linux/drivers/net, branch v2.6.13-rc7

[PATCH] emac netpoll fix

2005-08-24T01:43:45Z

netpoll is void(struct net_device *), not int(struct net_device *) Signed-off-by: Al Viro Signed-off-by: Linus Torvalds

[PATCH] Kconfig fix (emac dependencient)

2005-08-24T01:43:43Z

emac doesn't build modular; ibm_emac_debug doesn't build at all (missing headers). Signed-off-by: Al Viro Signed-off-by: Linus Torvalds

[PATCH] Kconfig fix (airo_cs on m32r)

2005-08-24T01:43:41Z

airo_cs is broken on m32r; marked as such. [Proper fix would involve separating PCI-dependent parts and making sure they don't get in the way _and_ arranging for asm/scatterlist.h getting picked on m32r] Signed-off-by: Al Viro Signed-off-by: Linus Torvalds

[PATCH] Kconfig fix (tms380tr and ISA_DMA_API)

2005-08-24T01:43:41Z

ISA parts of tms380tr are using ISA DMA helpers and should depend on ISA_DMA_API. Signed-off-by: Al Viro Signed-off-by: Linus Torvalds

[TG3]: Update driver version and reldate.

2005-08-19T19:57:31Z

Signed-off-by: David S. Miller

[TG3]: Fix SerDes detection

2005-08-19T19:54:29Z

A problem was reported by Grant Grundler on an HP rx8620 using IOX Core LAN partno(A7109-6) 5701 copper NIC. The tg3 driver mistakenly detects this NIC as having a SerDes PHY and link does not come up as a result. The problem was caused by an incorrectly programmed eeprom that set the NIC_SRAM_DATA_CFG_PHY_TYPE_FIBER bit in the NIC_SRAM_DATA_CFG location. This patch will override the NIC_SRAM_DATA_CFG_PHY_TYPE_FIBER bit if a valid PHY ID is read from the MII registers on older 570x chips where the MII interface is not used on SerDes chips. On newer chips such as the 5780 that use MII for both copper and SerDes, SerDes detection must rely on the eeprom. This patch will make the SerDes detection identical to versions 3.25 and older. Signed-off-by: Michael Chan Acked-by: Grant Grundler Signed-off-by: David S. Miller

Merge head 'upstream-fixes' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6

2005-08-18T21:58:21Z

[NET]: Fix comment in loopback driver.

2005-08-18T21:05:18Z

Signed-off-by: Ralf Baechle Signed-off-by: Andrew Morton Signed-off-by: David S. Miller

[PATCH] DM9000 - incorrect ioctl() handling

2005-08-18T20:59:14Z

The DM9000 driver is responding to ioctl() calls it should not be. This can cause problems with the wireless tools incorrectly indentifying the device as wireless capable, and crashing under certain operations. This patch also moves the version printk() to the init call, so that you only get it once for multiple devices, and to show it is loaded if there are no defined dm9000s Signed-off-by: Ben Dooks Signed-off-by: Jeff Garzik

[PATCH] DM9000 - spinlock fixes

2005-08-18T20:59:14Z

Fix DM9000 driver usage of spinlocks, which mainly came to light when running a kernel with spinlock debugging. These come down to: 1) Un-initialised spin lock 2) Several cases of using spin_xxx(lock) and not spin_xxx(&lock) 3) move the locking around the phy reg for read/write to only keep the lock when actually reading or writing to the phy. Signed-off-by: Ben Dooks Signed-off-by: Jeff Garzik