Linux 2.4.21-rc2

Linux 2.4.21-rc2

[Marcelo Tosatti]

Hi,

Here goes release canditate 2. The aic7xxx problems should be fixed.

Driver ordering from boot: or lilo.conf

[Mr. James W. Laferriere]

	Hello All ,  I am having a difficulty with the dpt_120 driver &
	the sym2-53c8xx driver .  I have just placed a Adaptec 2010s raid
	controller in my system & it now wants to take device 0800 away
	from the sym2 driver at boot time .  ALL drivers are statically
	built into the kernel .  Is there a way at either of the methods
	stated in the subjec to get the order back so that I have my real
	root file system back on 0800 ?  Tia ,  JimL
-- 
       +------------------------------------------------------------------+
       | James   W.   Laferriere | System    Techniques | Give me VMS     |
       | Network        Engineer |     P.O. Box 854     |  Give me Linux  |
       | babydr@baby-dragons.com | Coudersport PA 16915 |   only  on  AXP |
       +------------------------------------------------------------------+

[patch] 2.4.21-rc2: fix sound/kahlua.c .text.exit error

[Adrian Bunk]

drivers/sound/kahlua.c in 2.4.21-rc2 causes a .text.exit error at the 
final linking when compiled statically into a kernel with CONFIG_HOTPLUG 
enabled. The pointer to remove_one is __devexit_p and not discarded but 
the __exit remove_one is discarded.

The following patch is needed:


--- linux-2.4.21-rc2-full/drivers/sound/kahlua.c.old	2003-05-10 19:03:12.000000000 +0200
+++ linux-2.4.21-rc2-full/drivers/sound/kahlua.c	2003-05-10 19:03:41.000000000 +0200
@@ -182,7 +182,7 @@
 	return 0;
 }
 
-static void __exit remove_one(struct pci_dev *pdev)
+static void __devexit remove_one(struct pci_dev *pdev)
 {
 	struct address_info *hw_config = pci_get_drvdata(pdev);
 	sb_dsp_unload(hw_config, 0);
@@ -219,7 +219,7 @@
 	return pci_module_init(&kahlua_driver);
 }
 
-static void __exit kahlua_cleanup_module(void)
+static void __devexit kahlua_cleanup_module(void)
 {
 	return pci_unregister_driver(&kahlua_driver);
 }


Please apply for -rc3
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

[patch] 2.4.21-rc2: fix ips.c .text.exit error

[Adrian Bunk]

ips_remove_device in drivers/scsi/ips.c is __devexit but it's called 
from the __devinit ips_insert_device resulting in a compile error when 
cmpiled statically into a kernel without CONFIG_HOTPLUG enables.

The patch below solves this by removing the __devexit from 
ips_remove_device.

Please apply for -rc3
Adrian

--- linux-2.4.21-pre6-full-nohotplug/drivers/scsi/ips.c.old	2003-03-27 22:25:49.000000000 +0100
+++ linux-2.4.21-pre6-full-nohotplug/drivers/scsi/ips.c	2003-03-27 22:28:51.000000000 +0100
@@ -246,9 +246,6 @@
     #define IPS_SG_ADDRESS(sg)       ((sg)->address)
     #define IPS_LOCK_SAVE(lock,flags) spin_lock_irqsave(&io_request_lock,flags)
     #define IPS_UNLOCK_RESTORE(lock,flags) spin_unlock_irqrestore(&io_request_lock,flags)
-    #ifndef __devexit_p
-        #define __devexit_p(x) x
-    #endif
 #else
     #define IPS_SG_ADDRESS(sg)      (page_address((sg)->page) ? \
                                      page_address((sg)->page)+(sg)->offset : 0)
@@ -338,48 +335,48 @@
    static char ips_hot_plug_name[] = "ips";
    
    static int __devinit  ips_insert_device(struct pci_dev *pci_dev, const struct pci_device_id *ent);
-   static void __devexit ips_remove_device(struct pci_dev *pci_dev);
+   static void ips_remove_device(struct pci_dev *pci_dev);
    
