Hello,
Here goes -pre7.
This is a feature freeze, only bugfixes will be accepted from now on.
-rc1 should be out in a couple of weeks.
-pre7 contains a bunch of updates (networking, sh/sparc64 merge, etc)
including another IO scheduler change which should improve overall
performance a lot.
Summary of changes from v2.4.22-pre6 to v2.4.22-pre7
============================================
<ja:ssi.bg>:
o [IPV4/IPV6]: Fix use-after-free bugs in tunneling drivers
<lethal:unusual.internal.linux-sh.org>:
o SH Merge
o SH update
<tgraf:suug.ch>:
o [NET]: Make {send,recv}msg return EMSGSIZE when msg_iovelen is too big, as per 1003.1
o [NET]: Return EDESTADDRREQ as appropriate in sendmsg implementations
Alan Cox:
o add quota autoload
o typo bits
Ben Collins:
o [SPARC64]: Fix OBP 4.6+ PCI probing, use pcic_present() consistently
o Fix ALi15x3 DMA on sparc64 (maybe others)
Benjamin Herrenschmidt:
o radeonfb: fix artifacts during boot
Chas Williams:
o [ATM]: Add reference counting to atm_dev
o [ATM]: Make ATM buildable as a module
o [ATM]: Eliminate cli, make function names sane in net/atm/lec.c
Christoph Hellwig:
o vmap() backport
Dave Kleikamp:
o JFS: Possible trap/data loss when fixing directory index table
David S. Miller:
o [SUNHME]: Set RXMAX/TXMAX large enough to handle VLAN frames
o [NET]: Ok, sunhme is VLAN challenged after all
o [SUNRPC]: Fix compiler warning in svcsock.c
o [NETFILTER]: Fix build warnings in ipv6 modules, thanks Geert
o [ATM]: Fix build, missing lec_priv member
o [ATM]: Fix lec.c warning with bridging disabled
o [SPARC64]: Fix assumptions about data section ordering and objects ending up in .data vs .bss
o [SPARC{,64}]: Add barrier() to cpu_relax() for consistency with 2.5.x
o [SPARC64]: Update defconfig
o [Bluetooth]: Fix buggy CONFIG_ISDN test in cmtp Config.in
o [SPARC64]: Do not break out of PCI controller probing loop too early
David Stevens:
o [IPV4]: Do not sent IGMP leave messages unless IFF_UP
Gerd Knorr:
o bttv driver update
o tuner driver update
o bttv documentation update
o Update tv card i2c helper modules
Ivan Kokshaysky:
o typecast bug in sched.c bites reschedule_idle
James Morris:
o [NETLINK]: Just drop packets for kernel netlink socket with no data_ready handler
Jens Axboe:
o more iosched work
Maksim Krasnyanskiy:
o [Bluetooth] CMTP protocol depends on ISDN and ISDN CAPI
Marcel Holtmann:
o [Bluetooth] Make READ_TRANSMIT_POWER_LEVEL available for normal users
o [Bluetooth] Support for inquiry with unlimited responses
o [Bluetooth] Support for AVM BlueFRITZ! USB
o [Bluetooth] Add l2cap_load() function
o [Bluetooth] Handle command complete event for inquiry cancel
o [Bluetooth] Declare the function l2cap_load()
o [Bluetooth] Update the maintainer entries for the Bluetooth subsystem
Marcelo Tosatti:
o Changed EXTRAVERSION to -pre7
o buffer.c: remove unused out_putf label
Mikael Pettersson:
o clean crc temp files in lib/
Oleg Drokin:
o Fix link/unlink race. By Chris Mason concurrent link/unlinks can create savelinks for files that still exist, the fix here is to be somewhat smarter about when we change the link count
Paul Mackerras:
o PPC32: Fix the debug check in kunmap_atomic
o PPC32: Fix IRQ sense and polarity setting on 405 and 440 cpus
Paul Mundt:
o sh64: Fix ATM module build
o sh64: defconfig update
o sh64: Cayman IRQ handler updates
Roman Zippel:
o hfs+: update copyright
o hfs+: remove some smaller files
o hfs+: volume/permission fixes
o hfs+: fix rename of links
o hfs+: check size of inode and sb info
o hfs+: various cleanups
o hfs+: link hfsplus before hfs
o hfs+: export mark_page_accessed
o hfs+: Makefile update
Tom Rini:
o PPC32: Add support for the Motorola PowerPlus family of boards
o PPC32: Remove trailing whitespace in numerous files
On Fri, 2003-07-18 at 21:53, Marcelo Tosatti wrote:
> Hello,
>
> Here goes -pre7.
Will ACL/xattr support get its way onto mainstream 2.4 soon?
On 07.18, Marcelo Tosatti wrote:
>
> Hello,
>
> Here goes -pre7.
