Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754531Ab0KXNbt (ORCPT ); Wed, 24 Nov 2010 08:31:49 -0500 Received: from metis.ext.pengutronix.de ([92.198.50.35]:35264 "EHLO metis.ext.pengutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754187Ab0KXNbr (ORCPT ); Wed, 24 Nov 2010 08:31:47 -0500 Message-ID: <4CED13B5.5020309@pengutronix.de> Date: Wed, 24 Nov 2010 14:31:33 +0100 From: Marc Kleine-Budde Organization: Pengutronix User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.15) Gecko/20101027 Thunderbird/3.0.10 MIME-Version: 1.0 To: Tomoya MORINAGA CC: Wolfgang Grandegger , Wolfram Sang , Christian Pellegrin , Barry Song <21cnbao@gmail.com>, Samuel Ortiz , socketcan-core@lists.berlios.de, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, "David S. Miller" , andrew.chih.howe.khor@intel.com, qi.wang@intel.com, margie.foster@intel.com, yong.y.wang@intel.com, kok.howg.ewe@intel.com, joel.clark@intel.com Subject: Re: [PATCH net-next-2.6 6/17 v3] can: EG20T PCH: Fix endianness issue References: <4CED023E.1070009@dsn.okisemi.com> In-Reply-To: <4CED023E.1070009@dsn.okisemi.com> X-Enigmail-Version: 1.0.1 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="------------enigF36C65A1CE5F6BAE3857FD71" X-SA-Exim-Connect-IP: 2001:6f8:1178:4:5054:ff:fe8d:eefb X-SA-Exim-Mail-From: mkl@pengutronix.de X-SA-Exim-Scanned: No (on metis.ext.pengutronix.de); SAEximRunCond expanded to false X-PTX-Original-Recipient: linux-kernel@vger.kernel.org Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 17004 Lines: 533 This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --------------enigF36C65A1CE5F6BAE3857FD71 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable On 11/24/2010 01:17 PM, Tomoya MORINAGA wrote: > Fix endianness issue. > there is endianness issue both Tx and Rx. > Currently, data is set like below. > Register: > MSB--LSB > x x D0 D1 > x x D2 D3 > x x D4 D5 > x x D6 D7 >=20 > But Data to be sent must be set like below. > Register: > MSB--LSB > x x D1 D0 > x x D3 D2 > x x D5 D4 > x x D7 D6 (x means reserved area.) The endianess handling looks good now! (the & 0xff can be removed, though, see linline) >=20 >=20 > Modify netif_rx() to netif_receive_skb(). Please give a reason, I suppose because the function is used from NAPI. (Frankly speaking this has nothing to do with endianess issues). > For easy to read, some sub-functions are created. >=20 >=20 > Modify complex "goto" to do~while. What about the next_flag in the rx_normal function? Can you get rid of it, too? >=20 > Signed-off-by: Tomoya MORINAGA > --- > From 468cf576aadc0ce26496eaf13fbac4bb2670a9de Mon Sep 17 00:00:00 2001 > From: Tomoya MORINAGA > Date: Thu, 18 Nov 2010 10:23:32 +0900 > Subject: [PATCH] CAN : Improve RxTx processing >=20 > Signed-off-by: Tomoya MORINAGA > --- > drivers/net/can/pch_can.c | 306 > +++++++++++++++++++++++---------------------- > 1 files changed, 156 insertions(+), 150 deletions(-) >=20 > diff --git a/drivers/net/can/pch_can.c