From ec04fd60fd74a9db9c63fe11d519be3642cabfdd Mon Sep 17 00:00:00 2001 From: Rick Farrington Date: Tue, 1 Jul 2008 09:20:33 +0800 Subject: [PATCH 01/46] iwlwifi: fix incorrect 5GHz rates reported in monitor mode This patch fixes the rates reported in monitor mode operation (Wireshark) for iwlwifi. Previously, packets with rates of 6M..24M would be reported incorrectly and packets with rates of 36M..54M would not passed up the stack. Signed-off-by: Rick Farrington Signed-off-by: Zhu Yi Signed-off-by: John W. Linville --- drivers/net/wireless/iwlwifi/iwl-3945.c | 6 +++++- drivers/net/wireless/iwlwifi/iwl-4965.c | 6 +++++- 2 files changed, 10 insertions(+), 2 deletions(-) diff --git a/drivers/net/wireless/iwlwifi/iwl-3945.c b/drivers/net/wireless/iwlwifi/iwl-3945.c index 62a3d8f8563e..f5387a7a76c0 100644 --- a/drivers/net/wireless/iwlwifi/iwl-3945.c +++ b/drivers/net/wireless/iwlwifi/iwl-3945.c @@ -588,8 +588,12 @@ static void iwl3945_add_radiotap(struct iwl3945_priv *priv, if (rate == -1) iwl3945_rt->rt_rate = 0; - else + else { + if (stats->band == IEEE80211_BAND_5GHZ) + rate += IWL_FIRST_OFDM_RATE; + iwl3945_rt->rt_rate = iwl3945_rates[rate].ieee; + } /* antenna number */ antenna = phy_flags_hw & RX_RES_PHY_FLAGS_ANTENNA_MSK; diff --git a/drivers/net/wireless/iwlwifi/iwl-4965.c b/drivers/net/wireless/iwlwifi/iwl-4965.c index bf19eb8aafd0..de330ae0ca95 100644 --- a/drivers/net/wireless/iwlwifi/iwl-4965.c +++ b/drivers/net/wireless/iwlwifi/iwl-4965.c @@ -3528,8 +3528,12 @@ static void iwl4965_add_radiotap(struct iwl_priv *priv, if (rate == -1) iwl4965_rt->rt_rate = 0; - else + else { + if (stats->band == IEEE80211_BAND_5GHZ) + rate += IWL_FIRST_OFDM_RATE; + iwl4965_rt->rt_rate = iwl4965_rates[rate].ieee; + } /* * "antenna number" From 6afe6828b19b4567768264831d101026cb5510ff Mon Sep 17 00:00:00 2001 From: Zhu Yi Date: Tue, 1 Jul 2008 09:20:34 +0800 Subject: [PATCH 02/46] iwlwifi: drop skb silently for Tx request in monitor mode This patch fixes the problem to keep mac80211 resubmitting SKBs when Tx request cannot be met in monitor mode. Signed-off-by: Zhu Yi Signed-off-by: John W. Linville --- drivers/net/wireless/iwlwifi/iwl3945-base.c | 3 ++- drivers/net/wireless/iwlwifi/iwl4965-base.c | 3 ++- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/net/wireless/iwlwifi/iwl3945-base.c b/drivers/net/wireless/iwlwifi/iwl3945-base.c index b1b3c523185d..6027e1119c3f 100644 --- a/drivers/net/wireless/iwlwifi/iwl3945-base.c +++ b/drivers/net/wireless/iwlwifi/iwl3945-base.c @@ -6687,7 +6687,8 @@ static int iwl3945_mac_tx(struct ieee80211_hw *hw, struct sk_buff *skb, if (priv->iw_mode == IEEE80211_IF_TYPE_MNTR) { IWL_DEBUG_MAC80211("leave - monitor\n"); - return -1; + dev_kfree_skb_any(skb); + return 0; } IWL_DEBUG_TX("dev->xmit(%d bytes) at rate 0x%02x\n", skb->len, diff --git a/drivers/net/wireless/iwlwifi/iwl4965-base.c b/drivers/net/wireless/iwlwifi/iwl4965-base.c index 5ed16ce78468..0bd55bb19739 100644 --- a/drivers/net/wireless/iwlwifi/iwl4965-base.c +++ b/drivers/net/wireless/iwlwifi/iwl4965-base.c @@ -6237,7 +6237,8 @@ static int iwl4965_mac_tx(struct ieee80211_hw *hw, struct sk_buff *skb, if (priv->iw_mode == IEEE80211_IF_TYPE_MNTR) { IWL_DEBUG_MAC80211("leave - monitor\n"); - return -1; + dev_kfree_skb_any(skb); + return 0; } IWL_DEBUG_TX("dev->xmit(%d bytes) at rate 0x%02x\n", skb->len, From 7b58ccfe32f40eca8c8ca29aa723a5d0e814f0c9 Mon Sep 17 00:00:00 2001 From: "andrey@cozybit.com" Date: Tue, 1 Jul 2008 11:43:53 -0700 Subject: [PATCH 03/46] libertas: support USB persistence on suspend/resume (resend) Handle .reset_resume() so that libertas can survive suspend/resume without reloading the firmware. Signed-off-by: Andrey Yurovsky Acked-by: Deepak Saxena Acked-by: Dan Williams Signed-off-by: John W. Linville --- drivers/net/wireless/libertas/if_usb.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/wireless/libertas/if_usb.c b/drivers/net/wireless/libertas/if_usb.c index 8032df72aaab..36288b29abf7 100644 --- a/drivers/net/wireless/libertas/if_usb.c +++ b/drivers/net/wireless/libertas/if_usb.c @@ -925,6 +925,7 @@ static struct usb_driver if_usb_driver = { .id_table = if_usb_table, .suspend = if_usb_suspend, .resume = if_usb_resume, + .reset_resume = if_usb_resume, }; static int __init if_usb_init_module(void) From 81b23b4a7acd9b37a269c62d02479d4f645dd20a Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Thu, 3 Jul 2008 03:22:02 -0700 Subject: [PATCH 04/46] tcp: net/ipv4/tcp.c needs linux/scatterlist.h alpha: net/ipv4/tcp.c: In function 'tcp_calc_md5_hash': net/ipv4/tcp.c:2479: error: implicit declaration of function 'sg_init_table' net/ipv4/tcp.c:2482: error: implicit declaration of function 'sg_set_buf' net/ipv4/tcp.c:2507: error: implicit declaration of function 'sg_mark_end' Signed-off-by: Andrew Morton Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 850825dc86e6..de53024664e4 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -255,6 +255,7 @@ #include #include #include +#include #include #include #include From 374e7b59498ce0785b3727794b351221528a5159 Mon Sep 17 00:00:00 2001 From: Octavian Purdila Date: Thu, 3 Jul 2008 03:31:21 -0700 Subject: [PATCH 05/46] tcp: fix a size_t < 0 comparison in tcp_read_sock should be of type int (not size_t) since recv_actor can return negative values and it is also used in a < 0 comparison. Signed-off-by: Octavian Purdila Signed-off-by: David S. Miller --- net/ipv4/tcp.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index de53024664e4..1d723de18686 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -1209,7 +1209,8 @@ int tcp_read_sock(struct sock *sk, read_descriptor_t *desc, return -ENOTCONN; while ((skb = tcp_recv_skb(sk, seq, &offset)) != NULL) { if (offset < skb->len) { - size_t used, len; + int used; + size_t len; len = skb->len - offset; /* Stop reading if we hit a patch of urgent data */ From ab1b20467cd2214ad89a95d007047cd2a6b5bf5d Mon Sep 17 00:00:00 2001 From: Patrick McHardy Date: Thu, 3 Jul 2008 03:53:42 -0700 Subject: [PATCH 06/46] bridge: fix use-after-free in br_cleanup_bridges() Unregistering a bridge device may cause virtual devices stacked on the bridge, like vlan or macvlan devices, to be unregistered as well. br_cleanup_bridges() uses for_each_netdev_safe() to iterate over all devices during cleanup. This is not enough however, if one of the additionally unregistered devices is next in the list to the bridge device, it will get freed as well and the iteration continues on the freed element. Restart iteration after each bridge device removal from the beginning to fix this, similar to what rtnl_link_unregister() does. Signed-off-by: Patrick McHardy Acked-by: Stephen Hemminger Signed-off-by: David S. Miller --- net/bridge/br_if.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index c2397f503b0f..f38cc5317b88 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -442,12 +442,16 @@ int br_del_if(struct net_bridge *br, struct net_device *dev) void __exit br_cleanup_bridges(void) { - struct net_device *dev, *nxt; + struct net_device *dev; rtnl_lock(); - for_each_netdev_safe(&init_net, dev, nxt) - if (dev->priv_flags & IFF_EBRIDGE) +restart: + for_each_netdev(&init_net, dev) { + if (dev->priv_flags & IFF_EBRIDGE) { del_br(dev->priv); + goto restart; + } + } rtnl_unlock(); } From 54364b752ef0c91fe92684df000cc4593d1e8963 Mon Sep 17 00:00:00 2001 From: Krzysztof Halasa Date: Sun, 29 Jun 2008 21:48:11 +0200 Subject: [PATCH 07/46] Add missing skb->dev assignment in Frame Relay RX code Commit 4c13eb6657fe9ef7b4dc8f1a405c902e9e5234e0 ([ETH]: Make eth_type_trans set skb->dev like the other *_type_trans) removed skb->dev assignment from hdlc_fr.c:fr_rx(). Unfortunately it was also needed for cases other than eth_type_trans(). Adding it back. It's quite serious and may be a security risk as it causes a wrong input interface indication (the physical hdlcX instead of logical pvcX). Probably -stable class fix. Signed-off-by: Krzysztof Halasa Signed-off-by: Jeff Garzik --- drivers/net/wan/hdlc_fr.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/wan/hdlc_fr.c b/drivers/net/wan/hdlc_fr.c index 520bb0b1a9a2..6d35155c7145 100644 --- a/drivers/net/wan/hdlc_fr.c +++ b/drivers/net/wan/hdlc_fr.c @@ -1008,6 +1008,7 @@ static int fr_rx(struct sk_buff *skb) stats->rx_bytes += skb->len; if (pvc->state.becn) stats->rx_compressed++; + skb->dev = dev; netif_rx(skb); return NET_RX_SUCCESS; } else { From 97bff0953dd45a633fa69e1a650d612f5610a60b Mon Sep 17 00:00:00 2001 From: Tobias Diedrich Date: Thu, 3 Jul 2008 23:54:56 -0700 Subject: [PATCH 08/46] forcedeth: fix lockdep warning on ethtool -s After enabling CONFIG_LOCKDEP and CONFIG_PROVE_LOCKING I get the following warning when ethtool -s is first called on one of the forcedeth ports: ================================= [ INFO: inconsistent lock state ] 2.6.26-rc4 #28 --------------------------------- inconsistent {in-hardirq-W} -> {hardirq-on-W} usage. ethtool/1985 [HC0[0]:SC0[1]:HE1:SE0] takes: (&np->lock){++..}, at: [] nv_set_settings+0xc8/0x3de [forcedeth] {in-hardirq-W} state was registered at: [] 0xffffffffffffffff irq event stamp: 3606 hardirqs last enabled at (3605): [] _spin_unlock_irqrestore+0x3f/0x68 hardirqs last disabled at (3604): [] _spin_lock_irqsave+0x13/0x46 softirqs last enabled at (3534): [] __do_softirq+0xbc/0xc5 softirqs last disabled at (3606): [] _spin_lock_bh+0x11/0x41 other info that might help us debug this: 2 locks held by ethtool/1985: #0: (rtnl_mutex){--..}, at: [] rtnl_lock+0x12/0x14 #1: (_xmit_ETHER){-+..}, at: [] nv_set_settings+0xb3/0x3de [forcedeth] stack backtrace: Pid: 1985, comm: ethtool Not tainted 2.6.26-rc4 #28 Call Trace: [] print_usage_bug+0x162/0x173 [] mark_lock+0x231/0x41f [] __lock_acquire+0x4e7/0xcac [] ? trace_hardirqs_on+0xf1/0x115 [] ? disable_irq_nosync+0x6f/0x7b [] lock_acquire+0x55/0x6e [] ? :forcedeth:nv_set_settings+0xc8/0x3de [] _spin_lock+0x2f/0x3c [] :forcedeth:nv_set_settings+0xc8/0x3de [] dev_ethtool+0x186/0xea3 [] ? mutex_lock_nested+0x243/0x275 [] ? debug_mutex_free_waiter+0x46/0x4a [] ? mutex_lock_nested+0x266/0x275 [] dev_ioctl+0x4eb/0x600 [] ? _spin_unlock_irqrestore+0x3f/0x68 [] sock_ioctl+0x1f5/0x202 [] vfs_ioctl+0x2a/0x77 [] do_vfs_ioctl+0x25b/0x270 [] ? trace_hardirqs_on_thunk+0x35/0x3a [] sys_ioctl+0x42/0x65 [] system_call_after_swapgs+0x7b/0x80 This is caused by the following snippet in nv_set_settings: netif_carrier_off(dev); if (netif_running(dev)) { nv_disable_irq(dev); netif_tx_lock_bh(dev); spin_lock(&np->lock); /* stop engines */ nv_stop_rxtx(dev); spin_unlock(&np->lock); netif_tx_unlock_bh(dev); } Because of nv_disable_irq this is probably not really a problem though (I guess) and replacing the spin_lock with spin_lock_irqsave could keep interrupts disabled for a longer period of time because of delays in nv_stop_rx and nv_stop_tx. Signed-off-by: Tobias Diedrich Cc: Ayaz Abdulla Signed-off-by: Andrew Morton Signed-off-by: Jeff Garzik --- drivers/net/forcedeth.