INSTALLATION
 
-This distribution was prepared for Linux kernel version 1.3.41.  Installation
+This distribution was prepared for Linux kernel version 1.3.51.  Installation
 in later versions will probably be successful as well, though BusLogic.patch
-may not be required once this driver becomes part of the standard development
-kernel; installation in earlier versions should not be attempted as 1.3.41
-contains changes I made to the common code that are essential for correct error
-recovery.
+may not be required.  Installation in Linux 1.3.41 - 1.3.50 will probably be
+successful, but may require minor modifications.
 
 To install the BusLogic SCSI driver, you may use the following commands,
 replacing "/usr/src" with wherever you keep your Linux kernel source tree:
 
   cd /usr/src
-  tar -xvzf BusLogic-1.3.0.tar.gz
+  tar -xvzf BusLogic-1.3.1.tar.gz
   mv README.BusLogic BusLogic.[ch] linux/drivers/scsi
   patch -p < BusLogic.patch
   cd linux

drivers/scsi/hosts.c

@@ -290,8 +290,6 @@ struct Scsi_Host * scsi_register(Scsi_Host_Template * tpnt, int j){
 
     retval->unique_id = 0;
     retval->io_port = 0;
-    retval->forbidden_addr = 0;
-    retval->forbidden_size = 0;
     retval->hostt = tpnt;
     retval->next = NULL;
 #ifdef DEBUG
@@ -404,28 +402,6 @@ unsigned int scsi_init()
     return 0;
 }
 
-
-void scsi_mem_init(unsigned long memory_end)
-{
-    struct Scsi_Host *Host;
-    long High8, Low24;
-    for (Host = scsi_hostlist; Host != NULL; Host = Host->next) {
-	if (Host->forbidden_addr > 0 && Host->forbidden_size > 0) {
-	    for (High8 = 1<<24; High8 < memory_end; High8 += 1<<24) {
-		for (Low24 = Host->forbidden_addr;
-		     Low24 < Host->forbidden_addr + Host->forbidden_size;
-		     Low24 += PAGE_SIZE) {
-		    unsigned long ForbiddenAddress = High8 + Low24;
-		    if (ForbiddenAddress >= memory_end) goto next_host;
-		    mem_map[MAP_NR(ForbiddenAddress)].reserved = 1;
-		}
-	    }
-	}
-    next_host:
-	continue;
-    }
-}
-
 /*
  * Overrides for Emacs so that we follow Linus's tabbing style.
  * Emacs will notice this stuff at the end of the file and automatically

drivers/scsi/hosts.h

@@ -282,14 +282,6 @@ struct Scsi_Host
      * initialized to 0 in scsi_register.
      */
     unsigned int unique_id;
-
-    /*
-     * Set these if there are conflicts between memory
-     * in the < 1mb region and regions at 16mb multiples.
-     * The address must be on a page boundary.
-     */
-    unsigned long forbidden_addr;
-    unsigned long forbidden_size;
     
     /*
      * The rest can be copied from the template, or specifically
@@ -313,7 +305,7 @@ struct Scsi_Host
      */
     unsigned suggest_bus_reset:1;
  
-    int hostdata[0];  /* Used for storage of host specific stuff */
+    unsigned long hostdata[0];  /* Used for storage of host specific stuff */
 };
 
 extern struct Scsi_Host * scsi_hostlist;

drivers/scsi/scsi.c

@@ -802,8 +802,8 @@ static void scsi_times_out (Scsi_Cmnd * SCpnt, int pid)
 	if (!scsi_abort (SCpnt, DID_TIME_OUT, pid))
 	    return;
     case IN_ABORT:
-	printk("SCSI host %d abort() timed out - resetting\n",
-	       SCpnt->host->host_no);
+	printk("SCSI host %d abort (pid %ld) timed out - resetting\n",
+	       SCpnt->host->host_no, SCpnt->pid);
 	if (!scsi_reset (SCpnt, FALSE))
 	    return;
     case IN_RESET:

drivers/scsi/ultrastor.c

@@ -63,7 +63,7 @@
  * the 24F, cause the SCSI bus to do odd things and generally lead to
  * kernel panics and machine hangs.  This is like the Adaptec code.
  *
- * Check I/O ports 14f, and 34f before allocating them to avoid conflicts.
+ * Use check/snarf_region for 14f, 34f to avoid I/O space address conflicts.
  */
 