>
> This is a feature freeze, only bugfixes will be accepted from now on.
>
Could you include at least this, for completeness ?
--- linux-2.4.21-bp1/arch/i386/Makefile.orig 2003-06-18 23:40:25.000000000 +0200
+++ linux-2.4.21-bp1/arch/i386/Makefile 2003-06-18 23:59:25.000000000 +0200
@@ -53,11 +53,11 @@
endif
ifdef CONFIG_MPENTIUMIII
-CFLAGS += -march=i686
+CFLAGS += $(call check_gcc,-march=pentium3,-march=i686)
endif
ifdef CONFIG_MPENTIUM4
-CFLAGS += -march=i686
+CFLAGS += $(call check_gcc,-march=pentium4,-march=i686)
endif
ifdef CONFIG_MK6
And how about the 'inline -> always_inline' patch ? It can be considered
as a bug that breaks the build with gcc3. If you consider it appropiate,
I will resend.
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.22-pre6-jam1m (gcc 3.3.1 (Mandrake Linux 9.2 3.3.1-0.3mdk))
On 07.18, Marcelo Tosatti wrote:
>
> Hello,
>
> Here goes -pre7.
>
> This is a feature freeze, only bugfixes will be accepted from now on.
>
HyperThreaded P4s will still fail to set AT_PLATFORM correctly.
Do you want to kill this bug from .22 ? ;)
--- linux/fs/binfmt_elf.c.orig 2002-12-28 00:12:32.000000000 +0100
+++ linux/fs/binfmt_elf.c 2002-12-28 00:32:37.000000000 +0100
@@ -116,11 +116,14 @@
elf_caddr_t *argv;
elf_caddr_t *envp;
elf_addr_t *sp, *csp;
+ char *stack_top;
char *k_platform, *u_platform;
long hwcap;
size_t platform_len = 0;
size_t len;
+ stack_top = p;
+
/*
* Get hold of platform and hardware capabilities masks for
* the machine we are running on. In some cases (Sparc),
@@ -135,8 +138,8 @@
platform_len = strlen(k_platform) + 1;
u_platform = p - platform_len;
__copy_to_user(u_platform, k_platform, platform_len);
- } else
- u_platform = p;
+ stack_top = u_platform;
+ }
#if defined(__i386__) && defined(CONFIG_SMP)
/*
@@ -149,15 +152,14 @@
* processors. This keeps Mr Marcelo Person happier but should be
* removed for 2.5
*/
-
if(smp_num_siblings > 1)
- u_platform = u_platform - ((current->pid % 64) << 7);
+ stack_top -= ((current->pid % NR_CPUS) << 7);
#endif
/*
* Force 16 byte _final_ alignment here for generality.
*/
- sp = (elf_addr_t *)(~15UL & (unsigned long)(u_platform));
+ sp = (elf_addr_t *)(~15UL & (unsigned long)(stack_top));
csp = sp;
csp -= (1+DLINFO_ITEMS)*2 + (k_platform ? 2 : 0);
#ifdef DLINFO_ARCH_ITEMS
--
J.A. Magallon <[email protected]> \ Software is like sex:
werewolf.able.es \ It's better when it's free
Mandrake Linux release 9.2 (Cooker) for i586
Linux 2.4.22-pre6-jam1m (gcc 3.3.1 (Mandrake Linux 9.2 3.3.1-0.3mdk))
on alpha:
gcc -D__KERNEL__ -I/src/build/linux-2.4.22-pre7-part-build/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-fomit-frame-pointer -pipe -mno-fp-regs -ffixed-8 -mcpu=ev67 -Wa,-mev6
-nostdinc -iwithprefix include -DKBUILD_BASENAME=api -DEXPORT_SYMTAB -c
api.c
In file included from api.c:21:
internal.h:19:28: asm/kmap_types.h: No such file or directory
In file included from api.c:21:
internal.h:24: error: return type is an incomplete type
internal.h: In function rypto_kmap_type':
internal.h:25: error: invalid use of undefined type num km_type'
internal.h:25: warning: eturn' with a value, in function returning void
--
Tom Vier <[email protected]>
DSA Key ID 0xE6CB97DA
On Sat, Jul 19, 2003 at 01:06:41AM +0200, Felipe Alfaro Solana wrote:
> On Fri, 2003-07-18 at 21:53, Marcelo Tosatti wrote:
> > Hello,
> >
> > Here goes -pre7.
>
> Will ACL/xattr support get its way onto mainstream 2.4 soon?
Doubt it.
Unless it gets into -ac or -aa for a long while and a whole bunch of users
clamor for it.
So, is acl only working with ext[23] & XFS? What about reiserfs or jfs?
I was thinking of giving the acl patch a try one of these days. There are a
couple things where the ugo model doesn't work at my company.
