Linux 2.4.28-rc2

Linux 2.4.28-rc2

[Marcelo Tosatti]

Hi,

Here goes the second release canditate of 2.4.28.

It contains a network update (which is composed of several smaller changes 
- IPVS, SCTP, Netfilter), a collection of ACPI bugfixes, a SMBFS client 
buffer overflow fix (which is very hard to exploit - the attacker needs to 
control packets from the server), amongst others.

This is going to be become 2.4.28 if nothing bad shows up.

Please help with testing! 


Summary of changes from v2.4.28-rc1 to v2.4.28-rc2
============================================

<arnouten:bzzt.net>:
  o [TCP]: Add /proc/net/tcp{,6} layout documentation

<lkml:rtr.ca>:
  o delkin_cb: new carbus IDE driver

<ruber:engr.es>:
  o [CRYPTO]: Add Tnepres cipher support

Aaron Grothe:
  o [CRYPTO]: Put khazad back into tcrypt table

Adrian Bunk:
  o add SCSI_SATA_ULI help text
  o Adrian Bunk CREDITS entry

Andrea Arcangeli:
  o [NET]: Accept should return ENFILE not EMFILE

Chris Wright:
  o compile fix for neighbour scalability backport
  o compile fix for neighbour scalability backport

David S. Miller:
  o [PKT_SCHED]: sch_netem.c needs linux/init.h
  o [AF_UNIX]: Remove spurious len test in unix_mkname
  o [CRYPTO]: Fix typo in Kconfig
  o [TG3]: Update driver version and reldate
  o [AF_PACKET]: Set VM_IO for mmap areas
  o [CRYPTO]: Delete MODULE_ALIAS line

Eric Sandeen:
  o fix for large direct I/O

Greg Banks:
  o [NET]: Fix race between neigh-timer_handler and neigh_event_send

Harald Welte:
  o [NET]: Backport neighbour scalability fixes from 2.6.x
  o [NETFILTER]: fix ipt_ULOG bogus error messages

Ivan Kokshaysky:
  o Alpha: fixes for bootp/bootpz targets

James Morris:
  o [CRYPTO]: Add __init and __initdata to aes.c

Keith Owens:
  o Avoid oops in proc_delete_inode

Len Brown:
  o [ACPI] reserve IOPORTS for ACPI (David Shaohua Li) http://bugzilla.kernel.org/show_bug.cgi?id=2641
  o [ACPI] boot option fixes from 2.6 "acpi_serialize" "acpi_wake_gpes_always_on" "acpi_osi=" http://bugzilla.kernel.org/show_bug.cgi?id=2534
  o set acpi_gbl_leave_wake_gpes_disabled to FALSE for 2.4 because it would take a backport of big 2.6 changes to make this code work and 2.4 doesn't support suspend/resume anyway.
  o [ACPI] Enter ACPI mode earlier Fixes two common boot failures due to buggy SMM BIOS code
  o [ACPI] fix build warnings
  o [ACPI] build fix
  o [ACPI] If BIOS disabled the LAPIC, believe it by default

Maciej W. Rozycki:
  o [NET]: Fix fddi_statistics for 64-bit
  o [IPV4]: Set ARP hw type correctly for BOOTP over FDDI
  o [IPV4]: Permit the official ARP hw type in SIOCSARP for FDDI

Marcelo Tosatti:
  o Jakub Bogusz: missing include in farsync WAN driver
  o mcp: Fix proc_delete_inode oops bug correction typo
  o Urban Widmark: Fix smbfs client overflow
  o Changed EXTRAVERSION to -rc2

Patrick Caulfield:
  o [DECNET]: Mark myself as maintainer

Patrick McHardy:
  o [PKT_SCHED]: Fix netem qlen accounting

Paul Fulghum:
  o serial receive lockup fix
  o usb serial write fix

Pete Zaitcev:
  o USB: update unusual_devs.h

Randy Dunlap:
  o [TG3]: tg3_nvram_read_using_eeprom cannot be __init

Sridhar Samudrala:
  o [SCTP] Adaption layer indication support
  o [SCTP] Update cwnd/ssthresh as per the sctpimpguide modifications
  o [SCTP] When an address is deleted, update any transports that are caching it as a source adddress.
  o [SCTP] Fix HEARTBEAT_ACKs being sent to wrong dest. ip address in a multi-homing scenario after a failback.