 /* Changes from version 1.8 to version 1.9

fs/buffer.c

@@ -25,7 +25,9 @@
 #include <linux/locks.h>
 #include <linux/errno.h>
 #include <linux/malloc.h>
+#include <linux/pagemap.h>
 #include <linux/swapctl.h>
+#include <linux/smp.h>
 
 #include <asm/system.h>
 #include <asm/segment.h>
@@ -278,6 +280,7 @@ void invalidate_buffers(kdev_t dev)
 			if (bh->b_count)
 				continue;
 			bh->b_flushtime = 0;
+			clear_bit(BH_Protected, &bh->b_state);
 			clear_bit(BH_Uptodate, &bh->b_state);
 			clear_bit(BH_Dirty, &bh->b_state);
 			clear_bit(BH_Req, &bh->b_state);
@@ -837,6 +840,7 @@ void __bforget(struct buffer_head * buf)
 {
 	wait_on_buffer(buf);
 	mark_buffer_clean(buf);
+	clear_bit(BH_Protected, &buf->b_state);
 	buf->b_count--;
 	remove_from_hash_queue(buf);
 	buf->b_dev = NODEV;
@@ -1023,7 +1027,7 @@ static void read_buffers(struct buffer_head * bh[], int nrbuf)
 	} while (nrbuf > 0);
 }
 
-int bread_page(unsigned long address, kdev_t dev, int b[], int size)
+static int bread_page(unsigned long address, kdev_t dev, int b[], int size)
 {
 	struct buffer_head *bh, *next, *arr[MAX_BUF_PER_PAGE];
 	int block, nr;
@@ -1072,6 +1076,34 @@ int bread_page(unsigned long address, kdev_t dev, int b[], int size)
 	return 0;
 }
 
+/*
+ * Generic "readpage" function for block devices that have the
+ * normal bmap functionality. This is most of the block device
+ * filesystems.
+ */
+int generic_readpage(struct inode * inode, struct page * page)
+{
+	unsigned long block;
+	int *p, nr[PAGE_SIZE/512];
+	int i;
+
+	i = PAGE_SIZE >> inode->i_sb->s_blocksize_bits;
+	block = page->offset >> inode->i_sb->s_blocksize_bits;
+	p = nr;
+	do {
+		*p = inode->i_op->bmap(inode, block);
+		i--;
+		block++;
+		p++;
+	} while (i > 0);
+
+	/* We should make this asynchronous, but this is good enough for now.. */
+	bread_page(page_address(page), inode->i_dev, nr, inode->i_sb->s_blocksize);
+	page->uptodate = 1;
+	wake_up(&page->wait);
+	return 0;
+}
+
 #if 0
 /*
  * bwrite_page writes a page out to the buffer cache and/or the physical device.
@@ -1781,22 +1813,35 @@ asmlinkage int sys_bdflush(int func, long data)
  * the syscall above, but now we launch it ourselves internally with
  * kernel_thread(...)  directly after the first thread in init/main.c */
 
-int bdflush(void * unused) {
-	
+int bdflush(void * unused) 
+{
 	int i;
 	int ndirty;
 	int nlist;
 	int ncount;
 	struct buffer_head * bh, *next;
 
-	/* We have a bare-bones task_struct, and really should fill
-	in a few more things so "top" and /proc/2/{exe,root,cwd}
-	display semi-sane things. Not real crucial though...  */
+	/*
+	 *	We have a bare-bones task_struct, and really should fill
+	 *	in a few more things so "top" and /proc/2/{exe,root,cwd}
+	 *	display semi-sane things. Not real crucial though...  
+	 */
 
 	current->session = 1;
 	current->pgrp = 1;
 	sprintf(current->comm, "kernel bdflush");
 