Mike
On Sat, Jul 19, 2003 at 01:06:41AM +0200, Felipe Alfaro Solana wrote:
> >
> > Will ACL/xattr support get its way onto mainstream 2.4 soon?
> On l?, 2003-07-19 at 02:38, Mike Fedyk wrote:
> Doubt it.
>
> Unless it gets into -ac or -aa for a long while and a whole bunch of users
> clamor for it.
>
> So, is acl only working with ext[23] & XFS? What about reiserfs or jfs?
There is one patch floating around for reiserfs, but thats probably
_very_ unofficial code. JFS, no idea.
> I was thinking of giving the acl patch a try one of these days. There are a
> couple things where the ugo model doesn't work at my company.
I've been using it for 1-2 years now in a few production setups, never
had any problems with it. That is, other than waiting for a new release
or doing a dirty forwardport myself breaking things each time mainstream
2.4 updates. The distro's I use also ships libraries to support it :-)
--
Cheers,
Andr? Tomt
[email protected]
On 2003-07-19T02:53:11, Andre Tomt <[email protected]> said:
> > So, is acl only working with ext[23] & XFS? What about reiserfs or jfs?
> There is one patch floating around for reiserfs, but thats probably
> _very_ unofficial code. JFS, no idea.
SuSE has been shipping with ACLs in ext[23], XFS, reiserfs, JFS for
almost a year now. Andreas Gruenbacher did a lot of work on it.
Sincerely,
Lars Marowsky-Br?e <[email protected]>
--
SuSE Labs - Research & Development, SuSE Linux AG
"If anything can go wrong, it will." "Chance favors the prepared (mind)."
-- Capt. Edward A. Murphy -- Louis Pasteur
On Sat, Jul 19, 2003 at 01:06:41AM +0200, Felipe Alfaro Solana wrote:
> > Here goes -pre7.
>
> Will ACL/xattr support get its way onto mainstream 2.4 soon?
xattr support is in 2.4 mainline although only JFS currently support
it.
On Fri, Jul 18, 2003 at 05:38:24PM -0700, Mike Fedyk wrote:
> On Sat, Jul 19, 2003 at 01:06:41AM +0200, Felipe Alfaro Solana wrote:
> > On Fri, 2003-07-18 at 21:53, Marcelo Tosatti wrote:
> > > Hello,
> > >
> > > Here goes -pre7.
> >
> > Will ACL/xattr support get its way onto mainstream 2.4 soon?
>
> Doubt it.
>
> Unless it gets into -ac or -aa for a long while and a whole bunch of users
> clamor for it.
The VFS infrastructure for ACLs and the XFS ACL code is in -aa for a long
time.
> So, is acl only working with ext[23] & XFS? What about reiserfs or jfs?
ACLs work with XFS, ext[23] and JFS in 2.4 and 2.4+patches. In addition
SuSE ships patches for xattr/ACL on reiserfs but namesys doesn't like them.
Tom Vier wrote:
> on alpha:
>
> internal.h:19:28: asm/kmap_types.h: No such file or directory
Same on s390 and some other platforms.We should just get rid of
this problem by providing a generic kmap_types header.
Marcello, please consider this patch.
Arnd <><
--
D: Add generic <asm/kmap_types.h> for all architectures that are missing it.
D: This is needed to build the crypto drivers.
diff -Nru a/include/asm-generic/kmap_types.h b/include/asm-generic/kmap_types.h
--- /dev/null Wed Dec 31 16:00:00 1969
+++ b/include/asm-generic/kmap_types.h Sat Jul 19 13:33:30 2003
@@ -0,0 +1,22 @@
+#ifndef _ASM_KMAP_TYPES_H
+#define _ASM_KMAP_TYPES_H
+
+/* This header defines a generic km_type that can be
+ * used by all architectures that do not support
+ * high memory.