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/drivers/net/forcedeth.c b/drivers/net/forcedeth.c index 2cb244763292..20d4fe96a81c 100644 --- a/drivers/net/forcedeth.c +++ b/drivers/net/forcedeth.c @@ -4194,12 +4194,23 @@ static int nv_set_settings(struct net_device *dev, struct ethtool_cmd *ecmd) netif_carrier_off(dev); if (netif_running(dev)) { + unsigned long flags; + nv_disable_irq(dev); netif_tx_lock_bh(dev); - spin_lock(&np->lock); + /* with plain spinlock lockdep complains */ + spin_lock_irqsave(&np->lock, flags); /* stop engines */ + /* FIXME: + * this can take some time, and interrupts are disabled + * due to spin_lock_irqsave, but let's hope no daemon + * is going to change the settings very often... + * Worst case: + * NV_RXSTOP_DELAY1MAX + NV_TXSTOP_DELAY1MAX + * + some minor delays, which is up to a second approximately + */ nv_stop_rxtx(dev); - spin_unlock(&np->lock); + spin_unlock_irqrestore(&np->lock, flags); netif_tx_unlock_bh(dev); } From 5c2cec143ac54c1960e54bc320fa7d13ac8e0f4a Mon Sep 17 00:00:00 2001 From: Jan-Bernd Themann Date: Thu, 3 Jul 2008 15:18:45 +0100 Subject: [PATCH 09/46] ehea: fix might sleep problem A mutex has to be replaced by spinlocks as it can be called from a context which does not allow sleeping. The kzalloc flag GFP_KERNEL has to be replaced by GFP_ATOMIC for the same reason. Signed-off-by: Jan-Bernd Themann Signed-off-by: Jeff Garzik --- drivers/net/ehea/ehea.h | 2 +- drivers/net/ehea/ehea_main.c | 22 +++++++++++----------- 2 files changed, 12 insertions(+), 12 deletions(-) diff --git a/drivers/net/ehea/ehea.h b/drivers/net/ehea/ehea.h index fe872fbd671e..bf57e1532f5e 100644 --- a/drivers/net/ehea/ehea.h +++ b/drivers/net/ehea/ehea.h @@ -452,7 +452,7 @@ struct ehea_bcmc_reg_entry { struct ehea_bcmc_reg_array { struct ehea_bcmc_reg_entry *arr; int num_entries; - struct mutex lock; + spinlock_t lock; }; #define EHEA_PORT_UP 1 diff --git a/drivers/net/ehea/ehea_main.c b/drivers/net/ehea/ehea_main.c index 075fd547421e..489de9ba5fc0 100644 --- a/drivers/net/ehea/ehea_main.c +++ b/drivers/net/ehea/ehea_main.c @@ -241,7 +241,7 @@ static void ehea_update_bcmc_registrations(void) } if (num_registrations) { - arr = kzalloc(num_registrations * sizeof(*arr), GFP_KERNEL); + arr = kzalloc(num_registrations * sizeof(*arr), GFP_ATOMIC); if (!arr) return; /* Keep the existing array */ } else @@ -301,7 +301,7 @@ static struct net_device_stats *ehea_get_stats(struct net_device *dev) memset(stats, 0, sizeof(*stats)); - cb2 = kzalloc(PAGE_SIZE, GFP_KERNEL); + cb2 = kzalloc(PAGE_SIZE, GFP_ATOMIC); if (!cb2) { ehea_error("no mem for cb2"); goto out; @@ -1763,7 +1763,7 @@ static int ehea_set_mac_addr(struct net_device *dev, void *sa) memcpy(dev->dev_addr, mac_addr->sa_data, dev->addr_len); - mutex_lock(&ehea_bcmc_regs.lock); + spin_lock(&ehea_bcmc_regs.lock); /* Deregister old MAC in pHYP */ if (port->state == EHEA_PORT_UP) { @@ -1785,7 +1785,7 @@ static int ehea_set_mac_addr(struct net_device *dev, void *sa) out_upregs: ehea_update_bcmc_registrations(); - mutex_unlock(&ehea_bcmc_regs.lock); + spin_unlock(&ehea_bcmc_regs.lock); out_free: kfree(cb0); out: @@ -1947,7 +1947,7 @@ static void ehea_set_multicast_list(struct net_device *dev) } ehea_promiscuous(dev, 0); - mutex_lock(&ehea_bcmc_regs.lock); + spin_lock(&ehea_bcmc_regs.lock); if (dev->flags & IFF_ALLMULTI) { ehea_allmulti(dev, 1); @@ -1978,7 +1978,7 @@ static void ehea_set_multicast_list(struct net_device *dev) } out: ehea_update_bcmc_registrations(); - mutex_unlock(&ehea_bcmc_regs.lock); + spin_unlock(&ehea_bcmc_regs.lock); return; } @@ -2497,7 +2497,7 @@ static int ehea_up(struct net_device *dev) } } - mutex_lock(&ehea_bcmc_regs.lock); + spin_lock(&ehea_bcmc_regs.lock); ret = ehea_broadcast_reg_helper(port, H_REG_BCMC); if (ret) { @@ -2520,7 +2520,7 @@ out: ehea_info("Failed starting %s. ret=%i", dev->name, ret); ehea_update_bcmc_registrations(); - mutex_unlock(&ehea_bcmc_regs.lock); + spin_unlock(&ehea_bcmc_regs.lock); ehea_update_firmware_handles(); mutex_unlock(&ehea_fw_handles.lock); @@ -2575,7 +2575,7 @@ static int ehea_down(struct net_device *dev) mutex_lock(&ehea_fw_handles.lock); - mutex_lock(&ehea_bcmc_regs.lock); + spin_lock(&ehea_bcmc_regs.lock); ehea_drop_multicast_list(dev); ehea_broadcast_reg_helper(port, H_DEREG_BCMC); @@ -2584,7 +2584,7 @@ static int ehea_down(struct net_device *dev) port->state = EHEA_PORT_DOWN; ehea_update_bcmc_registrations(); - mutex_unlock(&ehea_bcmc_regs.lock); + spin_unlock(&ehea_bcmc_regs.lock); ret = ehea_clean_all_portres(port); if (ret) @@ -3590,7 +3590,7 @@ int __init ehea_module_init(void) memset(&ehea_bcmc_regs, 0, sizeof(ehea_bcmc_regs)); mutex_init(&ehea_fw_handles.lock); - mutex_init(&ehea_bcmc_regs.lock); + spin_lock_init(&ehea_bcmc_regs.lock); ret = check_module_parm(); if (ret) From b0afffe89be619f42ae4215554ed66e67de7bb0e Mon Sep 17 00:00:00 2001 From: Jan-Bernd Themann Date: Thu, 3 Jul 2008 15:18:48 +0100 Subject: [PATCH 10/46] ehea: add MODULE_DEVICE_TABLE Required to allow distros to easily detect when ehea module needs to be loaded Signed-off-by: Jan-Bernd Themann Signed-off-by: Jeff Garzik --- drivers/net/ehea/ehea_main.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ehea/ehea_main.c b/drivers/net/ehea/ehea_main.c index 489de9ba5fc0..6ebcb221cd5e 100644 --- a/drivers/net/ehea/ehea_main.c +++ b/drivers/net/ehea/ehea_main.c @@ -118,6 +118,7 @@ static struct of_device_id ehea_device_table[] = { }, {}, }; +MODULE_DEVICE_TABLE(of, ehea_device_table); static struct of_platform_driver ehea_driver = { .name = "ehea", From 2f69ae01c83a94af5dc3c20e8135b974687ed004 Mon Sep 17 00:00:00 2001 From: Jan-Bernd Themann Date: Thu, 3 Jul 2008 15:18:51 +0100 Subject: [PATCH 11/46] ehea: fix race condition When ehea_stop is called the function cancel_work_sync(&port->reset_task) is used to ensure that the reset task is not running anymore. We need an additional flag to ensure that it can not be scheduled after this call again for a certain time. Signed-off-by: Jan-Bernd Themann Signed-off-by: Jeff Garzik --- drivers/net/ehea/ehea.h | 6 ++++-- drivers/net/ehea/ehea_main.c | 17 ++++++++++++----- 2 files changed, 16 insertions(+), 7 deletions(-) diff --git a/drivers/net/ehea/ehea.h b/drivers/net/ehea/ehea.h index bf57e1532f5e..e01926b7b5b7 100644 --- a/drivers/net/ehea/ehea.h +++ b/drivers/net/ehea/ehea.h @@ -40,7 +40,7 @@ #include #define DRV_NAME "ehea" -#define DRV_VERSION "EHEA_0091" +#define DRV_VERSION "EHEA_0092" /* eHEA capability flags */ #define DLPAR_PORT_ADD_REM 1 @@ -478,6 +478,7 @@ struct ehea_port { int num_add_tx_qps; int num_mcs; int resets; + u64 flags; u64 mac_addr; u32 logical_port_id; u32 port_speed; @@ -501,7 +502,8 @@ struct port_res_cfg { }; enum ehea_flag_bits { - __EHEA_STOP_XFER + __EHEA_STOP_XFER, + __EHEA_DISABLE_PORT_RESET }; void ehea_set_ethtool_ops(struct net_device *netdev); diff --git a/drivers/net/ehea/ehea_main.c b/drivers/net/ehea/ehea_main.c index 6ebcb221cd5e..49ba6a9a7b06 100644 --- a/drivers/net/ehea/ehea_main.c +++ b/drivers/net/ehea/ehea_main.c @@ -138,6 +138,12 @@ void ehea_dump(void *adr, int len, char *msg) } } +void ehea_schedule_port_reset(struct ehea_port *port) +{ + if (!test_bit(__EHEA_DISABLE_PORT_RESET, &port->flags)) + schedule_work(&port->reset_task); +} + static void ehea_update_firmware_handles(void) { struct ehea_fw_handle_entry *arr = NULL; @@ -588,7 +594,7 @@ static int ehea_treat_poll_error(struct ehea_port_res *pr, int rq, "Resetting port.", pr->qp->init_attr.qp_nr); ehea_dump(cqe, sizeof(*cqe), "CQE"); } - schedule_work(&pr->port->reset_task); + ehea_schedule_port_reset(pr->port); return 1; } @@ -766,7 +772,7 @@ static struct ehea_cqe *ehea_proc_cqes(struct ehea_port_res *pr, int my_quota) ehea_error("Send Completion Error: Resetting port"); if (netif_msg_tx_err(pr->port)) ehea_dump(cqe, sizeof(*cqe), "Send CQE"); - schedule_work(&pr->port->reset_task); + ehea_schedule_port_reset(pr->port); break; } @@ -886,7 +892,7 @@ static irqreturn_t ehea_qp_aff_irq_handler(int irq, void *param) eqe = ehea_poll_eq(port->qp_eq); } - schedule_work(&port->reset_task); + ehea_schedule_port_reset(port); return IRQ_HANDLED; } @@ -2606,13 +2612,14 @@ static int ehea_stop(struct net_device *dev) if (netif_msg_ifdown(port)) ehea_info("disabling port %s", dev->name); + set_bit(__EHEA_DISABLE_PORT_RESET, &port->flags); cancel_work_sync(&port->reset_task); - mutex_lock(&port->port_lock); netif_stop_queue(dev); port_napi_disable(port); ret = ehea_down(dev); mutex_unlock(&port->port_lock); + clear_bit(__EHEA_DISABLE_PORT_RESET, &port->flags); return ret; } @@ -2942,7 +2949,7 @@ static void ehea_tx_watchdog(struct net_device *dev) if (netif_carrier_ok(dev) && !test_bit(__EHEA_STOP_XFER, &ehea_driver_flags)) - schedule_work(&port->reset_task); + ehea_schedule_port_reset(port); } int ehea_sense_adapter_attr(struct ehea_adapter *adapter) From 3ff2cd230de31e3b7bf2efd254e0393e5fc1d15d Mon Sep 17 00:00:00 2001 From: Roland Dreier Date: Tue, 1 Jul 2008 10:20:33 -0700 Subject: [PATCH 12/46] ehea: Access iph->tot_len with correct endianness iph->tot_len is stored in network byte order, so access it using ntohs(). This doesn't have any real world impact on ehea, since ehea only exists for big-endian platfroms (at the moment at least) but fixing this gets rid of a sparse warning and avoids having a bad example in the tree. Signed-off-by: Roland Dreier Signed-off-by: Jeff Garzik --- drivers/net/ehea/ehea_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ehea/ehea_main.c b/drivers/net/ehea/ehea_main.c index 49ba6a9a7b06..0920b796bd78 100644 --- a/drivers/net/ehea/ehea_main.c +++ b/drivers/net/ehea/ehea_main.c @@ -623,7 +623,7 @@ static int get_skb_hdr(struct sk_buff *skb, void **iphdr, *tcph = tcp_hdr(skb); /* check if ip header and tcp header are complete */ - if (iph->tot_len < ip_len + tcp_hdrlen(skb)) + if (ntohs(iph->tot_len) < ip_len + tcp_hdrlen(skb)) return -1; *hdr_flags = LRO_IPV4 | LRO_TCP; From 773212337941c5e26a05989532943877d72a2c83 Mon Sep 17 00:00:00 2001 From: Roland Dreier Date: Tue, 1 Jul 2008 10:22:45 -0700 Subject: [PATCH 13/46] pasemi_mac: Access iph->tot_len with correct endianness iph->tot_len is stored in network byte order, so access it using ntohs(). This doesn't have any real world impact on pasemi_mac, since the device only exists as part of a big-endian system-on-chip, but fixing this gets rid of a sparse warning and avoids having a bad example in the tree. Signed-off-by: Roland Dreier Signed-off-by: Jeff Garzik --- drivers/net/pasemi_mac.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/pasemi_mac.c b/drivers/net/pasemi_mac.c index 3b2a6c598088..993d87c9296f 100644 --- a/drivers/net/pasemi_mac.c +++ b/drivers/net/pasemi_mac.c @@ -277,7 +277,7 @@ static int get_skb_hdr(struct sk_buff *skb, void **iphdr, *tcph = tcp_hdr(skb); /* check if ip header and tcp header are complete */ - if (iph->tot_len < ip_len + tcp_hdrlen(skb)) + if (ntohs(iph->tot_len) < ip_len + tcp_hdrlen(skb)) return -1; *hdr_flags = LRO_IPV4 | LRO_TCP; From ab9b30cc3ec868fab8764d710193107fbeedbd0f Mon Sep 17 00:00:00 2001 From: Sathya Narayanan Date: Tue, 1 Jul 2008 10:58:05 +0200 Subject: [PATCH 14/46] ibm_newemac: Fixes kernel crashes when speed of cable connected changes The descriptor pointers were not initialized to NIL values, so it was poiniting to some random addresses which was completely invalid. This fix takes care of initializing the descriptor to NIL values and clearing the valid descriptors on clean ring operation. Signed-off-by: Sathya Narayanan Signed-off-by: Stefan Roese Signed-off-by: Jeff Garzik --- drivers/net/ibm_newemac/core.