+	/*
+	 *	As a kernel thread we want to tamper with system buffers
+	 *	and other internals and thus be subject to the SMP locking
+	 *	rules. (On a uniprocessor box this does nothing).
+	 */
+	 
+#ifdef __SMP__
+	lock_kernel();
+	syscall_count++;
+#endif
+		 
 	for (;;) {
 #ifdef DEBUG
 		printk("bdflush() activated...");

fs/ext/file.c

@@ -64,7 +64,7 @@ struct inode_operations ext_file_inode_operations = {
 	NULL,			/* rename */
 	NULL,			/* readlink */
 	NULL,			/* follow_link */
-	NULL,			/* readpage */
+	generic_readpage,	/* readpage */
 	NULL,			/* writepage */
 	ext_bmap,		/* bmap */
 	ext_truncate,		/* truncate */

fs/ext2/balloc.c

@@ -227,7 +227,7 @@ void ext2_free_blocks (const struct inode * inode, unsigned long block,
 				      block);
 		else {
 			if (sb->dq_op)
-				sb->dq_op->free_block(inode, fs_to_dq_blocks(1, inode->i_blksize));
+				sb->dq_op->free_block(inode, fs_to_dq_blocks(1, sb->s_blocksize));
 			gdp->bg_free_blocks_count++;
 			es->s_free_blocks_count++;
 		}
@@ -404,7 +404,7 @@ int ext2_new_block (const struct inode * inode, unsigned long goal,
 	 * Check quota for allocation of this block.
 	 */
 	if (sb->dq_op)
-		if (sb->dq_op->alloc_block (inode, fs_to_dq_blocks(1, inode->i_blksize))) {
+		if (sb->dq_op->alloc_block (inode, fs_to_dq_blocks(1, sb->s_blocksize))) {
 			unlock_super (sb);
 			*err = -EDQUOT;
 			return 0;
@@ -424,7 +424,7 @@ int ext2_new_block (const struct inode * inode, unsigned long goal,
 		ext2_warning (sb, "ext2_new_block",
 			      "bit already set for block %d", j);
 		if (sb->dq_op)
-			sb->dq_op->free_block(inode, fs_to_dq_blocks(1, inode->i_blksize));
+			sb->dq_op->free_block(inode, fs_to_dq_blocks(1, sb->s_blocksize));
 		goto repeat;
 	}
 
@@ -440,11 +440,11 @@ int ext2_new_block (const struct inode * inode, unsigned long goal,
 		for (k = 1;
 		     k < 8 && (j + k) < EXT2_BLOCKS_PER_GROUP(sb); k++) {
 			if (sb->dq_op)
-				if (sb->dq_op->alloc_block(inode, fs_to_dq_blocks(1, inode->i_blksize)))
+				if (sb->dq_op->alloc_block(inode, fs_to_dq_blocks(1, sb->s_blocksize)))
 					break;
 			if (set_bit (j + k, bh->b_data)) {
 				if (sb->dq_op)
-					sb->dq_op->free_block(inode, fs_to_dq_blocks(1, inode->i_blksize));
+					sb->dq_op->free_block(inode, fs_to_dq_blocks(1, sb->s_blocksize));
 				break;
 			}
 			(*prealloc_count)++;

fs/ext2/file.c

@@ -36,8 +36,6 @@
 #include <linux/fs.h>
 #include <linux/ext2_fs.h>
 
-static int ext2_readpage(struct inode *, unsigned long, char *);
-static int ext2_file_read (struct inode *, struct file *, char *, int);
 static int ext2_file_write (struct inode *, struct file *, const char *, int);
 static void ext2_release_file (struct inode *, struct file *);
 