+ */
+
+enum km_type {
+ KM_BOUNCE_READ,
+ KM_SKB_SUNRPC_DATA,
+ KM_SKB_DATA_SOFTIRQ,
+ KM_USER0,
+ KM_USER1,
+ KM_BH_IRQ,
+ KM_SOFTIRQ0,
+ KM_SOFTIRQ1,
+
+ KM_TYPE_NR
+};
+
+#endif
diff -Nru a/include/asm-alpha/kmap_types.h b/include/asm-alpha/kmap_types.h
--- /dev/null Wed Dec 31 16:00:00 1969
+++ b/include/asm-alpha/kmap_types.h Sat Jul 19 13:33:30 2003
@@ -0,0 +1 @@
+#include <asm-generic/kmap_types.h>
diff -Nru a/include/asm-arm/kmap_types.h b/include/asm-arm/kmap_types.h
--- /dev/null Wed Dec 31 16:00:00 1969
+++ b/include/asm-arm/kmap_types.h Sat Jul 19 13:33:30 2003
@@ -0,0 +1 @@
+#include <asm-generic/kmap_types.h>
diff -Nru a/include/asm-cris/kmap_types.h b/include/asm-cris/kmap_types.h
--- /dev/null Wed Dec 31 16:00:00 1969
+++ b/include/asm-cris/kmap_types.h Sat Jul 19 13:33:30 2003
@@ -0,0 +1 @@
+#include <asm-generic/kmap_types.h>
diff -Nru a/include/asm-ia64/kmap_types.h b/include/asm-ia64/kmap_types.h
--- /dev/null Wed Dec 31 16:00:00 1969
+++ b/include/asm-ia64/kmap_types.h Sat Jul 19 13:33:30 2003
@@ -0,0 +1 @@
+#include <asm-generic/kmap_types.h>
diff -Nru a/include/asm-s390/kmap_types.h b/include/asm-s390/kmap_types.h
--- /dev/null Wed Dec 31 16:00:00 1969
+++ b/include/asm-s390/kmap_types.h Sat Jul 19 13:33:30 2003
@@ -0,0 +1 @@
+#include <asm-generic/kmap_types.h>
diff -Nru a/include/asm-s390x/kmap_types.h b/include/asm-s390x/kmap_types.h
--- /dev/null Wed Dec 31 16:00:00 1969
+++ b/include/asm-s390x/kmap_types.h Sat Jul 19 13:33:30 2003
@@ -0,0 +1 @@
+#include <asm-generic/kmap_types.h>
diff -Nru a/include/asm-sh/kmap_types.h b/include/asm-sh/kmap_types.h
--- /dev/null Wed Dec 31 16:00:00 1969
+++ b/include/asm-sh/kmap_types.h Sat Jul 19 13:33:30 2003
@@ -0,0 +1 @@
+#include <asm-generic/kmap_types.h>
diff -Nru a/include/asm-sh64/kmap_types.h b/include/asm-sh64/kmap_types.h
--- /dev/null Wed Dec 31 16:00:00 1969
+++ b/include/asm-sh64/kmap_types.h Sat Jul 19 13:33:30 2003
@@ -0,0 +1 @@
+#include <asm-generic/kmap_types.h>
On Saturday 19 July 2003 13:41, Arnd Bergmann wrote:
Hi Arnd,
> > on alpha:
> > internal.h:19:28: asm/kmap_types.h: No such file or directory
> Same on s390 and some other platforms.We should just get rid of
> this problem by providing a generic kmap_types header.
I agree on this but your patch is bogus. Please use the attached one instead.
Marcello, please apply this one instead. Against 2.4.22-BK (2.4.21-bk14)
ciao, Marc
Marcelo Tosatti <[email protected]> wrote:
| Here goes -pre7.
|
| This is a feature freeze, only bugfixes will be accepted from now on.
OK, do you accept the following bugfix for /proc/cmdline?
--- linux-2.4.21/fs/proc/proc_misc.c.orig Fri Jun 27 11:35:06 2003
+++ linux-2.4.21/fs/proc/proc_misc.c Fri Jun 27 11:37:15 2003
@@ -423,9 +423,9 @@
int count, int *eof, void *data)
{
extern char saved_command_line[];
- int len;
+ int len = 0;
- len = snprintf(page, count, "%s\n", saved_command_line);
+ proc_sprintf(page, &off, &len, "%s\n", saved_command_line);
return proc_calc_metrics(page, start, off, count, eof, len);
}
Since kernel version 2.4.19, a read() from /proc/cmdline with a
non-zero offset doesn't work anymore: try "dd bs=1 < /proc/cmdline".
Because of this bug, the following fails in ash and the busybox shell:
$ read line < /proc/cmdline
I've posted this fix to lkml in januari, and sent it to you several
times, but got no reaction. Alan Cox included this patch in version
2.4.21pre5-ac4 of his tree. Could you please include this obvious fix
in 2.4.22?
--
Dick Streefland //// De Bilt
[email protected] (@ @) The Netherlands
------------------------------oOO--(_)--OOo------------------
On Sat, Jul 19, 2003 at 09:21:03AM +0100, Christoph Hellwig wrote:
> On Sat, Jul 19, 2003 at 01:06:41AM +0200, Felipe Alfaro Solana wrote:
> > > Here goes -pre7.
> >
> > Will ACL/xattr support get its way onto mainstream 2.4 soon?
>
> xattr support is in 2.4 mainline although only JFS currently support
> it.
Does that include ACL Support?
On Mon, Jul 21, 2003 at 10:24:53AM -0700, Mike Fedyk wrote:
> > xattr support is in 2.4 mainline although only JFS currently support
> > it.
>
> Does that include ACL Support?
No.