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/ibm_newemac/core.c b/drivers/net/ibm_newemac/core.c index 5d2108c5ac7c..ca63fa000ee7 100644 --- a/drivers/net/ibm_newemac/core.c +++ b/drivers/net/ibm_newemac/core.c @@ -2719,6 +2719,8 @@ static int __devinit emac_probe(struct of_device *ofdev, /* Clean rings */ memset(dev->tx_desc, 0, NUM_TX_BUFF * sizeof(struct mal_descriptor)); memset(dev->rx_desc, 0, NUM_RX_BUFF * sizeof(struct mal_descriptor)); + memset(dev->tx_skb, 0, NUM_TX_BUFF * sizeof(struct sk_buff *)); + memset(dev->rx_skb, 0, NUM_RX_BUFF * sizeof(struct sk_buff *)); /* Attach to ZMII, if needed */ if (emac_has_feature(dev, EMAC_FTR_HAS_ZMII) && From 6c688f4294dec3f2228fd46be67604508177a1c3 Mon Sep 17 00:00:00 2001 From: Sathya Narayanan Date: Tue, 1 Jul 2008 10:58:19 +0200 Subject: [PATCH 15/46] ibm_newemac: Fixes entry of short packets Short packets has to be discarded by the driver. So this patch addresses the issue of discarding the short packets of size lesser then ethernet header size. Signed-off-by: Sathya Narayanan Signed-off-by: Stefan Roese Signed-off-by: Jeff Garzik --- drivers/net/ibm_newemac/core.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/net/ibm_newemac/core.c b/drivers/net/ibm_newemac/core.c index ca63fa000ee7..babc79ad490b 100644 --- a/drivers/net/ibm_newemac/core.c +++ b/drivers/net/ibm_newemac/core.c @@ -1636,6 +1636,12 @@ static int emac_poll_rx(void *param, int budget) goto next; } + if (len < ETH_HLEN) { + ++dev->estats.rx_dropped_stack; + emac_recycle_rx_skb(dev, slot, len); + goto next; + } + if (len && len < EMAC_RX_COPY_THRESH) { struct sk_buff *copy_skb = alloc_skb(len + EMAC_RX_SKB_HEADROOM + 2, GFP_ATOMIC); From c5a78ac00c400df29645e59938700301efb371d0 Mon Sep 17 00:00:00 2001 From: Laurent Pinchart Date: Thu, 26 Jun 2008 11:48:22 +0200 Subject: [PATCH 16/46] fs_enet: restore promiscuous and multicast settings in restart() The restart() function is called when the link state changes and resets multicast and promiscuous settings. This patch restores those settings at the end of restart(). Signed-off-by: Laurent Pinchart Signed-off-by: Jeff Garzik --- drivers/net/fs_enet/mac-fcc.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/net/fs_enet/mac-fcc.c b/drivers/net/fs_enet/mac-fcc.c index e36321152d50..8268b3535b30 100644 --- a/drivers/net/fs_enet/mac-fcc.c +++ b/drivers/net/fs_enet/mac-fcc.c @@ -463,6 +463,9 @@ static void restart(struct net_device *dev) else C32(fccp, fcc_fpsmr, FCC_PSMR_FDE | FCC_PSMR_LPB); + /* Restore multicast and promiscuous settings */ + set_multicast_list(dev); + S32(fccp, fcc_gfmr, FCC_GFMR_ENR | FCC_GFMR_ENT); } From 7f2d38eb7a42bea1c1df51bbdaa2ca0f0bdda07f Mon Sep 17 00:00:00 2001 From: Oliver Hartkopp Date: Sat, 5 Jul 2008 23:38:43 -0700 Subject: [PATCH 17/46] can: add sanity checks Even though the CAN netlayer only deals with CAN netdevices, the netlayer interface to the userspace and to the device layer should perform some sanity checks. This patch adds several sanity checks that mainly prevent userspace apps to send broken content into the system that may be misinterpreted by some other userspace application. Signed-off-by: Oliver Hartkopp Signed-off-by: Urs Thuermann Acked-by: Andre Naujoks Signed-off-by: David S. Miller --- net/can/af_can.c | 10 ++++++++++ net/can/bcm.c | 23 +++++++++++++++++++---- net/can/raw.c | 3 +++ 3 files changed, 32 insertions(+), 4 deletions(-) diff --git a/net/can/af_can.c b/net/can/af_can.c index 7e8ca2836452..484bbf6dd032 100644 --- a/net/can/af_can.c +++ b/net/can/af_can.c @@ -205,12 +205,19 @@ static int can_create(struct net *net, struct socket *sock, int protocol) * -ENOBUFS on full driver queue (see net_xmit_errno()) * -ENOMEM when local loopback failed at calling skb_clone() * -EPERM when trying to send on a non-CAN interface + * -EINVAL when the skb->data does not contain a valid CAN frame */ int can_send(struct sk_buff *skb, int loop) { struct sk_buff *newskb = NULL; + struct can_frame *cf = (struct can_frame *)skb->data; int err; + if (skb->len != sizeof(struct can_frame) || cf->can_dlc > 8) { + kfree_skb(skb); + return -EINVAL; + } + if (skb->dev->type != ARPHRD_CAN) { kfree_skb(skb); return -EPERM; @@ -605,6 +612,7 @@ static int can_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, struct net_device *orig_dev) { struct dev_rcv_lists *d; + struct can_frame *cf = (struct can_frame *)skb->data; int matches; if (dev->type != ARPHRD_CAN || dev_net(dev) != &init_net) { @@ -612,6 +620,8 @@ static int can_rcv(struct sk_buff *skb, struct net_device *dev, return 0; } + BUG_ON(skb->len != sizeof(struct can_frame) || cf->can_dlc > 8); + /* update statistics */ can_stats.rx_frames++; can_stats.rx_frames_delta++; diff --git a/net/can/bcm.c b/net/can/bcm.c index d9a3a9d13bed..72c2ce904f83 100644 --- a/net/can/bcm.c +++ b/net/can/bcm.c @@ -298,7 +298,7 @@ static void bcm_send_to_user(struct bcm_op *op, struct bcm_msg_head *head, if (head->nframes) { /* can_frames starting here */ - firstframe = (struct can_frame *) skb_tail_pointer(skb); + firstframe = (struct can_frame *)skb_tail_pointer(skb); memcpy(skb_put(skb, datalen), frames, datalen); @@ -826,6 +826,10 @@ static int bcm_tx_setup(struct bcm_msg_head *msg_head, struct msghdr *msg, for (i = 0; i < msg_head->nframes; i++) { err = memcpy_fromiovec((u8 *)&op->frames[i], msg->msg_iov, CFSIZ); + + if (op->frames[i].can_dlc > 8) + err = -EINVAL; + if (err < 0) return err; @@ -858,6 +862,10 @@ static int bcm_tx_setup(struct bcm_msg_head *msg_head, struct msghdr *msg, for (i = 0; i < msg_head->nframes; i++) { err = memcpy_fromiovec((u8 *)&op->frames[i], msg->msg_iov, CFSIZ); + + if (op->frames[i].can_dlc > 8) + err = -EINVAL; + if (err < 0) { if (op->frames != &op->sframe) kfree(op->frames); @@ -1164,9 +1172,12 @@ static int bcm_tx_send(struct msghdr *msg, int ifindex, struct sock *sk) skb->dev = dev; skb->sk = sk; - can_send(skb, 1); /* send with loopback */ + err = can_send(skb, 1); /* send with loopback */ dev_put(dev); + if (err) + return err; + return CFSIZ + MHSIZ; } @@ -1185,6 +1196,10 @@ static int bcm_sendmsg(struct kiocb *iocb, struct socket *sock, if (!bo->bound) return -ENOTCONN; + /* check for valid message length from userspace */ + if (size < MHSIZ || (size - MHSIZ) % CFSIZ) + return -EINVAL; + /* check for alternative ifindex for this bcm_op */ if (!ifindex && msg->msg_name) { @@ -1259,8 +1274,8 @@ static int bcm_sendmsg(struct kiocb *iocb, struct socket *sock, break; case TX_SEND: - /* we need at least one can_frame */ - if (msg_head.nframes < 1) + /* we need exactly one can_frame behind the msg head */ + if ((msg_head.nframes != 1) || (size != CFSIZ + MHSIZ)) ret = -EINVAL; else ret = bcm_tx_send(msg, ifindex, sk); diff --git a/net/can/raw.c b/net/can/raw.c index 69877b8e7e9c..3e46ee36a1aa 100644 --- a/net/can/raw.c +++ b/net/can/raw.c @@ -632,6 +632,9 @@ static int raw_sendmsg(struct kiocb *iocb, struct socket *sock, } else ifindex = ro->ifindex; + if (size != sizeof(struct can_frame)) + return -EINVAL; + dev = dev_get_by_index(&init_net, ifindex); if (!dev) return -ENXIO; From 0f9bfa569d46f2346a53a940b2b9e49a38635732 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 6 Jul 2008 16:06:25 -0700 Subject: [PATCH 18/46] vsprintf: split out '%s' handling logic The actual code is the same, just split out into a helper function. This makes it easier to read, and allows for future sharing of the string code. Signed-off-by: Linus Torvalds --- lib/vsprintf.c | 57 +++++++++++++++++++++++++++----------------------- 1 file changed, 31 insertions(+), 26 deletions(-) diff --git a/lib/vsprintf.c b/lib/vsprintf.c index 6021757a4496..926c7e00e2dc 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -482,6 +482,35 @@ static char *number(char *buf, char *end, unsigned long long num, int base, int return buf; } +static char *string(char *buf, char *end, char *s, int field_width, int precision, int flags) +{ + int len, i; + + if ((unsigned long)s < PAGE_SIZE) + s = ""; + + len = strnlen(s, precision); + + if (!(flags & LEFT)) { + while (len < field_width--) { + if (buf < end) + *buf = ' '; + ++buf; + } + } + for (i = 0; i < len; ++i) { + if (buf < end) + *buf = *s; + ++buf; ++s; + } + while (len < field_width--) { + if (buf < end) + *buf = ' '; + ++buf; + } + return buf; +} + /** * vsnprintf - Format a string and place it in a buffer * @buf: The buffer to place the result into @@ -502,11 +531,9 @@ static char *number(char *buf, char *end, unsigned long long num, int base, int */ int vsnprintf(char *buf, size_t size, const char *fmt, va_list args) { - int len; unsigned long long num; - int i, base; + int base; char *str, *end, c; - const char *s; int flags; /* flags to number() */ @@ -622,29 +649,7 @@ int vsnprintf(char *buf, size_t size, const char *fmt, va_list args) continue; case 's': - s = va_arg(args, char *); - if ((unsigned long)s < PAGE_SIZE) - s = ""; - - len = strnlen(s, precision); - - if (!(flags & LEFT)) { - while (len < field_width--) { - if (str < end) - *str = ' '; - ++str; - } - } - for (i = 0; i < len; ++i) { - if (str < end) - *str = *s; - ++str; ++s; - } - while (len < field_width--) { - if (str < end) - *str = ' '; - ++str; - } + str = string(str, end, va_arg(args, char *), field_width, precision, flags); continue; case 'p': From 78a8bf69b32980879975f7e31d30386c50bfe851 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 6 Jul 2008 16:16:15 -0700 Subject: [PATCH 19/46] vsprintf: split out '%p' handling logic The actual code is the same, just split out into a helper function. This makes it easier to read, and allows for simple future extension of %p handling. Signed-off-by: Linus Torvalds --- lib/vsprintf.