@@ -47,7 +45,7 @@ static void ext2_release_file (struct inode *, struct file *);
  */
 static struct file_operations ext2_file_operations = {
 	NULL,			/* lseek - default */
-	ext2_file_read,		/* read */
+	generic_file_read,	/* read */
 	ext2_file_write,	/* write */
 	NULL,			/* readdir - bad */
 	NULL,			/* select - default */
@@ -74,7 +72,7 @@ struct inode_operations ext2_file_inode_operations = {
 	NULL,			/* rename */
 	NULL,			/* readlink */
 	NULL,			/* follow_link */
-	ext2_readpage,		/* readpage */
+	generic_readpage,	/* readpage */
 	NULL,			/* writepage */
 	ext2_bmap,		/* bmap */
 	ext2_truncate,		/* truncate */
@@ -82,120 +80,6 @@ struct inode_operations ext2_file_inode_operations = {
 	NULL			/* smap */
 };
 
-static int ext2_readpage(struct inode * inode, unsigned long offset, char * page)
-{
-	int *p, nr[PAGE_SIZE/512];
-	int i;
-
-	i = PAGE_SIZE >> inode->i_sb->s_blocksize_bits;
-	offset >>= inode->i_sb->s_blocksize_bits;
-	p = nr;
-	do {
-		*p = ext2_bmap(inode, offset);
-		i--;
-		offset++;
-		p++;
-	} while (i > 0);
-	return bread_page((unsigned long) page, inode->i_dev, nr, inode->i_sb->s_blocksize);
-}
-
-/*
- * This is a generic file read routine, and uses the
- * inode->i_op->readpage() function for the actual low-level
- * stuff. We can put this into the other filesystems too
- * once we've debugged it a bit more.
- */
-static int ext2_file_read (struct inode * inode, struct file * filp,   
-		char * buf, int count)
-{
-	int read = 0;
-	unsigned long pos;
-	unsigned long addr;
-	unsigned long cached_page = 0;
-	struct page *page;
-
-	if (count <= 0)
-		return 0;
-
-	pos = filp->f_pos;
-	
-	for (;;) {
-		unsigned long offset, nr;
-
-		if (pos >= inode->i_size)
-			break;
-
-		offset = pos & ~PAGE_MASK;
-		nr = PAGE_SIZE - offset;
-		if (nr > count)
-			nr = count;
-
-		/* is it already cached? */
-		page = find_page(inode, pos & PAGE_MASK);
-		if (page)
-			goto found_page;
-
-		/* not cached, have to read it in.. */
-		if (!(addr = cached_page)) {
-			addr = cached_page = __get_free_page(GFP_KERNEL);
-			if (!addr) {
-				if (!read)
-					read = -ENOMEM;
-				break;
-			}
-		}
-		inode->i_op->readpage(inode, pos & PAGE_MASK, (char *) addr);
-
-		/* while we did that, things may have changed.. */
-		if (pos >= inode->i_size)
-			break;
-		page = find_page(inode, pos & PAGE_MASK);
-		if (page)
-			goto found_page;
-
-		/* nope, this is the only copy.. */
-		cached_page = 0;
-		page = mem_map + MAP_NR(addr);
-		page->offset = pos & PAGE_MASK;
-		add_page_to_inode_queue(inode, page);
-		add_page_to_hash_queue(inode, page);
-
-found_page:
-		if (nr > inode->i_size - pos)
-			nr = inode->i_size - pos;
-		page->count++;
-		addr = page_address(page);
-		memcpy_tofs(buf, (void *) (addr + offset), nr);
-		free_page(addr);
-		buf += nr;
-		pos += nr;
-		read += nr;
-		count -= nr;
-		if (!count)
-			break;
-	}
-	filp->f_pos = pos;
-	if (cached_page)
-		free_page(cached_page);
-	if (!IS_RDONLY(inode)) {
-		inode->i_atime = CURRENT_TIME;
-		inode->i_dirt = 1;
-	}
-	return read;
-}
-
-static inline void update_vm_cache(struct inode * inode, unsigned long pos,
-	char * buf, int count)
-{
-	struct page * page;
-
-	page = find_page(inode, pos & PAGE_MASK);
-	if (page) {
-		pos = (pos & ~PAGE_MASK) + page_address(page);
-		memcpy((void *) pos, buf, count);
-	}
-}
-
 static int ext2_file_write (struct inode * inode, struct file * filp,
 			    const char * buf, int count)
 {

