2002-10-04 01:43:20

by Marcelo Tosatti

[permalink] [raw]
Subject: Linux 2.4.20-pre9

The Athlon problems introduced in pre3 should be gone now.

pre9 has some JFS/ext3 fixes, USB fixes and several network drivers fixes.

There are still some pending issues to be solved for 2.4.20 which I hope
get worked out on the next -pre's...

Summary of changes from v2.4.20-pre8 to v2.4.20-pre9

<[email protected]>:
o fs/partitions/sun.c: raid autodetect for sun disk labels

<[email protected]>:
o Update for JMTek USBDrive

<[email protected]>:
o net/sched/sch_htb.c: Verify classid and direct_qlen properly

<[email protected]>:
o Support get-MII-data ioctls in 8139cp net driver

<[email protected]>:
o fix drm ioctl ABI default

<[email protected]>:
o pc_keyb.c: hook for keybd controller detection
o Configure.help update (1/2)
o AGPGART 2/5: size AGP mem correctly when memory is
o ati_pcigart: support 16K and 64K page size
o Configure.help update (2/2)
o PCI ID database update

<[email protected]>:
o acenic net drvr fix: remove '=' typo in intr mask writel()

<[email protected]>:
o Place buffer dirtied in truncate() on inode's dirty data list (Eric Sandeen)
o Enable CONFIG_DRM_I810_XFREE_41 so we are compatible with XFree 4.1 as default
o Changed EXTRAVERSION to -pre9

<[email protected]>:
o USB 2.0 HDD Walker / ST-HW-818SLIM usb-storage fix

<[email protected]>:
o net/ipv4/proc.c: Dont print dummy member of icmp_mib

<[email protected]>:
o 2.4.20-pre4/ext3: Bump ext3 version number
o 2.4.20-pre4/ext3: Fix LVM snapshot deadlock
o 2.4.20-pre4/ext3: jbd commit interval tuning
o Sanity check for Intermezzo/ext3
o ext3 commit notification for Intermezzo
o Fix the order of inodes being marked dirty in a couple of corner cases

<[email protected]>:
o Kernel TUN/TAP Documentation rework

<[email protected]>:
o [IPv6]: Verify ND options properly
o net/ipv6/addrconf.c: Refine IPv6 Address Validation Timer
o net/ipv6/ndisc.c: Add missing credits
o net/ipv6/ip6_fib.c: Default route support on router

Adrian Bunk <[email protected]>:
o add "If unsure, say N" to CONFIG_X86_TSC_DISABLE
o Configure.help entry for the e100 driver (fwd)

Alan Cox <[email protected]>:
o security - various bits in iee1394
o Fix config.in breakage from mips people
o resend maestro3 update

Andi Kleen <[email protected]>:
o Fix disabling of x86 capabilities from command line
o Fix pageattr with mem=nopentium

Dave Kleikamp <[email protected]>:
o JFS: Fix problems with NFS
o JFS: detect and fix invalid directory index values
o JFS: Remove assert(i < MAX_ACTIVE)

David S. Miller <[email protected]>:
o net/ipv4/netfilter/ip_conntrack_proto_tcp.c: Include linux/string.h
o drivers/block/ll_rw_blk.c: u64 is not necessarily long long
o init/do_mounts.c: Protect cramfs stuff with CONFIG_BLK_DEV_RAM too
o drivers/net/acenic.h: readl is not an lvalue
o drivers/net/ppp_generic.c:ppp_receive_frame Kill unused local label

Edward Peng <[email protected]>:
o update sundance driver to support building on older kernel

Greg Kroah-Hartman <[email protected]>:
o USB: added Palm Zire support to the visor driver

Harald Welte <[email protected]>:
o [NETFILTER]: Trivial fixes

James Morris <[email protected]>:
o net/ipv4/netfilter/ipchains_core.c: Use GFP_ATOMIC under ip_fw_lock

Javier Achirica <[email protected]>:
o sync airo wireless driver with 2.5
o airo wireless: use ETH_ALEN constant where appropriate
o airo wireless: disable access to card while prom flashing in progress [note: more work needs to be done here, but this is better than nothing -jgarzik]
o airo wireless: more verbose MAC-enable errors
o airo wireless: power down on if down. add local 'ai' to fix build
o airo wireless: fix "non-probe mode" setup
o airo wireless: Fixes signal level retrieval in SPY mode (releases memory block after read out)
o airo wireless net drvr: add Cisco MIC support Conditionally enabled when out-of-tree, but open source, crypto lib is present.