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) diff --git a/lib/vsprintf.c b/lib/vsprintf.c index 926c7e00e2dc..f569feb7662e 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -511,6 +511,16 @@ static char *string(char *buf, char *end, char *s, int field_width, int precisio return buf; } +static char *pointer(char *buf, char *end, void *ptr, int field_width, int precision, int flags) +{ + flags |= SMALL; + if (field_width == -1) { + field_width = 2*sizeof(void *); + flags |= ZEROPAD; + } + return number(buf, end, (unsigned long) ptr, 16, field_width, precision, flags); +} + /** * vsnprintf - Format a string and place it in a buffer * @buf: The buffer to place the result into @@ -653,17 +663,9 @@ int vsnprintf(char *buf, size_t size, const char *fmt, va_list args) continue; case 'p': - flags |= SMALL; - if (field_width == -1) { - field_width = 2*sizeof(void *); - flags |= ZEROPAD; - } - str = number(str, end, - (unsigned long) va_arg(args, void *), - 16, field_width, precision, flags); + str = pointer(str, end, va_arg(args, void *), field_width, precision, flags); continue; - case 'n': /* FIXME: * What does C99 say about the overflow case here? */ From 4d8a743cdd2690c0bc8d1b8cbd02cffb1ead849f Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 6 Jul 2008 16:24:57 -0700 Subject: [PATCH 20/46] vsprintf: add infrastructure support for extended '%p' specifiers This expands the kernel '%p' handling with an arbitrary alphanumberic specifier extension string immediately following the '%p'. Right now it's just being ignored, but the next commit will start adding some specific pointer type extensions. NOTE! The reason the extension is appended to the '%p' is to allow minimal gcc type checking: gcc will still see the '%p' and will check that the argument passed in is indeed a pointer, and yet will not complain about the extended information that gcc doesn't understand about (on the other hand, it also won't actually check that the pointer type and the extension are compatible). Alphanumeric characters were chosen because there is no sane existing use for a string format with a hex pointer representation immediately followed by alphanumerics (which is what such a format string would have traditionally resulted in). Signed-off-by: Linus Torvalds --- lib/vsprintf.c | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) diff --git a/lib/vsprintf.c b/lib/vsprintf.c index f569feb7662e..5d6f0718b6d9 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -511,7 +511,14 @@ static char *string(char *buf, char *end, char *s, int field_width, int precisio return buf; } -static char *pointer(char *buf, char *end, void *ptr, int field_width, int precision, int flags) +/* + * Show a '%p' thing. A kernel extension is that the '%p' is followed + * by an extra set of alphanumeric characters that are extended format + * specifiers. + * + * Right now don't actually handle any such, but we will.. + */ +static char *pointer(const char *fmt, char *buf, char *end, void *ptr, int field_width, int precision, int flags) { flags |= SMALL; if (field_width == -1) { @@ -663,7 +670,12 @@ int vsnprintf(char *buf, size_t size, const char *fmt, va_list args) continue; case 'p': - str = pointer(str, end, va_arg(args, void *), field_width, precision, flags); + str = pointer(fmt+1, str, end, + va_arg(args, void *), + field_width, precision, flags); + /* Skip all alphanumeric pointer suffixes */ + while (isalnum(fmt[1])) + fmt++; continue; case 'n': From 0fe1ef24f7bd0020f29ffe287dfdb9ead33ca0b2 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 6 Jul 2008 16:43:12 -0700 Subject: [PATCH 21/46] vsprintf: add support for '%pS' and '%pF' pointer formats They print out a pointer in symbolic format, if possible (ie using symbolic KALLSYMS information). The '%pS' format is for regular direct pointers (which can point to data or code and that you find on the stack during backtraces etc), while '%pF' is for C function pointer types. On most architectures, the two mean exactly the same thing, but some architectures use an indirect pointer for C function pointers, where the function pointer points to a function descriptor (which in turn contains the actual pointer to the code). The '%pF' code automatically does the appropriate function descriptor dereference on such architectures. Signed-off-by: Linus Torvalds --- lib/vsprintf.c | 41 ++++++++++++++++++++++++++++++++++++++++- 1 file changed, 40 insertions(+), 1 deletion(-) diff --git a/lib/vsprintf.c b/lib/vsprintf.c index 5d6f0718b6d9..1dc2d1d18fa8 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -22,6 +22,8 @@ #include #include #include +#include +#include #include /* for PAGE_SIZE */ #include @@ -511,15 +513,52 @@ static char *string(char *buf, char *end, char *s, int field_width, int precisio return buf; } +static inline void *dereference_function_descriptor(void *ptr) +{ +#if defined(CONFIG_IA64) || defined(CONFIG_PPC64) + void *p; + if (!probe_kernel_address(ptr, p)) + ptr = p; +#endif + return ptr; +} + +static char *symbol_string(char *buf, char *end, void *ptr, int field_width, int precision, int flags) +{ + unsigned long value = (unsigned long) ptr; +#ifdef CONFIG_KALLSYMS + char sym[KSYM_SYMBOL_LEN]; + sprint_symbol(sym, value); + return string(buf, end, sym, field_width, precision, flags); +#else + field_width = 2*sizeof(void *); + flags |= SPECIAL | SMALL | ZEROPAD; + return number(buf, end, value, 16, field_width, precision, flags); +#endif +} + /* * Show a '%p' thing. A kernel extension is that the '%p' is followed * by an extra set of alphanumeric characters that are extended format * specifiers. * - * Right now don't actually handle any such, but we will.. + * Right now we just handle 'F' (for symbolic Function descriptor pointers) + * and 'S' (for Symbolic direct pointers), but this can easily be + * extended in the future (network address types etc). + * + * The difference between 'S' and 'F' is that on ia64 and ppc64 function + * pointers are really function descriptors, which contain a pointer the + * real address. */ static char *pointer(const char *fmt, char *buf, char *end, void *ptr, int field_width, int precision, int flags) { + switch (*fmt) { + case 'F': + ptr = dereference_function_descriptor(ptr); + /* Fallthrough */ + case 'S': + return symbol_string(buf, end, ptr, field_width, precision, flags); + } flags |= SMALL; if (field_width == -1) { field_width = 2*sizeof(void *); From 3bc5ab9b7f2760d2892fd0a0589e1077e869d4f5 Mon Sep 17 00:00:00 2001 From: Benjamin Herrenschmidt Date: Mon, 7 Jul 2008 16:39:50 +1000 Subject: [PATCH 22/46] powerpc: Fix unterminated of_device_id array in legacy_serial.c A recent patch to legacy_serial.c factored out some code by using the of_match_node() facility to match a node against an array of possible matches. However, the patch didn't properly terminate the array causing potential crashes in cases where no match is found. In addition, the name of the array was poorly chosen for a static symbol making debugging harder. Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Linus Torvalds --- arch/powerpc/kernel/legacy_serial.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/arch/powerpc/kernel/legacy_serial.c b/arch/powerpc/kernel/legacy_serial.c index cf37f5ca4b71..4d96e1db55ee 100644 --- a/arch/powerpc/kernel/legacy_serial.c +++ b/arch/powerpc/kernel/legacy_serial.c @@ -33,13 +33,14 @@ static struct legacy_serial_info { phys_addr_t taddr; } legacy_serial_infos[MAX_LEGACY_SERIAL_PORTS]; -static struct __initdata of_device_id parents[] = { +static struct __initdata of_device_id legacy_serial_parents[] = { {.type = "soc",}, {.type = "tsi-bridge",}, {.type = "opb", }, {.compatible = "ibm,opb",}, {.compatible = "simple-bus",}, {.compatible = "wrs,epld-localbus",}, + {}, }; static unsigned int legacy_serial_count; @@ -327,7 +328,7 @@ void __init find_legacy_serial_ports(void) struct device_node *parent = of_get_parent(np); if (!parent) continue; - if (of_match_node(parents, parent) != NULL) { + if (of_match_node(legacy_serial_parents, parent) != NULL) { index = add_legacy_soc_port(np, np); if (index >= 0 && np == stdout) legacy_serial_console = index; From 4f81c5350b44bcc501ab6f8a089b16d064b4d2f6 Mon Sep 17 00:00:00 2001 From: Jeff Dike Date: Mon, 7 Jul 2008 13:36:56 -0400 Subject: [PATCH 23/46] [UML] fix gcc ICEs and unresolved externs There are various constraints on the use of unit-at-a-time: - i386 uses no-unit-at-a-time for pre-4.0 (not 4.3) - x86_64 uses unit-at-a-time always Uli reported a crash on x86_64 with gcc 4.1.2 with unit-at-a-time, resulting in commit c0a18111e571138747a98af18b3a2124df56a0d1 Ingo reported a gcc internal error with gcc 4.3 with no-unit-at-a-timem, resulting in 22eecde2f9034764a3fd095eecfa3adfb8ec9a98 Benny Halevy is seeing extern inlines not resolved with gcc 4.3 with no-unit-at-a-time This patch reintroduces unit-at-a-time for gcc >= 4.0, bringing back the possibility of Uli's crash. If that happens, we'll debug it. I started seeing both the internal compiler errors and unresolved inlines on Fedora 9. This patch fixes both problems, without so far reintroducing the crash reported by Uli. Signed-off-by: Jeff Dike Cc: Benny Halevy Cc: Adrian Bunk Cc: Ingo Molnar Cc: Ulrich Drepper Signed-off-by: Linus Torvalds --- arch/um/Makefile | 1 - arch/um/Makefile-i386 | 7 +++++++ arch/um/Makefile-x86_64 | 3 +++ 3 files changed, 10 insertions(+), 1 deletion(-) diff --git a/arch/um/Makefile b/arch/um/Makefile index dbeab15e7bb7..ca40397017b9 100644 --- a/arch/um/Makefile +++ b/arch/um/Makefile @@ -77,7 +77,6 @@ include $(srctree)/$(ARCH_DIR)/Makefile-os-$(OS) KERNEL_DEFINES = $(strip -Derrno=kernel_errno -Dsigprocmask=kernel_sigprocmask \ -Dmktime=kernel_mktime $(ARCH_KERNEL_DEFINES)) KBUILD_CFLAGS += $(KERNEL_DEFINES) -KBUILD_CFLAGS += $(call cc-option,-fno-unit-at-a-time,) PHONY += linux diff --git a/arch/um/Makefile-i386 b/arch/um/Makefile-i386 index 561e373bd850..302cbe504543 100644 --- a/arch/um/Makefile-i386 +++ b/arch/um/Makefile-i386 @@ -32,4 +32,11 @@ cflags-y += $(call cc-option,-mpreferred-stack-boundary=2) # an unresolved reference. cflags-y += -ffreestanding +# Disable unit-at-a-time mode on pre-gcc-4.0 compilers, it makes gcc use +# a lot more stack due to the lack of sharing of stacklots. Also, gcc +# 4.3.0 needs -funit-at-a-time for extern inline functions. +KBUILD_CFLAGS += $(shell if [ $(call cc-version) -lt 0400 ] ; then \ + echo $(call cc-option,-fno-unit-at-a-time); \ + else echo $(call cc-option,-funit-at-a-time); fi ;) + KBUILD_CFLAGS += $(cflags-y) diff --git a/arch/um/Makefile-x86_64 b/arch/um/Makefile-x86_64 index 8ed362f93582..a9cd7e77a7ab 100644 --- a/arch/um/Makefile-x86_64 +++ b/arch/um/Makefile-x86_64 @@ -21,3 +21,6 @@ HEADER_ARCH := x86 LINK-$(CONFIG_LD_SCRIPT_DYN) += -Wl,-rpath,/lib64 LINK-y += -m64 + +# Do unit-at-a-time unconditionally on x86_64, following the host +KBUILD_CFLAGS += $(call cc-option,-funit-at-a-time) From 739db07f82767e7634176d18af2acbe77b11fd42 Mon Sep 17 00:00:00 2001 From: Jesse Barnes Date: Mon, 7 Jul 2008 09:55:26 -0700 Subject: [PATCH 24/46] Revert "PCI: Correct last two HP entries in the bfsort whitelist" This reverts commit a1676072558854b95336c8f7db76b0504e909a0a. It duplicates the change from 8d64c781f0c5fbfdf8016bd1634506ff2ad1376a and only one should be applied, otherwise some of the Dell quirks are lost. Thanks to Tony Camuso for catching this. Acked-by: Tony Camuso Signed-off-by: Jesse Barnes --- arch/x86/pci/common.