fs/ext2/ialloc.c

@@ -464,7 +464,7 @@ struct inode * ext2_new_inode (const struct inode * dir, int mode, int * err)
 		inode->i_gid = current->fsgid;
 	inode->i_dirt = 1;
 	inode->i_ino = j;
-	inode->i_blksize = sb->s_blocksize;
+	inode->i_blksize = PAGE_SIZE;	/* This is the optimal IO size (for stat), not the fs block size */
 	inode->i_blocks = 0;
 	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
 	inode->u.ext2_i.i_flags = dir->u.ext2_i.i_flags;

fs/ext2/inode.c

@@ -533,7 +533,7 @@ void ext2_read_inode (struct inode * inode)
 	inode->i_ctime = raw_inode->i_ctime;
 	inode->i_mtime = raw_inode->i_mtime;
 	inode->u.ext2_i.i_dtime = raw_inode->i_dtime;
-	inode->i_blksize = inode->i_sb->s_blocksize;
+	inode->i_blksize = PAGE_SIZE;	/* This is the optimal IO size (for stat), not the fs block size */
 	inode->i_blocks = raw_inode->i_blocks;
 	inode->i_version = ++event;
 	inode->u.ext2_i.i_flags = raw_inode->i_flags;

fs/hpfs/hpfs_fs.c

@@ -175,7 +175,7 @@ static const struct inode_operations hpfs_file_iops =
 	NULL,				/* rename */
 	NULL,				/* readlink */
 	NULL,				/* follow_link */
-	NULL,				/* readpage */
+	generic_readpage,		/* readpage */
 	NULL,				/* writepage */
 	(int (*)(struct inode *, int))
 	&hpfs_bmap,			/* bmap */

fs/inode.c

@@ -306,14 +306,22 @@ void inode_setattr(struct inode *inode, struct iattr *attr)
  * notify_change is called for inode-changing operations such as
  * chown, chmod, utime, and truncate.  It is guaranteed (unlike
  * write_inode) to be called from the context of the user requesting
- * the change.  It is not called for ordinary access-time updates.
- * NFS uses this to get the authentication correct.  -- jrs
+ * the change.
  */
 
 int notify_change(struct inode * inode, struct iattr *attr)
 {
 	int retval;
 
+	if (attr->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME)) {
+		unsigned long now = CURRENT_TIME;
+		if (!(attr->ia_valid & ATTR_ATIME_SET))
+			attr->ia_atime = now;
+		if (!(attr->ia_valid & ATTR_MTIME_SET))
+			attr->ia_mtime = now;
+		attr->ia_ctime = now;
+	}
+
 	if (inode->i_sb && inode->i_sb->s_op  &&
 	    inode->i_sb->s_op->notify_change) 
 		return inode->i_sb->s_op->notify_change(inode, attr);

fs/msdos/file.c

@@ -52,7 +52,7 @@ struct inode_operations msdos_file_inode_operations = {
 	NULL,			/* rename */
 	NULL,			/* readlink */
 	NULL,			/* follow_link */
-	NULL,			/* readpage */
+	generic_readpage,	/* readpage */
 	NULL,			/* writepage */
 	msdos_bmap,		/* bmap */
 	msdos_truncate,		/* truncate */

fs/ncpfs/file.c

 /*
  *  file.c
  *
- *  Copyright (C) 1995 by Paal-Kr. Engstad and Volker Lendecke
+ *  Copyright (C) 1995 by Volker Lendecke
  *
  */
 
@@ -82,7 +82,7 @@ ncp_make_open(struct inode *i, int right)
 static int 
 ncp_file_read(struct inode *inode, struct file *file, char *buf, int count)
 {
-	int bufsize, to_read, already_read;
+	int bufsize, already_read;
 	off_t pos;
         int errno;
 