Jean Tourrilhes <[email protected]>:
o irtty MODEM_BITS additional fix

Jeff Garzik <[email protected]>:
o Update eepro100 net driver's mdio_{read,write} functions to take 'struct net_device *' not 'long' as their first argument. This makes eepro100 compatible with the standard MII ethtool API, preparing it for that support.
o update eepro100 net driver to use standard MII phy API/lib, when implementing ethtool media ioctls.
o Add new MII lib functions mii_check_link, mii_check_media
o sundance net drvr: fix reset_tx logic (contributed by Edward Peng @ D-Link, cleaned up by me)
o sundance net drvr: fix DFE-580TX packet drop issue, further reset_tx fixes (contributed by Edward Peng @ D-Link)
o sundance net drvr: bump version to LK1.05
o [net drivers] fix MII lib force-media ethtool path (contributed by Edward Peng @ D-Link)
o sis900 net driver update
o [net drivers] MII lib update
o [net drivers] Rename MII lib API member, s/duplex_lock/force_media/, and update all drivers that reference this struct member.
o Add MII lib helper func generic_mii_ioctl, use it in 8139cp net drvr
o Use new MII lib helper generic_mii_ioctl in several net drivers
o [net drivers] Remove 'dev' argument from generic_mii_ioctl helper
o [net drivers] add optional duplex-changed arg to generic_mii_ioctl helper

Oliver Neukum <oli[email protected]>:
o USB: update of hpusbscsi
o USB: fixes for kaweth

Petr Vandrovec <[email protected]>:
o [NCPFS]: 32->64bit sparc64 conversions

Richard Henderson <[email protected]>:
o alpha strncpy fix

Rusty Russell <[email protected]>:
o Remove list_head typedef

Tim Schmielau <[email protected]>:
o Fix sb1000 jiffies usage: kill float constant, use time_after_eq()
o fix compares of jiffies

2002-10-06 20:55:11

by Marcus Sundberg

[permalink] [raw]
Subject: DFE-580TX packet drop persist (Re: Linux 2.4.20-pre9)

Marcelo Tosatti <[email protected]> writes:

> Summary of changes from v2.4.20-pre8 to v2.4.20-pre9
> ============================================

> o sundance net drvr: fix DFE-580TX packet drop issue, further reset_tx fixes (contributed by Edward Peng @ D-Link)


I'm sorry to say that the packet drop issue is still not solved
with a DF580-TX using four of these chips:
Class 0200: 1186:1002 (rev 12)
Subsystem: 1186:1012

Sending bi-directional streams of UDP-packets with 200-byte payload
through the machine I get zero packet loss and 15% CPU usage with
either eepro100 or tulip-based cards. Using the DFE-580TX I get
15% packet loss and 25% CPU usage. Ifconfig also shows packets
being dropped:

eth1 Link encap:Ethernet HWaddr 00:05:5D:E6:14:BE
inet addr: Bcast: Mask:
RX packets:2439797 errors:0 dropped:287381 overruns:0 frame:0
TX packets:2461746 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
Interrupt:18 Base address:0xa400

eth2 Link encap:Ethernet HWaddr 00:05:5D:E6:14:BF
inet addr: Bcast: Mask:
RX packets:2463601 errors:0 dropped:286106 overruns:0 frame:0
TX packets:2437637 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
Interrupt:19 Base address:0xa000

Marcus Sundberg <[email protected]> | Firewalls with SIP & NAT
Firewall Developer, Ingate Systems AB | http://www.ingate.com/