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index 940185ecaeda..6e64aaf00d1d 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -328,18 +328,18 @@ static struct dmi_system_id __devinitdata pciprobe_dmi_table[] = { #endif { .callback = set_bf_sort, - .ident = "HP ProLiant DL360", + .ident = "HP ProLiant DL385 G2", .matches = { DMI_MATCH(DMI_SYS_VENDOR, "HP"), - DMI_MATCH(DMI_PRODUCT_NAME, "ProLiant DL360"), + DMI_MATCH(DMI_PRODUCT_NAME, "ProLiant DL385 G2"), }, }, { .callback = set_bf_sort, - .ident = "HP ProLiant DL380", + .ident = "HP ProLiant DL585 G2", .matches = { DMI_MATCH(DMI_SYS_VENDOR, "HP"), - DMI_MATCH(DMI_PRODUCT_NAME, "ProLiant DL380"), + DMI_MATCH(DMI_PRODUCT_NAME, "ProLiant DL585 G2"), }, }, {} From 138c9021ca68d6f8a6fac3418995fee102a3cc22 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Tue, 8 Jul 2008 11:31:06 +0200 Subject: [PATCH 25/46] [S390] protect _PAGE_SPECIAL bit against mprotect Stop mprotect's pte_modify from wiping out the s390 pte_special bit, which caused oops thereafter when vm_normal_page thought X's abnormal was normal. Debugged-by: Ryan Hope Debugged-by: Zan Lynx Acked-by: Hugh Dickins Signed-off-by: Nick Piggin Signed-off-by: Martin Schwidefsky --- include/asm-s390/pgtable.h | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/include/asm-s390/pgtable.h b/include/asm-s390/pgtable.h index c7f4f8e3e297..bd0ea191dfa9 100644 --- a/include/asm-s390/pgtable.h +++ b/include/asm-s390/pgtable.h @@ -223,6 +223,9 @@ extern char empty_zero_page[PAGE_SIZE]; #define _PAGE_SPECIAL 0x004 /* SW associated with special page */ #define __HAVE_ARCH_PTE_SPECIAL +/* Set of bits not changed in pte_modify */ +#define _PAGE_CHG_MASK (PAGE_MASK | _PAGE_SPECIAL) + /* Six different types of pages. */ #define _PAGE_TYPE_EMPTY 0x400 #define _PAGE_TYPE_NONE 0x401 @@ -681,7 +684,7 @@ static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *pt */ static inline pte_t pte_modify(pte_t pte, pgprot_t newprot) { - pte_val(pte) &= PAGE_MASK; + pte_val(pte) &= _PAGE_CHG_MASK; pte_val(pte) |= pgprot_val(newprot); return pte; } From 86df86424939d316b1f6cfac1b6204f0c7dee317 Mon Sep 17 00:00:00 2001 From: David Gibson Date: Tue, 8 Jul 2008 15:58:16 +1000 Subject: [PATCH 26/46] Correct hash flushing from huge_ptep_set_wrprotect() As Andy Whitcroft recently pointed out, the current powerpc version of huge_ptep_set_wrprotect() has a bug. It just calls ptep_set_wrprotect() which in turn calls pte_update() then hpte_need_flush() with the 'huge' argument set to 0. This will cause hpte_need_flush() to flush the wrong hash entries (of any). Andy's fix for this is already in the powerpc tree as commit 016b33c4958681c24056abed8ec95844a0da80a3. I have confirmed this is a real bug, not masked by some other synchronization, with a new testcase for libhugetlbfs. A process write a (MAP_PRIVATE) hugepage mapping, fork(), then alter the mapping and have the child incorrectly see the second write. Therefore, this should be fixed for 2.6.26, and for the stable tree. Here is a suitable patch for 2.6.26, which I think will also be suitable for the stable tree (neither of the headers in question has been changed much recently). It is cut down slighlty from Andy's original version, in that it does not include a 32-bit version of huge_ptep_set_wrprotect(). Currently, hugepages are not supported on any 32-bit powerpc platform. When they are, a suitable 32-bit version can be added - the only 32-bit hardware which supports hugepages does not use the conventional hashtable MMU and so will have different needs anyway. Signed-off-by: Andy Whitcroft Signed-off-by: David Gibson Signed-off-by: Benjamin Herrenschmidt Signed-off-by: Linus Torvalds --- include/asm-powerpc/hugetlb.h | 6 ------ include/asm-powerpc/pgtable-ppc64.h | 10 ++++++++++ 2 files changed, 10 insertions(+), 6 deletions(-) diff --git a/include/asm-powerpc/hugetlb.h b/include/asm-powerpc/hugetlb.h index 649c6c3b87b3..be32ff02f4a0 100644 --- a/include/asm-powerpc/hugetlb.h +++ b/include/asm-powerpc/hugetlb.h @@ -49,12 +49,6 @@ static inline pte_t huge_pte_wrprotect(pte_t pte) return pte_wrprotect(pte); } -static inline void huge_ptep_set_wrprotect(struct mm_struct *mm, - unsigned long addr, pte_t *ptep) -{ - ptep_set_wrprotect(mm, addr, ptep); -} - static inline int huge_ptep_set_access_flags(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep, pte_t pte, int dirty) diff --git a/include/asm-powerpc/pgtable-ppc64.h b/include/asm-powerpc/pgtable-ppc64.h index cc6a43ba41d0..7686569a0bef 100644 --- a/include/asm-powerpc/pgtable-ppc64.h +++ b/include/asm-powerpc/pgtable-ppc64.h @@ -314,6 +314,16 @@ static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, old = pte_update(mm, addr, ptep, _PAGE_RW, 0); } +static inline void huge_ptep_set_wrprotect(struct mm_struct *mm, + unsigned long addr, pte_t *ptep) +{ + unsigned long old; + + if ((pte_val(*ptep) & _PAGE_RW) == 0) + return; + old = pte_update(mm, addr, ptep, _PAGE_RW, 1); +} + /* * We currently remove entries from the hashtable regardless of whether * the entry was young or dirty. The generic routines only flush if the From a861beb1401d65e3f095fee074c13645ab06490e Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Tue, 8 Jul 2008 19:27:22 +0200 Subject: [PATCH 27/46] ide: add __ide_default_irq() inline helper Add __ide_default_irq() inline helper and use it instead of ide_default_irq() in ide-probe.c and ns87415.c (all host drivers except IDE PCI ones always setup hwif->irq so it is enough to check only for I/O bases 0x1f0 and 0x170). This fixes post-2.6.25 regression since ide_default_irq() define could shadow ide_default_irq() inline. Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/ide-probe.c | 8 ++------ drivers/ide/pci/ns87415.c | 6 +----- include/linux/ide.h | 15 +++++++++++++++ 3 files changed, 18 insertions(+), 11 deletions(-) diff --git a/drivers/ide/ide-probe.c b/drivers/ide/ide-probe.c index d27061b39324..26e68b65b7cf 100644 --- a/drivers/ide/ide-probe.c +++ b/drivers/ide/ide-probe.c @@ -1218,16 +1218,12 @@ static void drive_release_dev (struct device *dev) complete(&drive->gendev_rel_comp); } -#ifndef ide_default_irq -#define ide_default_irq(irq) 0 -#endif - static int hwif_init(ide_hwif_t *hwif) { int old_irq; if (!hwif->irq) { - hwif->irq = ide_default_irq(hwif->io_ports.data_addr); + hwif->irq = __ide_default_irq(hwif->io_ports.data_addr); if (!hwif->irq) { printk("%s: DISABLED, NO IRQ\n", hwif->name); return 0; @@ -1257,7 +1253,7 @@ static int hwif_init(ide_hwif_t *hwif) * It failed to initialise. Find the default IRQ for * this port and try that. */ - hwif->irq = ide_default_irq(hwif->io_ports.data_addr); + hwif->irq = __ide_default_irq(hwif->io_ports.data_addr); if (!hwif->irq) { printk("%s: Disabled unable to get IRQ %d.\n", hwif->name, old_irq); diff --git a/drivers/ide/pci/ns87415.c b/drivers/ide/pci/ns87415.c index fec4955f449b..a7a41bb82778 100644 --- a/drivers/ide/pci/ns87415.c +++ b/drivers/ide/pci/ns87415.c @@ -225,10 +225,6 @@ static int ns87415_dma_setup(ide_drive_t *drive) return 1; } -#ifndef ide_default_irq -#define ide_default_irq(irq) 0 -#endif - static void __devinit init_hwif_ns87415 (ide_hwif_t *hwif) { struct pci_dev *dev = to_pci_dev(hwif->dev); @@ -288,7 +284,7 @@ static void __devinit init_hwif_ns87415 (ide_hwif_t *hwif) } if (!using_inta) - hwif->irq = ide_default_irq(hwif->io_ports.data_addr); + hwif->irq = __ide_default_irq(hwif->io_ports.data_addr); else if (!hwif->irq && hwif->mate && hwif->mate->irq) hwif->irq = hwif->mate->irq; /* share IRQ with mate */ diff --git a/include/linux/ide.h b/include/linux/ide.h index 9918772bf274..eddb6daadf4a 100644 --- a/include/linux/ide.h +++ b/include/linux/ide.h @@ -189,6 +189,21 @@ static inline void ide_std_init_ports(hw_regs_t *hw, hw->io_ports.ctl_addr = ctl_addr; } +/* for IDE PCI controllers in legacy mode, temporary */ +static inline int __ide_default_irq(unsigned long base) +{ + switch (base) { +#ifdef CONFIG_IA64 + case 0x1f0: return isa_irq_to_vector(14); + case 0x170: return isa_irq_to_vector(15); +#else + case 0x1f0: return 14; + case 0x170: return 15; +#endif + } + return 0; +} + #include #if !defined(MAX_HWIFS) || defined(CONFIG_EMBEDDED) From ffab6cf44e9058fe75a33aa86386b22e616a8f6f Mon Sep 17 00:00:00 2001 From: Sergei Shtylyov Date: Tue, 8 Jul 2008 19:27:22 +0200 Subject: [PATCH 28/46] palm_bk3710: fix IDECLK period calculation The driver uses completely bogus rounding formula for calculating period from the IDECLK frequency which gives one-off period values (e.g. 11 ns with 100 MHz IDECLK) which in turn can lead to overclocked IDE transfer timings. Actually, rounding is just wrong in this case, so use a mere division for a safe result. While at it, also: - give 'ide_palm_clk' variable a more suitable name; - get rid of the useless 'ideclkp' variable; - drop the LISP stype 'p' postfix from the 'clkp' variable's name. :-) Signed-off-by: Sergei Shtylyov Cc: mcherkashin@ru.mvista.com Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/arm/palm_bk3710.c | 38 +++++++++++++++++------------------ 1 file changed, 18 insertions(+), 20 deletions(-) diff --git a/drivers/ide/arm/palm_bk3710.c b/drivers/ide/arm/palm_bk3710.c index cc24803fadff..2f2b4f4cf229 100644 --- a/drivers/ide/arm/palm_bk3710.c +++ b/drivers/ide/arm/palm_bk3710.c @@ -76,7 +76,7 @@ struct palm_bk3710_udmatiming { #include "../ide-timing.h" -static long ide_palm_clk; +static unsigned ideclk_period; /* in nanoseconds */ static const struct palm_bk3710_udmatiming palm_bk3710_udmatimings[6] = { {160, 240}, /* UDMA Mode 0 */ @@ -86,8 +86,6 @@ static const struct palm_bk3710_udmatiming palm_bk3710_udmatimings[6] = { {85, 60}, /* UDMA Mode 4 */ }; -static struct clk *ideclkp; - static void palm_bk3710_setudmamode(void __iomem *base, unsigned int dev, unsigned int mode) { @@ -97,10 +95,10 @@ static void palm_bk3710_setudmamode(void __iomem *base, unsigned int dev, /* DMA Data Setup */ t0 = DIV_ROUND_UP(palm_bk3710_udmatimings[mode].cycletime, - ide_palm_clk) - 1; - tenv = DIV_ROUND_UP(20, ide_palm_clk) - 1; + ideclk_period) - 1; + tenv = DIV_ROUND_UP(20, ideclk_period) - 1; trp = DIV_ROUND_UP(palm_bk3710_udmatimings[mode].rptime, - ide_palm_clk) - 1; + ideclk_period) - 1; /* udmatim Register */ val16 = readw(base + BK3710_UDMATIM) & (dev ? 0xFF0F : 0xFFF0); @@ -141,8 +139,8 @@ static void palm_bk3710_setdmamode(void __iomem *base, unsigned int dev, cycletime = max_t(int, t->cycle, min_cycle); /* DMA Data Setup */ - t0 = DIV_ROUND_UP(cycletime, ide_palm_clk); - td = DIV_ROUND_UP(t->active, ide_palm_clk); + t0 = DIV_ROUND_UP(cycletime, ideclk_period); + td = DIV_ROUND_UP(t->active, ideclk_period); tkw = t0 - td - 1; td -= 1; @@ -168,9 +166,9 @@ static void palm_bk3710_setpiomode(void __iomem *base, ide_drive_t *mate, struct ide_timing *t; /* PIO Data Setup */ - t0 = DIV_ROUND_UP(cycletime, ide_palm_clk); + t0 = DIV_ROUND_UP(cycletime, ideclk_period); t2 = DIV_ROUND_UP(ide_timing_find_mode(XFER_PIO_0 + mode)->active, - ide_palm_clk); + ideclk_period); t2i = t0 - t2 - 1; t2 -= 1; @@ -192,8 +190,8 @@ static void palm_bk3710_setpiomode(void __iomem *base, ide_drive_t *mate, /* TASKFILE Setup */ t = ide_timing_find_mode(XFER_PIO_0 + mode); - t0 = DIV_ROUND_UP(t->cyc8b, ide_palm_clk); - t2 = DIV_ROUND_UP(t->act8b, ide_palm_clk); + t0 = DIV_ROUND_UP(t->cyc8b, ideclk_period); + t2 = DIV_ROUND_UP(t->act8b, ideclk_period); t2i = t0 - t2 - 1; t2 -= 1; @@ -350,22 +348,22 @@ static const struct ide_port_info __devinitdata palm_bk3710_port_info = { static int __devinit palm_bk3710_probe(struct platform_device *pdev) { - struct clk *clkp; + struct clk *clk; struct resource *mem, *irq; ide_hwif_t *hwif; - unsigned long base; + unsigned long base, rate; int i; hw_regs_t hw; u8 idx[4] = { 0xff, 0xff, 0xff, 0xff }; - clkp = clk_get(NULL, "IDECLK"); - if (IS_ERR(clkp)) + clk = clk_get(NULL, "IDECLK"); + if (IS_ERR(clk)) return -ENODEV; - ideclkp = clkp; - clk_enable(ideclkp); - ide_palm_clk = clk_get_rate(ideclkp)/100000; - ide_palm_clk = (10000/ide_palm_clk) + 1; + clk_enable(clk); + rate = clk_get_rate(clk); + ideclk_period = 1000000000UL / rate; + /* Register the IDE interface with Linux ATA Interface */ memset(&hw, 0, sizeof(hw)); From be305042b7a01a1ab03a8adfa95f57bc63e012e1 Mon Sep 17 00:00:00 2001 From: Bartlomiej Zolnierkiewicz Date: Tue, 8 Jul 2008 19:27:23 +0200 Subject: [PATCH 29/46] it8213: fix return value in it8213_init_one() Signed-off-by: Bartlomiej Zolnierkiewicz --- drivers/ide/pci/it8213.