@@ -126,17 +126,8 @@ ncp_file_read(struct inode *inode, struct file *file, char *buf, int count)
         while (already_read < count)
 	{
 		int read_this_time;
-
-		if ((pos % bufsize) != 0)
-		{
-			to_read = bufsize - (pos % bufsize);
-		}
-		else
-		{
-			to_read = bufsize;
-		}
-
-		to_read = min(to_read, count - already_read);
+		int to_read = min(bufsize - (pos % bufsize),
+				  count - already_read);
 
 		if (ncp_read(NCP_SERVER(inode), NCP_FINFO(inode)->file_handle,
 			     pos, to_read, buf, &read_this_time) != 0)
@@ -156,7 +147,11 @@ ncp_file_read(struct inode *inode, struct file *file, char *buf, int count)
 
         file->f_pos = pos;
 
-	if (!IS_RDONLY(inode)) inode->i_atime = CURRENT_TIME;
+	if (!IS_RDONLY(inode))
+	{
+		inode->i_atime = CURRENT_TIME;
+	}
+
 	inode->i_dirt = 1;
 
         DPRINTK("ncp_file_read: exit %s\n", NCP_ISTRUCT(inode)->entryName);
@@ -168,7 +163,7 @@ static int
 ncp_file_write(struct inode *inode, struct file *file, const char *buf,
 	       int count)
 {
-	int bufsize, to_write, already_written;
+	int bufsize, already_written;
         off_t pos;
         int errno;
 			  
@@ -211,17 +206,9 @@ ncp_file_write(struct inode *inode, struct file *file, const char *buf,
         while (already_written < count)
 	{
 		int written_this_time;
+		int to_write = min(bufsize - (pos % bufsize),
+				   count - already_written);
 
-		if ((pos % bufsize) != 0)
-		{
-			to_write = bufsize - (pos % bufsize);
-		}
-		else
-		{
-			to_write = bufsize;
-		}
-
-		to_write = min(to_write, count - already_written);
 		if (ncp_write(NCP_SERVER(inode), NCP_FINFO(inode)->file_handle,
 			      pos, to_write, buf, &written_this_time) != 0)
 		{

fs/nfs/inode.c

@@ -268,13 +268,13 @@ int nfs_notify_change(struct inode *inode, struct iattr *attr)
 	else
 		sattr.size = (unsigned) -1;
 
-	if (attr->ia_valid & ATTR_MTIME) {
+	if (attr->ia_valid & ATTR_MTIME_SET) {
 		sattr.mtime.seconds = attr->ia_mtime;
 		sattr.mtime.useconds = 0;
 	} else 
 		sattr.mtime.seconds = sattr.mtime.useconds = (unsigned) -1;
 
-	if (attr->ia_valid & ATTR_ATIME) {
+	if (attr->ia_valid & ATTR_ATIME_SET) {
 		sattr.atime.seconds = attr->ia_atime;
 		sattr.atime.useconds = 0;
 	} else

fs/sysv/file.c

@@ -64,7 +64,7 @@ struct inode_operations sysv_file_inode_operations = {
 	NULL,			/* rename */
 	NULL,			/* readlink */
 	NULL,			/* follow_link */
-	NULL,			/* readpage */
+	generic_readpage,	/* readpage */
 	NULL,			/* writepage */
 	sysv_bmap,		/* bmap */
 	sysv_truncate,		/* truncate */

fs/umsdos/file.c

@@ -87,7 +87,7 @@ struct inode_operations umsdos_file_inode_operations = {
 	NULL,			/* rename */
 	NULL,			/* readlink */
 	NULL,			/* follow_link */
-	NULL,			/* readpage */
+	generic_readpage,	/* readpage */
 	NULL,			/* writepage */
 	msdos_bmap,		/* bmap */
 	UMSDOS_truncate,/* truncate */

fs/xiafs/file.c

@@ -61,7 +61,7 @@ struct inode_operations xiafs_file_inode_operations = {
     NULL,			/* rename */
     NULL,			/* readlink */
     NULL,			/* follow_link */
-    NULL,			/* readpage */
+    generic_readpage,		/* readpage */
     NULL,			/* writepage */
     xiafs_bmap,			/* bmap */
     xiafs_truncate,		/* truncate */

include/asm-alpha/pgtable.h

@@ -112,7 +112,6 @@ static inline void invalidate_range(struct mm_struct *mm,
 #define _PAGE_UWE	0x2000	/* used to do the dirty bit in software */
 