    struct pci_driver ips_pci_driver = {
        name:		ips_hot_plug_name,
        id_table:	ips_pci_table,
        probe:		ips_insert_device,
-       remove:		__devexit_p(ips_remove_device),
+       remove:		ips_remove_device,
    }; 
            
    struct pci_driver ips_pci_driver_anaconda = {
        name:		ips_hot_plug_name,
        id_table:	ips_pci_table_anaconda,
        probe:		ips_insert_device,
-       remove:		__devexit_p(ips_remove_device),
+       remove:		ips_remove_device,
    }; 
 
    struct pci_driver ips_pci_driver_5i = {
        name:		ips_hot_plug_name,
        id_table:	ips_pci_table_5i,
        probe:		ips_insert_device,
-       remove:		__devexit_p(ips_remove_device),
+       remove:		ips_remove_device,
    };
            
    struct pci_driver ips_pci_driver_6i = {
        name:		ips_hot_plug_name,
        id_table:	ips_pci_table_6i,
        probe:		ips_insert_device,
-       remove:		__devexit_p(ips_remove_device),
+       remove:		ips_remove_device,
    };
 
    struct pci_driver ips_pci_driver_i960 = {
        name:		ips_hot_plug_name,
        id_table:	ips_pci_table_i960,
        probe:		ips_insert_device,
-       remove:		__devexit_p(ips_remove_device),
+       remove:		ips_remove_device,
    };
 
    struct pci_driver ips_pci_driver_adaptec = {
        name:		ips_hot_plug_name,
        id_table:	ips_pci_table_adaptec,
        probe:		ips_insert_device,
-       remove:		__devexit_p(ips_remove_device),
+       remove:		ips_remove_device,
    };
 
 #endif
@@ -7346,7 +7343,7 @@
 /*   Routine Description:                                                    */
 /*     Remove one Adapter ( Hot Plugging )                                   */
 /*---------------------------------------------------------------------------*/
-static void __devexit ips_remove_device(struct pci_dev *pci_dev)
+static void ips_remove_device(struct pci_dev *pci_dev)
 {
    int    i;
    struct Scsi_Host *sh;

Linux v2.4.21-rc2 changelog (was: Linux 2.4.21-rc2)

[Matthias Andree]

On Thu, 08 May 2003, Marcelo Tosatti wrote:

> Here goes release canditate 2. The aic7xxx problems should be fixed.

It's late, but here goes the changelog summary with changes since
2.4.21-rc1 (the usual bk set -d -rv2.4.21-rc1, listing the first lines
of the changesets that make up 2.4.21-rc2 but aren't in 2.4.21-rc1, IOW,
the rc1 -> rc2 changes, grouped by author and sorted by surname:

<lucy:innosys.com>:
  o USB: keyspan driver fixes

Muli Ben-Yehuda:
  o [NETFILTER]: ip_queue memory leaks

Neil Brown:
  o Update umem driver for newer cards
  o Return correct result for ACCESS(READ) on eXecute-only

Ben Collins:
  o Fix highmem_io for sbp2
  o More firewire/IEEE1394 fixes
  o Fix IEEE1394 locking problems + cleanups