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/ide/pci/it8213.c b/drivers/ide/pci/it8213.c index 9053c8771e6e..2b71bdf74e73 100644 --- a/drivers/ide/pci/it8213.c +++ b/drivers/ide/pci/it8213.c @@ -184,8 +184,7 @@ static const struct ide_port_info it8213_chipsets[] __devinitdata = { static int __devinit it8213_init_one(struct pci_dev *dev, const struct pci_device_id *id) { - ide_setup_pci_device(dev, &it8213_chipsets[id->driver_data]); - return 0; + return ide_setup_pci_device(dev, &it8213_chipsets[id->driver_data]); } static const struct pci_device_id it8213_pci_tbl[] = { From b32dfbb9c54393af32761add16e249664193621f Mon Sep 17 00:00:00 2001 From: Shane McDonald Date: Sat, 5 Jul 2008 17:19:42 -0600 Subject: [PATCH 30/46] [MIPS] Atlas, decstation: Fix section mismatches triggered by defconfigs Resolve these mismatches by defining affected functions with the __cpuinit attribute, rather than __init. Signed-off-by: Shane McDonald Signed-off-by: Ralf Baechle --- arch/mips/mm/c-r3k.c | 6 +++--- arch/mips/mm/sc-rm7k.c | 4 ++-- 2 files changed, 5 insertions(+), 5 deletions(-) diff --git a/arch/mips/mm/c-r3k.c b/arch/mips/mm/c-r3k.c index 76935e320214..27a5b466c85c 100644 --- a/arch/mips/mm/c-r3k.c +++ b/arch/mips/mm/c-r3k.c @@ -26,7 +26,7 @@ static unsigned long icache_size, dcache_size; /* Size in bytes */ static unsigned long icache_lsize, dcache_lsize; /* Size in bytes */ -unsigned long __init r3k_cache_size(unsigned long ca_flags) +unsigned long __cpuinit r3k_cache_size(unsigned long ca_flags) { unsigned long flags, status, dummy, size; volatile unsigned long *p; @@ -61,7 +61,7 @@ unsigned long __init r3k_cache_size(unsigned long ca_flags) return size * sizeof(*p); } -unsigned long __init r3k_cache_lsize(unsigned long ca_flags) +unsigned long __cpuinit r3k_cache_lsize(unsigned long ca_flags) { unsigned long flags, status, lsize, i; volatile unsigned long *p; @@ -90,7 +90,7 @@ unsigned long __init r3k_cache_lsize(unsigned long ca_flags) return lsize * sizeof(*p); } -static void __init r3k_probe_cache(void) +static void __cpuinit r3k_probe_cache(void) { dcache_size = r3k_cache_size(ST0_ISC); if (dcache_size) diff --git a/arch/mips/mm/sc-rm7k.c b/arch/mips/mm/sc-rm7k.c index fc227f3b1199..e3abfb2d7e86 100644 --- a/arch/mips/mm/sc-rm7k.c +++ b/arch/mips/mm/sc-rm7k.c @@ -86,7 +86,7 @@ static void rm7k_sc_inv(unsigned long addr, unsigned long size) /* * This function is executed in uncached address space. */ -static __init void __rm7k_sc_enable(void) +static __cpuinit void __rm7k_sc_enable(void) { int i; @@ -107,7 +107,7 @@ static __init void __rm7k_sc_enable(void) } } -static __init void rm7k_sc_enable(void) +static __cpuinit void rm7k_sc_enable(void) { if (read_c0_config() & RM7K_CONF_SE) return; From 14defd90f5281da8a1bf43bc789efbafe5991cd8 Mon Sep 17 00:00:00 2001 From: Thomas Bogendoerfer Date: Tue, 8 Jul 2008 14:46:34 +0200 Subject: [PATCH 31/46] [MIPS] Fix 32bit kernels on R4k with 128 byte cache line size The generated copy_page for R4k CPU with a 128 byte cache line size used Create Dirty Exclusive cache line operations even if only part of the cache line was filled. This change avoids generating cache operations, if only part of the cache line size is copied in one loop. It also increases the maxmimum loop size, because the generated code even fits into the available space for r4k CPUs with 128 byte cache line size. Signed-off-by: Thomas Bogendoerfer Signed-off-by: Ralf Baechle --- arch/mips/mm/page.c | 65 ++++++++++++++++++++++++--------------------- 1 file changed, 34 insertions(+), 31 deletions(-) diff --git a/arch/mips/mm/page.c b/arch/mips/mm/page.c index 1edf0cbbeede..1417c6494858 100644 --- a/arch/mips/mm/page.c +++ b/arch/mips/mm/page.c @@ -235,13 +235,12 @@ static void __cpuinit set_prefetch_parameters(void) } /* * Too much unrolling will overflow the available space in - * clear_space_array / copy_page_array. 8 words sounds generous, - * but a R4000 with 128 byte L2 line length can exceed even that. + * clear_space_array / copy_page_array. */ - half_clear_loop_size = min(8 * clear_word_size, + half_clear_loop_size = min(16 * clear_word_size, max(cache_line_size >> 1, 4 * clear_word_size)); - half_copy_loop_size = min(8 * copy_word_size, + half_copy_loop_size = min(16 * copy_word_size, max(cache_line_size >> 1, 4 * copy_word_size)); } @@ -263,21 +262,23 @@ static inline void __cpuinit build_clear_pref(u32 **buf, int off) if (pref_bias_clear_store) { uasm_i_pref(buf, pref_dst_mode, pref_bias_clear_store + off, A0); - } else if (cpu_has_cache_cdex_s) { - uasm_i_cache(buf, Create_Dirty_Excl_SD, off, A0); - } else if (cpu_has_cache_cdex_p) { - if (R4600_V1_HIT_CACHEOP_WAR && cpu_is_r4600_v1_x()) { - uasm_i_nop(buf); - uasm_i_nop(buf); - uasm_i_nop(buf); - uasm_i_nop(buf); + } else if (cache_line_size == (half_clear_loop_size << 1)) { + if (cpu_has_cache_cdex_s) { + uasm_i_cache(buf, Create_Dirty_Excl_SD, off, A0); + } else if (cpu_has_cache_cdex_p) { + if (R4600_V1_HIT_CACHEOP_WAR && cpu_is_r4600_v1_x()) { + uasm_i_nop(buf); + uasm_i_nop(buf); + uasm_i_nop(buf); + uasm_i_nop(buf); + } + + if (R4600_V2_HIT_CACHEOP_WAR && cpu_is_r4600_v2_x()) + uasm_i_lw(buf, ZERO, ZERO, AT); + + uasm_i_cache(buf, Create_Dirty_Excl_D, off, A0); + } } - - if (R4600_V2_HIT_CACHEOP_WAR && cpu_is_r4600_v2_x()) - uasm_i_lw(buf, ZERO, ZERO, AT); - - uasm_i_cache(buf, Create_Dirty_Excl_D, off, A0); - } } void __cpuinit build_clear_page(void) @@ -403,20 +404,22 @@ static inline void build_copy_store_pref(u32 **buf, int off) if (pref_bias_copy_store) { uasm_i_pref(buf, pref_dst_mode, pref_bias_copy_store + off, A0); - } else if (cpu_has_cache_cdex_s) { - uasm_i_cache(buf, Create_Dirty_Excl_SD, off, A0); - } else if (cpu_has_cache_cdex_p) { - if (R4600_V1_HIT_CACHEOP_WAR && cpu_is_r4600_v1_x()) { - uasm_i_nop(buf); - uasm_i_nop(buf); - uasm_i_nop(buf); - uasm_i_nop(buf); + } else if (cache_line_size == (half_copy_loop_size << 1)) { + if (cpu_has_cache_cdex_s) { + uasm_i_cache(buf, Create_Dirty_Excl_SD, off, A0); + } else if (cpu_has_cache_cdex_p) { + if (R4600_V1_HIT_CACHEOP_WAR && cpu_is_r4600_v1_x()) { + uasm_i_nop(buf); + uasm_i_nop(buf); + uasm_i_nop(buf); + uasm_i_nop(buf); + } + + if (R4600_V2_HIT_CACHEOP_WAR && cpu_is_r4600_v2_x()) + uasm_i_lw(buf, ZERO, ZERO, AT); + + uasm_i_cache(buf, Create_Dirty_Excl_D, off, A0); } - - if (R4600_V2_HIT_CACHEOP_WAR && cpu_is_r4600_v2_x()) - uasm_i_lw(buf, ZERO, ZERO, AT); - - uasm_i_cache(buf, Create_Dirty_Excl_D, off, A0); } } From 2aac05a91971fbd1bf6cbed78b8731eb7454b9b7 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 7 Jul 2008 13:26:10 -0400 Subject: [PATCH 32/46] NFS: Fix readdir cache invalidation invalidate_inode_pages2_range() takes page offset arguments, not byte ranges. Another thought is that individual pages might perhaps get evicted by VM pressure, in which case we might perhaps want to re-read not only the evicted page, but all subsequent pages too (in case the server returns more/less data per page so that the alignment of the next entry changes). We should therefore remove the condition that we only do this on page->index==0. Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 58d43daec084..982a2064fe4c 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -204,7 +204,7 @@ int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page) * Note: assumes we have exclusive access to this mapping either * through inode->i_mutex or some other mechanism. */ - if (page->index == 0 && invalidate_inode_pages2_range(inode->i_mapping, PAGE_CACHE_SIZE, -1) < 0) { + if (invalidate_inode_pages2_range(inode->i_mapping, page->index + 1, -1) < 0) { /* Should never happen */ nfs_zap_mapping(inode, inode->i_mapping); } From 0d3a34b48c87a374b37d7a21a60d257d076484f3 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 7 Jul 2008 12:18:52 -0400 Subject: [PATCH 33/46] SUNRPC: Fix a double-free in rpcbind It is wrong to be freeing up the rpcbind arguments if the call to rpcb_call_async() fails, since they should already have been freed up by rpcb_map_release(). Signed-off-by: Trond Myklebust --- net/sunrpc/rpcb_clnt.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index 0517967a68bf..21c698d7b774 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -365,18 +365,16 @@ void rpcb_getport_async(struct rpc_task *task) rpc_release_client(rpcb_clnt); if (IS_ERR(child)) { status = -EIO; + /* rpcb_map_release() has freed the arguments */ dprintk("RPC: %5u %s: rpc_run_task failed\n", task->tk_pid, __func__); - goto bailout; + goto bailout_nofree; } rpc_put_task(child); task->tk_xprt->stat.bind_count++; return; -bailout: - kfree(map); - xprt_put(xprt); bailout_nofree: rpcb_wake_rpcbind_waiters(xprt, status); bailout_nowake: From 803a9067e19714ea7b7da760fe92f0d53bfa6994 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 1 Jul 2008 15:20:55 -0400 Subject: [PATCH 34/46] SUNRPC: Fix an rpcbind breakage for the case of IPv6 lookups Now that rpcb_next_version has been split into an IPv4 version and an IPv6 version, we Oops when rpcb_call_async attempts to look up the IPv6-specific RPC procedure in rpcb_next_version. Fix the Oops simply by having rpcb_getport_async pass the correct RPC procedure as an argument. Signed-off-by: Trond Myklebust --- net/sunrpc/rpcb_clnt.