Stephen Hemminger:
  o [PKT_SCHED]: netem: Use timer to handle packets not rescheduling

Thomas Graf:
  o [PKT_SCHED]: Remove useless line in cbq_dump_class
  o [PKT_SCHED]: Make rate estimator work on all platforms
  o [PKT_SCHED]: CBQ; Destroy filters before destroying classes
  o [PKT_SCHED]: u32: Remove unused hgenerator field in tc_u_hnode
  o [PKT_SCHED]: Avoid duplicated TCA_STATS TLVs for HTB and HFSC
  o [PKT_SCHED]: Rename TCQ_F_INGRES to TCQ_F_INGRESS
  o filemap.c compile fix

Wensong Zhang:
  o [IPVS]: Fix endian problem on sync message size

Özkan Sezer:
  o 2.6 backport: binfmt_elf memleak fix error handling
  o 2.6 backport: tun sign mishandling

2.4.28-rc2: net/atm/proc.c compile error

[Adrian Bunk]

On Sun, Nov 07, 2004 at 03:37:53PM -0200, Marcelo Tosatti wrote:
>...
> Summary of changes from v2.4.28-rc1 to v2.4.28-rc2
> ============================================
>...
> Harald Welte:
>   o [NET]: Backport neighbour scalability fixes from 2.6.x
>...


This patch removes atm_lec_info but not the user of this function, 
resulting in the following compile error:


<--  snip  -->

...
gcc-2.95 -D__KERNEL__ -I/home/bunk/linux/kernel-2.4/linux-2.4.28-rc2-full/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe -mpreferred-stack-boundary=2 -march=i686 -malign-functions=4    -nostdinc -iwithprefix include -DKBUILD_BASENAME=proc  -c -o proc.o proc.c
proc.c: In function `atm_proc_init':
proc.c:624: `atm_lec_info' undeclared (first use in this function)
proc.c:624: (Each undeclared identifier is reported only once
proc.c:624: for each function it appears in.)
make[3]: *** [proc.o] Error 1
make[3]: Leaving directory `/home/bunk/linux/kernel-2.4/linux-2.4.28-rc2-full/net/atm'

<--  snip  -->

Re: 2.4.28-rc2: net/atm/proc.c compile error

[David S. Miller]

On Sun, 7 Nov 2004 22:42:46 +0100
Adrian Bunk <bunk@stusta.de> wrote:

> On Sun, Nov 07, 2004 at 03:37:53PM -0200, Marcelo Tosatti wrote:
> >...
> > Summary of changes from v2.4.28-rc1 to v2.4.28-rc2
> > ============================================
> >...
> > Harald Welte:
> >   o [NET]: Backport neighbour scalability fixes from 2.6.x
> >...
> 
> 
> This patch removes atm_lec_info but not the user of this function, 
> resulting in the following compile error:

You must have mispatched, here is a grep I just did in Marcelo's
current tree:

davem@nuts:/disk1/BK/marcelo-2.4/net/atm$ egrep atm_lec_info *.c
davem@nuts:/disk1/BK/marcelo-2.4/net/atm$ 

Re: 2.4.28-rc2: net/atm/proc.c compile error

[David S. Miller]

On Mon, 8 Nov 2004 06:15:22 +0100
Willy Tarreau <willy@w.ods.org> wrote:

> #define CREATE_ENTRY(name) \
>     name = create_proc_entry(#name,0,atm_proc_root); \
>     if (!name) goto cleanup; \
>     name->data = atm_##name##_info; \
>     name->proc_fops = &proc_spec_atm_operations; \
>     name->owner = THIS_MODULE
> ...
> 623: #if defined(CONFIG_ATM_LANE) || defined(CONFIG_ATM_LANE_MODULE)
> 624:        CREATE_ENTRY(lec);
> 625: #endif
> 
> That's why your grep did not find it ;-)

Indeed, good catch.

> Is it enough to remove these 3 lines ?

No, it isn't, I'll fix this up.