 /* .. and these are ours ... */
-#define _PAGE_COW	0x10000
 #define _PAGE_DIRTY	0x20000
 #define _PAGE_ACCESSED	0x40000
 
@@ -141,13 +140,13 @@ static inline void invalidate_range(struct mm_struct *mm,
  */
 #define PAGE_NONE	__pgprot(_PAGE_VALID | __ACCESS_BITS | _PAGE_FOR | _PAGE_FOW | _PAGE_FOE)
 #define PAGE_SHARED	__pgprot(_PAGE_VALID | __ACCESS_BITS)
-#define PAGE_COPY	__pgprot(_PAGE_VALID | __ACCESS_BITS | _PAGE_FOW | _PAGE_COW)
+#define PAGE_COPY	__pgprot(_PAGE_VALID | __ACCESS_BITS | _PAGE_FOW)
 #define PAGE_READONLY	__pgprot(_PAGE_VALID | __ACCESS_BITS | _PAGE_FOW)
 #define PAGE_KERNEL	__pgprot(_PAGE_VALID | _PAGE_ASM | _PAGE_KRE | _PAGE_KWE)
 
 #define _PAGE_NORMAL(x) __pgprot(_PAGE_VALID | __ACCESS_BITS | (x))
 
-#define _PAGE_P(x) _PAGE_NORMAL((x) | (((x) & _PAGE_FOW)?0:(_PAGE_FOW | _PAGE_COW)))
+#define _PAGE_P(x) _PAGE_NORMAL((x) | (((x) & _PAGE_FOW)?0:_PAGE_FOW))
 #define _PAGE_S(x) _PAGE_NORMAL(x)
 
 /*
@@ -268,20 +267,17 @@ extern inline int pte_write(pte_t pte)		{ return !(pte_val(pte) & _PAGE_FOW); }
 extern inline int pte_exec(pte_t pte)		{ return !(pte_val(pte) & _PAGE_FOE); }
 extern inline int pte_dirty(pte_t pte)		{ return pte_val(pte) & _PAGE_DIRTY; }
 extern inline int pte_young(pte_t pte)		{ return pte_val(pte) & _PAGE_ACCESSED; }
-extern inline int pte_cow(pte_t pte)		{ return pte_val(pte) & _PAGE_COW; }
 
 extern inline pte_t pte_wrprotect(pte_t pte)	{ pte_val(pte) |= _PAGE_FOW; return pte; }
 extern inline pte_t pte_rdprotect(pte_t pte)	{ pte_val(pte) |= _PAGE_FOR; return pte; }
 extern inline pte_t pte_exprotect(pte_t pte)	{ pte_val(pte) |= _PAGE_FOE; return pte; }
 extern inline pte_t pte_mkclean(pte_t pte)	{ pte_val(pte) &= ~(__DIRTY_BITS); return pte; }
 extern inline pte_t pte_mkold(pte_t pte)	{ pte_val(pte) &= ~(__ACCESS_BITS); return pte; }
-extern inline pte_t pte_uncow(pte_t pte)	{ pte_val(pte) &= ~_PAGE_COW; return pte; }
 extern inline pte_t pte_mkwrite(pte_t pte)	{ pte_val(pte) &= ~_PAGE_FOW; return pte; }
 extern inline pte_t pte_mkread(pte_t pte)	{ pte_val(pte) &= ~_PAGE_FOR; return pte; }
 extern inline pte_t pte_mkexec(pte_t pte)	{ pte_val(pte) &= ~_PAGE_FOE; return pte; }
 extern inline pte_t pte_mkdirty(pte_t pte)	{ pte_val(pte) |= __DIRTY_BITS; return pte; }
 extern inline pte_t pte_mkyoung(pte_t pte)	{ pte_val(pte) |= __ACCESS_BITS; return pte; }
-extern inline pte_t pte_mkcow(pte_t pte)	{ pte_val(pte) |= _PAGE_COW; return pte; }
 
 /* 
  * To set the page-dir. Note the self-mapping in the last entry