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) diff --git a/net/sunrpc/rpcb_clnt.c b/net/sunrpc/rpcb_clnt.c index 21c698d7b774..e6fb21b19b86 100644 --- a/net/sunrpc/rpcb_clnt.c +++ b/net/sunrpc/rpcb_clnt.c @@ -243,10 +243,10 @@ int rpcb_getport_sync(struct sockaddr_in *sin, u32 prog, u32 vers, int prot) } EXPORT_SYMBOL_GPL(rpcb_getport_sync); -static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbind_args *map, int version) +static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbind_args *map, struct rpc_procinfo *proc) { struct rpc_message msg = { - .rpc_proc = rpcb_next_version[version].rpc_proc, + .rpc_proc = proc, .rpc_argp = map, .rpc_resp = &map->r_port, }; @@ -271,6 +271,7 @@ static struct rpc_task *rpcb_call_async(struct rpc_clnt *rpcb_clnt, struct rpcbi void rpcb_getport_async(struct rpc_task *task) { struct rpc_clnt *clnt = task->tk_client; + struct rpc_procinfo *proc; u32 bind_version; struct rpc_xprt *xprt = task->tk_xprt; struct rpc_clnt *rpcb_clnt; @@ -280,7 +281,6 @@ void rpcb_getport_async(struct rpc_task *task) struct sockaddr *sap = (struct sockaddr *)&addr; size_t salen; int status; - struct rpcb_info *info; dprintk("RPC: %5u %s(%s, %u, %u, %d)\n", task->tk_pid, __func__, @@ -313,10 +313,12 @@ void rpcb_getport_async(struct rpc_task *task) /* Don't ever use rpcbind v2 for AF_INET6 requests */ switch (sap->sa_family) { case AF_INET: - info = rpcb_next_version; + proc = rpcb_next_version[xprt->bind_index].rpc_proc; + bind_version = rpcb_next_version[xprt->bind_index].rpc_vers; break; case AF_INET6: - info = rpcb_next_version6; + proc = rpcb_next_version6[xprt->bind_index].rpc_proc; + bind_version = rpcb_next_version6[xprt->bind_index].rpc_vers; break; default: status = -EAFNOSUPPORT; @@ -324,14 +326,13 @@ void rpcb_getport_async(struct rpc_task *task) task->tk_pid, __func__); goto bailout_nofree; } - if (info[xprt->bind_index].rpc_proc == NULL) { + if (proc == NULL) { xprt->bind_index = 0; status = -EPFNOSUPPORT; dprintk("RPC: %5u %s: no more getport versions available\n", task->tk_pid, __func__); goto bailout_nofree; } - bind_version = info[xprt->bind_index].rpc_vers; dprintk("RPC: %5u %s: trying rpcbind version %u\n", task->tk_pid, __func__, bind_version); @@ -361,7 +362,7 @@ void rpcb_getport_async(struct rpc_task *task) map->r_addr = rpc_peeraddr2str(rpcb_clnt, RPC_DISPLAY_UNIVERSAL_ADDR); map->r_owner = RPCB_OWNER_STRING; /* ignored for GETADDR */ - child = rpcb_call_async(rpcb_clnt, map, xprt->bind_index); + child = rpcb_call_async(rpcb_clnt, map, proc); rpc_release_client(rpcb_clnt); if (IS_ERR(child)) { status = -EIO; From eb35c218d83ec0780d9db869310f2e333f628702 Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Tue, 8 Jul 2008 14:37:06 -0400 Subject: [PATCH 35/46] reiserfs: discard prealloc in reiserfs_delete_inode With the removal of struct file from the xattr code, reiserfs_file_release() isn't used anymore, so the prealloc isn't discarded. This causes hangs later down the line. This patch adds it to reiserfs_delete_inode. In most cases it will be a no-op due to it already having been called, but will avoid hangs with xattrs. Signed-off-by: Jeff Mahoney Signed-off-by: Linus Torvalds --- fs/reiserfs/inode.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c index 57917932212e..192269698a8a 100644 --- a/fs/reiserfs/inode.c +++ b/fs/reiserfs/inode.c @@ -45,6 +45,8 @@ void reiserfs_delete_inode(struct inode *inode) goto out; reiserfs_update_inode_transaction(inode); + reiserfs_discard_prealloc(&th, inode); + err = reiserfs_delete_object(&th, inode); /* Do quota update inside a transaction for journaled quotas. We must do that From 6d8566ac131f3dc3e976de824dcb04feaea4b14c Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:19 +0200 Subject: [PATCH 36/46] x86/pci: Makefile merge: Removing Makefile_* No functional nor ordering changes here. Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/Makefile | 48 ++++++++++++++++++++++++++++++++++++++-- arch/x86/pci/Makefile_32 | 26 ---------------------- arch/x86/pci/Makefile_64 | 17 -------------- 3 files changed, 46 insertions(+), 45 deletions(-) delete mode 100644 arch/x86/pci/Makefile_32 delete mode 100644 arch/x86/pci/Makefile_64 diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index c5c8e485fc44..3a8fb7303ad1 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -1,5 +1,49 @@ ifeq ($(CONFIG_X86_32),y) -include ${srctree}/arch/x86/pci/Makefile_32 + +obj-y := i386.o init.o + +obj-$(CONFIG_PCI_BIOS) += pcbios.o +obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_32.o direct.o mmconfig-shared.o +obj-$(CONFIG_PCI_DIRECT) += direct.o +obj-$(CONFIG_PCI_OLPC) += olpc.o + +pci-y := fixup.o + +# Do not change the ordering here. There is a nasty init function +# ordering dependency which breaks when you move acpi.o below +# legacy/irq.o +pci-$(CONFIG_ACPI) += acpi.o +pci-y += legacy.o irq.o + +# Careful: VISWS overrule the pci-y above. The colons are +# therefor correct. This needs a proper fix by distangling the code. +pci-$(CONFIG_X86_VISWS) := visws.o fixup.o + +pci-$(CONFIG_X86_NUMAQ) += numa.o + +# Necessary for NUMAQ as well +pci-$(CONFIG_NUMA) += mp_bus_to_node.o + +obj-y += $(pci-y) common.o early.o +obj-y += amd_bus.o + else -include ${srctree}/arch/x86/pci/Makefile_64 + +# +# Makefile for X86_64 specific PCI routines +# +# Reuse the i386 PCI subsystem +# +EXTRA_CFLAGS += -Iarch/x86/pci + +obj-y := i386.o +obj-$(CONFIG_PCI_DIRECT)+= direct.o +obj-y += fixup.o init.o +obj-$(CONFIG_ACPI) += acpi.o +obj-y += legacy.o irq.o common.o early.o +# mmconfig has a 64bit special +obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_64.o direct.o mmconfig-shared.o + +obj-y += amd_bus.o + endif diff --git a/arch/x86/pci/Makefile_32 b/arch/x86/pci/Makefile_32 deleted file mode 100644 index a34fbf557926..000000000000 --- a/arch/x86/pci/Makefile_32 +++ /dev/null @@ -1,26 +0,0 @@ -obj-y := i386.o init.o - -obj-$(CONFIG_PCI_BIOS) += pcbios.o -obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_32.o direct.o mmconfig-shared.o -obj-$(CONFIG_PCI_DIRECT) += direct.o -obj-$(CONFIG_PCI_OLPC) += olpc.o - -pci-y := fixup.o - -# Do not change the ordering here. There is a nasty init function -# ordering dependency which breaks when you move acpi.o below -# legacy/irq.o -pci-$(CONFIG_ACPI) += acpi.o -pci-y += legacy.o irq.o - -# Careful: VISWS overrule the pci-y above. The colons are -# therefor correct. This needs a proper fix by distangling the code. -pci-$(CONFIG_X86_VISWS) := visws.o fixup.o - -pci-$(CONFIG_X86_NUMAQ) += numa.o - -# Necessary for NUMAQ as well -pci-$(CONFIG_NUMA) += mp_bus_to_node.o - -obj-y += $(pci-y) common.o early.o -obj-y += amd_bus.o diff --git a/arch/x86/pci/Makefile_64 b/arch/x86/pci/Makefile_64 deleted file mode 100644 index fd47068c95de..000000000000 --- a/arch/x86/pci/Makefile_64 +++ /dev/null @@ -1,17 +0,0 @@ -# -# Makefile for X86_64 specific PCI routines -# -# Reuse the i386 PCI subsystem -# -EXTRA_CFLAGS += -Iarch/x86/pci - -obj-y := i386.o -obj-$(CONFIG_PCI_DIRECT)+= direct.o -obj-y += fixup.o init.o -obj-$(CONFIG_ACPI) += acpi.o -obj-y += legacy.o irq.o common.o early.o -# mmconfig has a 64bit special -obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_64.o direct.o mmconfig-shared.o - -obj-y += amd_bus.o - From 7fa71c0e0e371b995fdf3a367542a014c1f37f29 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:20 +0200 Subject: [PATCH 37/46] x86/pci: Makefile merge: removing include dir flag Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/Makefile | 3 --- 1 file changed, 3 deletions(-) diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index 3a8fb7303ad1..8645094d7423 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -32,9 +32,6 @@ else # # Makefile for X86_64 specific PCI routines # -# Reuse the i386 PCI subsystem -# -EXTRA_CFLAGS += -Iarch/x86/pci obj-y := i386.o obj-$(CONFIG_PCI_DIRECT)+= direct.o From 0b40e84ab750b4ba5fc8d6712c892502dc93b212 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:21 +0200 Subject: [PATCH 38/46] x86/pci: Makefile merge: whitespace changes only Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/Makefile | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index 8645094d7423..30944bc5252d 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -33,14 +33,14 @@ else # Makefile for X86_64 specific PCI routines # -obj-y := i386.o -obj-$(CONFIG_PCI_DIRECT)+= direct.o -obj-y += fixup.o init.o -obj-$(CONFIG_ACPI) += acpi.o -obj-y += legacy.o irq.o common.o early.o +obj-y := i386.o +obj-$(CONFIG_PCI_DIRECT) += direct.o +obj-y += fixup.o init.o +obj-$(CONFIG_ACPI) += acpi.o +obj-y += legacy.o irq.o common.o early.o # mmconfig has a 64bit special -obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_64.o direct.o mmconfig-shared.o +obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_64.o direct.o mmconfig-shared.o -obj-y += amd_bus.o +obj-y += amd_bus.o endif From a5d4fb94f750e439a0722c0a76c9aade3a9199be Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:22 +0200 Subject: [PATCH 39/46] x86/pci: Makefile merge: decoupling options for mp_bus_to_node.o Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/Makefile | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index 30944bc5252d..044fa5bb2ff6 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -14,6 +14,7 @@ pci-y := fixup.o # legacy/irq.o pci-$(CONFIG_ACPI) += acpi.o pci-y += legacy.o irq.o +pci-$(CONFIG_NUMA) += mp_bus_to_node.o # Careful: VISWS overrule the pci-y above. The colons are # therefor correct. This needs a proper fix by distangling the code. @@ -21,9 +22,6 @@ pci-$(CONFIG_X86_VISWS) := visws.o fixup.o pci-$(CONFIG_X86_NUMAQ) += numa.o -# Necessary for NUMAQ as well -pci-$(CONFIG_NUMA) += mp_bus_to_node.o - obj-y += $(pci-y) common.o early.o obj-y += amd_bus.o From 8043c7be4b8dabaaeed8344dd0a4cbaa9a94fb71 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:23 +0200 Subject: [PATCH 40/46] x86/pci: Makefile merge: changing 64bit ordering This should be safe since mmconfig*.o and init.o do not contain *initcalls with the same level as in other files. Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/Makefile | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index 044fa5bb2ff6..d5f15eff2be0 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -31,14 +31,13 @@ else # Makefile for X86_64 specific PCI routines # -obj-y := i386.o -obj-$(CONFIG_PCI_DIRECT) += direct.o -obj-y += fixup.o init.o -obj-$(CONFIG_ACPI) += acpi.o -obj-y += legacy.o irq.o common.o early.o -# mmconfig has a 64bit special +obj-y := i386.o init.o obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_64.o direct.o mmconfig-shared.o - +obj-$(CONFIG_PCI_DIRECT) += direct.o +obj-y += fixup.o +obj-$(CONFIG_ACPI) += acpi.o +obj-y += legacy.o irq.o +obj-y += common.o early.o obj-y += amd_bus.o endif From 8eefc6640311bb56408fa0abe3f2930b3037f91a Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:24 +0200 Subject: [PATCH 41/46] x86/pci: Makefile merge: creating pci-y for 64 bit Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/Makefile | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index d5f15eff2be0..ed130a2ef671 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -34,10 +34,10 @@ else obj-y := i386.o init.o obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_64.o direct.o mmconfig-shared.o obj-$(CONFIG_PCI_DIRECT) += direct.o -obj-y += fixup.o -obj-$(CONFIG_ACPI) += acpi.o -obj-y += legacy.o irq.o -obj-y += common.o early.o +pci-y := fixup.o +pci-$(CONFIG_ACPI) += acpi.o +pci-y += legacy.o irq.o +obj-y += $(pci-y) common.o early.o obj-y += amd_bus.o endif From 42a4b427a8d7bff95d9622ffa2365cb56da8e7d6 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:25 +0200 Subject: [PATCH 42/46] x86/pci: mp_bus_to_node merge: moving code in amd_bus.c Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/amd_bus.c | 42 +++++++++++++++++++++--------------------- 1 file changed, 21 insertions(+), 21 deletions(-) diff --git a/arch/x86/pci/amd_bus.c b/arch/x86/pci/amd_bus.c index d02c598451ec..465b3a6f1c12 100644 --- a/arch/x86/pci/amd_bus.c +++ b/arch/x86/pci/amd_bus.c @@ -14,27 +14,6 @@ * also get peer root bus resource for io,mmio */ - -/* - * sub bus (transparent) will use entres from 3 to store extra from root, - * so need to make sure have enought slot there, increase PCI_BUS_NUM_RESOURCES? - */ -#define RES_NUM 16 -struct pci_root_info { - char name[12]; - unsigned int res_num; - struct resource res[RES_NUM]; - int bus_min; - int bus_max; - int node; - int link; -}; - -/* 4 at this time, it may become to 32 */ -#define PCI_ROOT_NR 4 -static int pci_root_num; -static struct pci_root_info pci_root_info[PCI_ROOT_NR]; - #ifdef CONFIG_NUMA #define BUS_NR 256 @@ -65,8 +44,29 @@ int