Re: 2.4.28-rc2: net/atm/proc.c compile error

[Willy Tarreau]

Hi David,

On Sun, Nov 07, 2004 at 05:42:47PM -0800, David S. Miller wrote:
 
> You must have mispatched, here is a grep I just did in Marcelo's
> current tree:
> 
> davem@nuts:/disk1/BK/marcelo-2.4/net/atm$ egrep atm_lec_info *.c
> davem@nuts:/disk1/BK/marcelo-2.4/net/atm$ 

No, he patched it right, I got it too and found where it broke :

gcc -D__KERNEL__ -I/data/projets/dev/linux/trees/linux-2.4.28-rc2/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe -mno-fp-regs -ffixed-8 -mcpu=ev6 -Wa,-mev6 -DMODULE  -nostdinc -iwithprefix include -DKBUILD_BASENAME=proc  -DEXPORT_SYMTAB -c proc.c
proc.c: In function `atm_proc_init':
proc.c:624: error: `atm_lec_info' undeclared (first use in this function)
proc.c:624: error: (Each undeclared identifier is reported only once
proc.c:624: error: for each function it appears in.)
make[2]: *** [proc.o] Error 1

----- net/atm/proc.c: -------

#define CREATE_ENTRY(name) \
    name = create_proc_entry(#name,0,atm_proc_root); \
    if (!name) goto cleanup; \
    name->data = atm_##name##_info; \
    name->proc_fops = &proc_spec_atm_operations; \
    name->owner = THIS_MODULE
...
623: #if defined(CONFIG_ATM_LANE) || defined(CONFIG_ATM_LANE_MODULE)
624:        CREATE_ENTRY(lec);
625: #endif

That's why your grep did not find it ;-)
Is it enough to remove these 3 lines ?

Regards,
Willy

Re: Linux 2.4.28-rc2

[Willy Tarreau]

Hi Marcelo,

just started a compile on alpha last night and looked at the results this
morning. 'make vmlinux' ends with the following errors. It is fairly possible
that it comes from my gcc/binutils combination (because I found that I compiled
previous kernels with gcc 3.2.3), in which case I will fix it later, but I'd
like someone with an alpha to check it on his side before the final release.

Cheers,
Willy

bash-2.03$ gcc -v
Reading specs from /usr/lib/gcc-lib/alphaev6-unknown-linux-gnu/3.3.4/specs
Configured with: ../gcc-3.3.4/configure --prefix=/usr --enable-version-specific-runtime-libs --enable-languages=c,c++ --disable-nls --disable-locale --enable-shared --enable-threads --program-suffix=-3.3 --enable-target-optspace --with-gnu-ld --with-gnu-as
Thread model: posix
gcc version 3.3.4

bash-2.03$ ld -v
GNU ld version 2.11.90.0.15 (with BFD 2.11.90.0.15)

bash-2.03$ make vmlinux
...
arch/alpha/mm/mm.o: In function `get_pgd_slow':
arch/alpha/mm/mm.o(.text+0x5c): relocation truncated to fit: GPRELHIGH rodata.cst8
ipc/ipc.o: In function `sys_semtimedop':
ipc/ipc.o(.text+0x3854): relocation truncated to fit: GPRELHIGH rodata.cst8
drivers/char/char.o: In function `kmem_vm_nopage':
drivers/char/char.o(.text+0xc48): relocation truncated to fit: GPRELHIGH rodata.cst8
drivers/char/char.o: In function `vt_ioctl':
drivers/char/char.o(.text+0xdf5c): relocation truncated to fit: GPRELHIGH rodata.cst8
drivers/char/char.o: In function `rs_wait_until_sent':
drivers/char/char.o(.text+0x1e4dc): relocation truncated to fit: GPRELHIGH rodata.cst8
drivers/block/block.o: In function `blk_seg_merge_ok':
drivers/block/block.o(.text+0x328): relocation truncated to fit: GPRELHIGH rodata.cst8
drivers/block/block.o: In function `ll_back_merge_fn':
drivers/block/block.o(.text+0x3d8): relocation truncated to fit: GPRELHIGH rodata.cst8
drivers/block/block.o: In function `ll_front_merge_fn':
drivers/block/block.o(.text+0x4a8): relocation truncated to fit: GPRELHIGH rodata.cst8
drivers/block/block.o: In function `ll_merge_requests_fn':
drivers/block/block.o(.text+0x578): relocation truncated to fit: GPRELHIGH rodata.cst8
drivers/block/block.o: In function `fdc_specify':
drivers/block/block.o(.text+0x56c0): relocation truncated to fit: GPRELHIGH rodata.cst8
/data/projets/dev/linux/trees/linux-2.4.28-rc2/lib/lib.a(bust_spinlocks.o): In function `bust_spinlocks':
bust_spinlocks.o(.text+0x60): relocation truncated to fit: GPRELHIGH rodata.str1.1
make: *** [vmlinux] Error 1