Alan Cox:
  o update hptraid
  o fix x.25 parsing
  o xdr warning (0 - any)
  o fix wrong type
  o maintainer updates
  o fix base handling in lib stuff
  o kill unneeded ifdefs, add rd/ and root=nbd
  o headers for sisfb update
  o put the ide idents back in working order
  o fix wrong types in if_shaper
  o sisfb ipdate
  o new sis fb idents
  o header for arcnet fixes
  o remove dead functions
  o fix the d_path error cases in umsdos
  o more /proc error cases
  o fix error cases in procfs
  o copy kernel not user object in ncpfs
  o handle error case in fs/namespace.c
  o fix lots of tdfxfb bugs
  o make sstfb work bigendian
  o Fix copy_to_user handling in vicam
  o make pegasus work on big endian
  o mdc800 copy_to_user handling fix
  o fix a race and a comment in via_audio
  o mpu401 copy_to_user handling fix
  o small fix for pcm alloc on i810
  o Fix get_user handling in cmpci
  o Fix copy_to_user handling in awe_wave
  o fix build with newer binutils
  o add another card id
  o fix qlogicisp leaks
  o fix nsp32 build with newer binutils
  o fix ide-scsi retry oops
  o fix cpqfc leak
  o fix time type in aha152x
  o fix 82092 crash cases
  o Fix copy_user handling in cosa
  o sis900 needs to know another PHY
  o fix roadrunner memory leak
  o fix compile of r8169 with newer binutils
  o Fix arcnet crashes with raw socket
  o Fix copy_user handling in z36120
  o /proc stuff for zoran
  o Fix copy_to_user handling in eicon
  o add blacklist for barracuda ata iv with CSB5
  o IDE: if 0 garbage removal
  o fix ide smp deadlock on settings sem
  o fix memory leak on rio
  o fix overrun in cdu31a

Oleg Drokin:
  o Memleak fix for DIGITAL EtherWORKS 3 ethernet driver

Jeff Garzik:
  o fix fealnx build on ia64 and other non-x86
  o tg3 fix

Christoph Hellwig:
  o add intelfb to Config.in

Stephen Hemminger:
  o [BRIDGE]: New maintainership

Benjamin Herrenschmidt:
  o Fix PPC build

Marcel Holtmann:
  o [Bluetooth] Fix L2CAP binding to local address
  o [Bluetooth] Respond correctly to RLS packets

Bernhard Kaindl:
  o Fixup 2.4 ptrace fix

Andi Kleen:
  o Another x86-64 build fix for gcc-3.3-hammer
  o Fix SMP x86-64 kernels on simics
  o Fix gcc 3.3 build for reverted aic7xxx driver
  o Critical fix for x86-64

Dave Kleikamp:
  o JFS: Avoid rare deadlock
  o JFS: jfs_lookup should check for bad inode returned from iget
  o JFS: Performance improvement

Maksim Krasnyanskiy:
  o [Bluetooth] Fix race condition in RFCOMM session and dlc scheduler
  o [Bluetooth] Improved RFCOMM TTY buffer management. Don't buffer
    more data than we have credits for.

Greg Kroah-Hartman:
  o IBM PCI Hotplug: fix up a number of memory leaks on the error path
  o IBM PCI Hotplug: fix up a lot of memory allocations and leaks just
    to figure out a slot name
  o i2c: bug fix for 2.4.21-rc1

Paul Mackerras:
  o Fix drivers/video/Config.in
  o update CREDITS
  o PPC32: Compile fix for ppc_ksyms.c - it needs the declaration of
    __div64_32
  o PPC32: Update the defconfigs

Torben Mathiasen:
  o PCI Hotplug: cpqphp 66/100/133MHz PCI-X support

Patrick McHardy:
  o [NETFILTER]: Multiple ipt_REJECT fixes

Adam Mercer:
  o Fix vesafb with large memory

David S. Miller:
  o [NET]: Fix hashing exploits in ipv4 routing, IP conntrack, and TCP
    synq
  o [NET]: SG without checksum support is illegal

James Morris:
  o [IPV4]: Choose new rt_hash_rnd every rt_run_flush
  o [NET]: Cosmetic cleanups of jhash code

Paulo Ornati:
  o explicit support for nVidia nForce

Ernie Petrides:
  o Orphan recovery error path fix

Stelian Pop:
  o sonypi fixes

Tom Rini:
  o PPC32: Export a missing symbol (__div64_32)

Nivedita Singhvi:
  o [AF_UNIX]: Fix max_dgram_qlen procfs permissions

Alan Stern:
  o USB: usb-storage fixes
  o USB: usb storage async unlink error code fix