get_mp_bus_to_node(int busnum) return node; } + #endif +/* + * sub bus (transparent) will use entres from 3 to store extra from root, + * so need to make sure have enought slot there, increase PCI_BUS_NUM_RESOURCES? + */ +#define RES_NUM 16 +struct pci_root_info { + char name[12]; + unsigned int res_num; + struct resource res[RES_NUM]; + int bus_min; + int bus_max; + int node; + int link; +}; + +/* 4 at this time, it may become to 32 */ +#define PCI_ROOT_NR 4 +static int pci_root_num; +static struct pci_root_info pci_root_info[PCI_ROOT_NR]; + void set_pci_bus_resources_arch_default(struct pci_bus *b) { int i; From d199a0482791bdf2f052081aa212d8c83ccef88a Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:26 +0200 Subject: [PATCH 43/46] x86/pci: merge: moving mp_bus_to_node.c to amd_bus.c Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/Makefile | 1 - arch/x86/pci/amd_bus.c | 32 +++++++++++++++++++++++++++++--- arch/x86/pci/mp_bus_to_node.c | 23 ----------------------- 3 files changed, 29 insertions(+), 27 deletions(-) delete mode 100644 arch/x86/pci/mp_bus_to_node.c diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index ed130a2ef671..f3643a7be85b 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -14,7 +14,6 @@ pci-y := fixup.o # legacy/irq.o pci-$(CONFIG_ACPI) += acpi.o pci-y += legacy.o irq.o -pci-$(CONFIG_NUMA) += mp_bus_to_node.o # Careful: VISWS overrule the pci-y above. The colons are # therefor correct. This needs a proper fix by distangling the code. diff --git a/arch/x86/pci/amd_bus.c b/arch/x86/pci/amd_bus.c index 465b3a6f1c12..a18141ae3f02 100644 --- a/arch/x86/pci/amd_bus.c +++ b/arch/x86/pci/amd_bus.c @@ -1,13 +1,13 @@ #include #include +#include #include "pci.h" #ifdef CONFIG_X86_64 - #include #include #include -#include +#endif /* * This discovers the pcibus <-> node mapping on AMD K8. @@ -18,6 +18,8 @@ #define BUS_NR 256 +#ifdef CONFIG_X86_64 + static int mp_bus_to_node[BUS_NR]; void set_mp_bus_to_node(int busnum, int node) @@ -45,7 +47,31 @@ int get_mp_bus_to_node(int busnum) return node; } -#endif +#else /* CONFIG_X86_32 */ + +static unsigned char mp_bus_to_node[BUS_NR]; + +void set_mp_bus_to_node(int busnum, int node) +{ + if (busnum >= 0 && busnum < BUS_NR) + mp_bus_to_node[busnum] = (unsigned char) node; +} + +int get_mp_bus_to_node(int busnum) +{ + int node; + + if (busnum < 0 || busnum > (BUS_NR - 1)) + return 0; + node = mp_bus_to_node[busnum]; + return node; +} + +#endif /* CONFIG_X86_32 */ + +#endif /* CONFIG_NUMA */ + +#ifdef CONFIG_X86_64 /* * sub bus (transparent) will use entres from 3 to store extra from root, diff --git a/arch/x86/pci/mp_bus_to_node.c b/arch/x86/pci/mp_bus_to_node.c deleted file mode 100644 index 022943999b84..000000000000 --- a/arch/x86/pci/mp_bus_to_node.c +++ /dev/null @@ -1,23 +0,0 @@ -#include -#include -#include - -#define BUS_NR 256 - -static unsigned char mp_bus_to_node[BUS_NR]; - -void set_mp_bus_to_node(int busnum, int node) -{ - if (busnum >= 0 && busnum < BUS_NR) - mp_bus_to_node[busnum] = (unsigned char) node; -} - -int get_mp_bus_to_node(int busnum) -{ - int node; - - if (busnum < 0 || busnum > (BUS_NR - 1)) - return 0; - node = mp_bus_to_node[busnum]; - return node; -} From f4d994b541198446fc3d5d4ec5de232e171bf6dd Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:27 +0200 Subject: [PATCH 44/46] x86/pci: Makefile merge: coalescing 32 and 64 bit Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/Makefile | 21 +-------------------- 1 file changed, 1 insertion(+), 20 deletions(-) diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index f3643a7be85b..1ba1236094a4 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -1,9 +1,7 @@ -ifeq ($(CONFIG_X86_32),y) - obj-y := i386.o init.o obj-$(CONFIG_PCI_BIOS) += pcbios.o -obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_32.o direct.o mmconfig-shared.o +obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_$(BITS).o direct.o mmconfig-shared.o obj-$(CONFIG_PCI_DIRECT) += direct.o obj-$(CONFIG_PCI_OLPC) += olpc.o @@ -23,20 +21,3 @@ pci-$(CONFIG_X86_NUMAQ) += numa.o obj-y += $(pci-y) common.o early.o obj-y += amd_bus.o - -else - -# -# Makefile for X86_64 specific PCI routines -# - -obj-y := i386.o init.o -obj-$(CONFIG_PCI_MMCONFIG) += mmconfig_64.o direct.o mmconfig-shared.o -obj-$(CONFIG_PCI_DIRECT) += direct.o -pci-y := fixup.o -pci-$(CONFIG_ACPI) += acpi.o -pci-y += legacy.o irq.o -obj-y += $(pci-y) common.o early.o -obj-y += amd_bus.o - -endif From ae28d70529b874f657afd0f70ca148599345ba2a Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:28 +0200 Subject: [PATCH 45/46] x86/pci: visws: renaming pcibios_init() Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/visws.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/x86/pci/visws.c b/arch/x86/pci/visws.c index c2df4e97eed6..c3dc9f26812c 100644 --- a/arch/x86/pci/visws.c +++ b/arch/x86/pci/visws.c @@ -85,7 +85,7 @@ void __init pcibios_update_irq(struct pci_dev *dev, int irq) pci_write_config_byte(dev, PCI_INTERRUPT_LINE, irq); } -static int __init pcibios_init(void) +static int __init pci_visws_init(void) { /* The VISWS supports configuration access type 1 only */ pci_probe = (pci_probe | PCI_PROBE_CONF1) & @@ -105,4 +105,4 @@ static int __init pcibios_init(void) return 0; } -subsys_initcall(pcibios_init); +subsys_initcall(pci_visws_init); From 8dd779b19ce5972072ad2372a86c8acbae4da768 Mon Sep 17 00:00:00 2001 From: Robert Richter Date: Wed, 2 Jul 2008 22:50:29 +0200 Subject: [PATCH 46/46] x86/pci: removing subsys_initcall ordering dependencies So far subsys_initcalls has been executed in this order depending on the object order in the Makefile: arch/x86/pci/visws.c:subsys_initcall(pcibios_init); arch/x86/pci/numa.c:subsys_initcall(pci_numa_init); arch/x86/pci/acpi.c:subsys_initcall(pci_acpi_init); arch/x86/pci/legacy.c:subsys_initcall(pci_legacy_init); arch/x86/pci/irq.c:subsys_initcall(pcibios_irq_init); arch/x86/pci/common.c:subsys_initcall(pcibios_init); This patch removes the ordering dependency. There is now only one subsys_initcall function that contains subsystem initialization code with a defined order. Signed-off-by: Robert Richter Acked-by: Jesse Barnes Signed-off-by: Ingo Molnar --- arch/x86/pci/Makefile | 4 ---- arch/x86/pci/acpi.c | 3 +-- arch/x86/pci/common.c | 4 +--- arch/x86/pci/init.c | 4 ++-- arch/x86/pci/irq.c | 5 +---- arch/x86/pci/legacy.c | 11 ++++++++++- arch/x86/pci/numa.c | 8 +++++++- arch/x86/pci/pci.h | 11 ++++++++--- arch/x86/pci/visws.c | 7 ++++++- 9 files changed, 36 insertions(+), 21 deletions(-) diff --git a/arch/x86/pci/Makefile b/arch/x86/pci/Makefile index 1ba1236094a4..99d9f095e4d4 100644 --- a/arch/x86/pci/Makefile +++ b/arch/x86/pci/Makefile @@ -6,10 +6,6 @@ obj-$(CONFIG_PCI_DIRECT) += direct.o obj-$(CONFIG_PCI_OLPC) += olpc.o pci-y := fixup.o - -# Do not change the ordering here. There is a nasty init function -# ordering dependency which breaks when you move acpi.o below -# legacy/irq.o pci-$(CONFIG_ACPI) += acpi.o pci-y += legacy.o irq.o diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c index 4fa52d3dc848..19af06927fbc 100644 --- a/arch/x86/pci/acpi.c +++ b/arch/x86/pci/acpi.c @@ -223,7 +223,7 @@ struct pci_bus * __devinit pci_acpi_scan_root(struct acpi_device *device, int do return bus; } -static int __init pci_acpi_init(void) +int __init pci_acpi_init(void) { struct pci_dev *dev = NULL; @@ -257,4 +257,3 @@ static int __init pci_acpi_init(void) return 0; } -subsys_initcall(pci_acpi_init); diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index 6e64aaf00d1d..20b9f59f95df 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -384,7 +384,7 @@ struct pci_bus * __devinit pcibios_scan_root(int busnum) extern u8 pci_cache_line_size; -static int __init pcibios_init(void) +int __init pcibios_init(void) { struct cpuinfo_x86 *c = &boot_cpu_data; @@ -411,8 +411,6 @@ static int __init pcibios_init(void) return 0; } -subsys_initcall(pcibios_init); - char * __devinit pcibios_setup(char *str) { if (!strcmp(str, "off")) { diff --git a/arch/x86/pci/init.c b/arch/x86/pci/init.c index b821f4462d99..d6c950f81858 100644 --- a/arch/x86/pci/init.c +++ b/arch/x86/pci/init.c @@ -4,7 +4,7 @@ /* arch_initcall has too random ordering, so call the initializers in the right sequence from here. */ -static __init int pci_access_init(void) +static __init int pci_arch_init(void) { #ifdef CONFIG_PCI_DIRECT int type = 0; @@ -40,4 +40,4 @@ static __init int pci_access_init(void) return 0; } -arch_initcall(pci_access_init); +arch_initcall(pci_arch_init); diff --git a/arch/x86/pci/irq.c b/arch/x86/pci/irq.c index f0859de23e20..dc568c6b83f8 100644 --- a/arch/x86/pci/irq.c +++ b/arch/x86/pci/irq.c @@ -1107,7 +1107,7 @@ static struct dmi_system_id __initdata pciirq_dmi_table[] = { { } }; -static int __init pcibios_irq_init(void) +int __init pcibios_irq_init(void) { DBG(KERN_DEBUG "PCI: IRQ init\n"); @@ -1142,9 +1142,6 @@ static int __init pcibios_irq_init(void) return 0; } -subsys_initcall(pcibios_irq_init); - - static void pirq_penalize_isa_irq(int irq, int active) { /* diff --git a/arch/x86/pci/legacy.c b/arch/x86/pci/legacy.c index a67921ce60af..3c1d795cbbe9 100644 --- a/arch/x86/pci/legacy.c +++ b/arch/x86/pci/legacy.c @@ -55,4 +55,13 @@ static int __init pci_legacy_init(void) return 0; } -subsys_initcall(pci_legacy_init); +int __init pci_subsys_init(void) +{ +#ifdef CONFIG_ACPI + pci_acpi_init(); +#endif + pci_legacy_init(); + pcibios_irq_init(); + pcibios_init(); +} +subsys_initcall(pci_subsys_init); diff --git a/arch/x86/pci/numa.c b/arch/x86/pci/numa.c index 99f1ecd485b5..e1620dc8649a 100644 --- a/arch/x86/pci/numa.c +++ b/arch/x86/pci/numa.c @@ -177,4 +177,10 @@ static int __init pci_numa_init(void) return 0; } -subsys_initcall(pci_numa_init); +static __init int pci_subsys_init(void) +{ + pci_numa_init(); + pcibios_irq_init(); + pcibios_init(); +} +subsys_initcall(pci_subsys_init); diff --git a/arch/x86/pci/pci.h b/arch/x86/pci/pci.h index ba263e626a68..58241748470f 100644 --- a/arch/x86/pci/pci.h +++ b/arch/x86/pci/pci.h @@ -39,9 +39,6 @@ enum pci_bf_sort_state { pci_dmi_bf, }; -extern void __init dmi_check_pciprobe(void); -extern void __init dmi_check_skip_isa_align(void); - /* pci-i386.c */ extern unsigned int pcibios_max_latency; @@ -99,10 +96,18 @@ extern struct pci_raw_ops *raw_pci_ext_ops; extern struct pci_raw_ops pci_direct_conf1; +/* arch_initcall level */ extern int pci_direct_probe(void); extern void pci_direct_init(int type); extern void pci_pcbios_init(void); extern int pci_olpc_init(void); +extern void __init dmi_check_pciprobe(void); +extern void __init dmi_check_skip_isa_align(void); + +/* some common used subsys_initcalls */ +extern int __init pci_acpi_init(void); +extern int __init pcibios_irq_init(void); +extern int __init pcibios_init(void); /* pci-mmconfig.c */ diff --git a/arch/x86/pci/visws.c b/arch/x86/pci/visws.c index c3dc9f26812c..16e52063ecb3 100644 --- a/arch/x86/pci/visws.c +++ b/arch/x86/pci/visws.c @@ -105,4 +105,9 @@ static int __init pci_visws_init(void) return 0; } -subsys_initcall(pci_visws_init); +static __init int pci_subsys_init(void) +{ + pci_visws_init(); + pcibios_init(); +} +subsys_initcall(pci_subsys_init);