Re: Linux 2.4.28-rc2

[Willy Tarreau]

Ok, just replying to myself.

I found an old gcc-2.91.66 which compiled it correctly on this machine. So
it might be a compiler/binutils issue.

Regards,
willy

On Mon, Nov 08, 200 at 06:32:17AM +0100, Willy Tarreau wrote:
> Hi Marcelo,
> 
> just started a compile on alpha last night and looked at the results this
> morning. 'make vmlinux' ends with the following errors. It is fairly possible
> that it comes from my gcc/binutils combination (because I found that I compiled
> previous kernels with gcc 3.2.3), in which case I will fix it later, but I'd
> like someone with an alpha to check it on his side before the final release.
> 
> Cheers,
> Willy
> 
> bash-2.03$ gcc -v
> Reading specs from /usr/lib/gcc-lib/alphaev6-unknown-linux-gnu/3.3.4/specs
> Configured with: ../gcc-3.3.4/configure --prefix=/usr --enable-version-specific-runtime-libs --enable-languages=c,c++ --disable-nls --disable-locale --enable-shared --enable-threads --program-suffix=-3.3 --enable-target-optspace --with-gnu-ld --with-gnu-as
> Thread model: posix
> gcc version 3.3.4
> 
> bash-2.03$ ld -v
> GNU ld version 2.11.90.0.15 (with BFD 2.11.90.0.15)
> 
> bash-2.03$ make vmlinux
> ...
> arch/alpha/mm/mm.o: In function `get_pgd_slow':
> arch/alpha/mm/mm.o(.text+0x5c): relocation truncated to fit: GPRELHIGH rodata.cst8
> ipc/ipc.o: In function `sys_semtimedop':
> ipc/ipc.o(.text+0x3854): relocation truncated to fit: GPRELHIGH rodata.cst8
> drivers/char/char.o: In function `kmem_vm_nopage':
> drivers/char/char.o(.text+0xc48): relocation truncated to fit: GPRELHIGH rodata.cst8
> drivers/char/char.o: In function `vt_ioctl':
> drivers/char/char.o(.text+0xdf5c): relocation truncated to fit: GPRELHIGH rodata.cst8
> drivers/char/char.o: In function `rs_wait_until_sent':
> drivers/char/char.o(.text+0x1e4dc): relocation truncated to fit: GPRELHIGH rodata.cst8
> drivers/block/block.o: In function `blk_seg_merge_ok':
> drivers/block/block.o(.text+0x328): relocation truncated to fit: GPRELHIGH rodata.cst8
> drivers/block/block.o: In function `ll_back_merge_fn':
> drivers/block/block.o(.text+0x3d8): relocation truncated to fit: GPRELHIGH rodata.cst8
> drivers/block/block.o: In function `ll_front_merge_fn':
> drivers/block/block.o(.text+0x4a8): relocation truncated to fit: GPRELHIGH rodata.cst8
> drivers/block/block.o: In function `ll_merge_requests_fn':
> drivers/block/block.o(.text+0x578): relocation truncated to fit: GPRELHIGH rodata.cst8
> drivers/block/block.o: In function `fdc_specify':
> drivers/block/block.o(.text+0x56c0): relocation truncated to fit: GPRELHIGH rodata.cst8
> /data/projets/dev/linux/trees/linux-2.4.28-rc2/lib/lib.a(bust_spinlocks.o): In function `bust_spinlocks':
> bust_spinlocks.o(.text+0x60): relocation truncated to fit: GPRELHIGH rodata.str1.1
> make: *** [vmlinux] Error 1
> 
> -
> 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/