Jun Sun:
  o kiobuf flush dcache properly

Marcelo Tosatti:
  o Changed EXTRAVERSION to -rc2 TAG: v2.4.21-rc2
  o aic7xxx PCI posting flush fix from Arjan
  o aic7xxx: Go back to old aic7xxx (pre3) since the new one lockups
    some cards on initialization. The new driver (aic79xx) is now a new
    directory. I know Justin will hate this, but I can't update the
    aic7xxx to a fully new driver in -rc stage.
  o Avoid is_dumpable() NULL pointer reference

Harald Welte:
  o [NETFILTER]: Trivial but important state fix for ipt_conntrack
  o [NETFILTER]: Makefile and build fixes

David Woodhouse:
  o JFFS2: Fix for_each_inode()

-- 
Matthias Andree

Re: Linux 2.4.21-rc2

[Jerome Chantelauze]

On Thu, May 08, 2003 at 10:15:12PM -0300, Marcelo Tosatti wrote:
> 
> Hi,
> 
> Here goes release canditate 2. The aic7xxx problems should be fixed.


Hi Marcello

Here is a patch I sent to lkml a few days ago. It tries to fix a problem
with 2.4.21-rc1 ide driver Makefile.

drivers/ide/Makefile is broken on 2.4.21-rc2 (it's broken at least since
2.4.421-pre6. 2.4.20 is OK).

I try to build a kernel with Old hard disk (MFM/RLL/IDE) support only
(I don't need the Enhanced IDE/MFM/RLL disk/cdrom/tape/floppy support).

Here is my .config:

#
# ATA/IDE/MFM/RLL support
#
CONFIG_IDE=y

#
# IDE, ATA and ATAPI Block devices
#
# CONFIG_BLK_DEV_IDE is not set
CONFIG_BLK_DEV_HD_ONLY=y
CONFIG_BLK_DEV_HD=y
CONFIG_IDEDMA_AUTO=y
# CONFIG_IDEDMA_IVB is not set
# CONFIG_DMA_NONPCI is not set
CONFIG_BLK_DEV_PDC202XX=y
CONFIG_BLK_DEV_IDE_MODES=y
# CONFIG_BLK_DEV_ATARAID is not set
# CONFIG_BLK_DEV_ATARAID_PDC is not set
# CONFIG_BLK_DEV_ATARAID_HPT is not set
# CONFIG_BLK_DEV_ATARAID_SII is not set

The resulting kernel has no support for ide hard disks:

# ls -l drivers/ide/*.o
-rw-r--r--    1 root     root            8 apr 23 08:56 drivers/ide/idedriver.o
#

The following patch fixes the problem for the x86 arch (sorry, I have no
access to other archs). 


*** linux-2.4.21-rc1/drivers/ide/Makefile.orig  Wed Apr 23 08:45:48 2003
--- linux-2.4.21-rc1/drivers/ide/Makefile       Wed Apr 23 09:20:14 2003
***************
*** 21,26 ****
--- 21,28 ----
  
  subdir-$(CONFIG_BLK_DEV_IDE) += legacy ppc arm raid pci
  
+ subdir-$(CONFIG_BLK_DEV_HD_ONLY) += legacy
+ 
  # First come modules that register themselves with the core
  
  ifeq ($(CONFIG_BLK_DEV_IDE),y)
***************
*** 50,55 ****
--- 52,60 ----
    obj-y               += arm/idedriver-arm.o
  endif
  
+ ifeq ($(CONFIG_BLK_DEV_HD_ONLY),y)
+   obj-y               += legacy/idedriver-legacy.o
+ endif
  
  ifeq ($(CONFIG_BLK_DEV_IDE),y)
  # RAID must be last of all


I used a x86 computer running a debian woody (gcc 2.95.4, glibc 2.2.5)
to build the kernel.

Regards

--
Jerome Chantelauze
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/