b/drivers/net/can/pch_can.c > index 843bbce..fa47707 100644 > --- a/drivers/net/can/pch_can.c > +++ b/drivers/net/can/pch_can.c > @@ -133,10 +133,7 @@ struct pch_can_if_regs { > u32 id1; > u32 id2; > u32 mcont; > - u32 dataa1; > - u32 dataa2; > - u32 datab1; > - u32 datab2; > + u32 data[4]; > u32 rsv[13]; > }; >=20 > @@ -421,10 +418,10 @@ static void pch_can_clear_buffers(struct > pch_can_priv *priv) > iowrite32(0x0, &priv->regs->ifregs[0].id1); > iowrite32(0x0, &priv->regs->ifregs[0].id2); > iowrite32(0x0, &priv->regs->ifregs[0].mcont); > - iowrite32(0x0, &priv->regs->ifregs[0].dataa1); > - iowrite32(0x0, &priv->regs->ifregs[0].dataa2); > - iowrite32(0x0, &priv->regs->ifregs[0].datab1); > - iowrite32(0x0, &priv->regs->ifregs[0].datab2); > + iowrite32(0x0, &priv->regs->ifregs[0].data[0]); > + iowrite32(0x0, &priv->regs->ifregs[0].data[1]); > + iowrite32(0x0, &priv->regs->ifregs[0].data[2]); > + iowrite32(0x0, &priv->regs->ifregs[0].data[3]); > iowrite32(PCH_CMASK_RDWR | PCH_CMASK_MASK | > PCH_CMASK_ARB | PCH_CMASK_CTRL, > &priv->regs->ifregs[0].cmask); > @@ -438,10 +435,10 @@ static void pch_can_clear_buffers(struct > pch_can_priv *priv) > iowrite32(0x0, &priv->regs->ifregs[1].id1); > iowrite32(0x0, &priv->regs->ifregs[1].id2); > iowrite32(0x0, &priv->regs->ifregs[1].mcont); > - iowrite32(0x0, &priv->regs->ifregs[1].dataa1); > - iowrite32(0x0, &priv->regs->ifregs[1].dataa2); > - iowrite32(0x0, &priv->regs->ifregs[1].datab1); > - iowrite32(0x0, &priv->regs->ifregs[1].datab2); > + iowrite32(0x0, &priv->regs->ifregs[1].data[0]); > + iowrite32(0x0, &priv->regs->ifregs[1].data[1]); > + iowrite32(0x0, &priv->regs->ifregs[1].data[2]); > + iowrite32(0x0, &priv->regs->ifregs[1].data[3]); > iowrite32(PCH_CMASK_RDWR | PCH_CMASK_MASK | > PCH_CMASK_ARB | PCH_CMASK_CTRL, > &priv->regs->ifregs[1].cmask); > @@ -683,7 +680,7 @@ static void pch_can_error(struct net_device *ndev, > u32 status) > cf->data[7] =3D (errc & PCH_REC) >> 8; >=20 > priv->can.state =3D state; > - netif_rx(skb); > + netif_receive_skb(skb); This is an unrelated fix. Please split into separate patch. >=20 > stats->rx_packets++; > stats->rx_bytes +=3D cf->can_dlc; > @@ -701,190 +698,202 @@ static irqreturn_t pch_can_interrupt(int irq, > void *dev_id) > return IRQ_HANDLED; > } >=20 > -static int pch_can_rx_normal(struct net_device *ndev, u32 int_stat) > +static void pch_fifo_thresh(struct pch_can_priv *priv, int obj_id) > +{ > + if (obj_id < PCH_FIFO_THRESH) { > + iowrite32(PCH_CMASK_RDWR | PCH_CMASK_CTRL | > + PCH_CMASK_ARB, &priv->regs->ifregs[0].cmask); > + > + /* Clearing the Dir bit. */ > + pch_can_bit_clear(&priv->regs->ifregs[0].id2, PCH_ID2_DIR); > + > + /* Clearing NewDat & IntPnd */ > + pch_can_bit_clear(&priv->regs->ifregs[0].mcont, > + PCH_IF_MCONT_INTPND); > + pch_can_check_if_busy(&priv->regs->ifregs[0].creq, obj_id); > + } else if (obj_id > PCH_FIFO_THRESH) { > + pch_can_int_clr(priv, obj_id); > + } else if (obj_id =3D=3D PCH_FIFO_THRESH) { > + int cnt; > + for (cnt =3D 0; cnt < PCH_FIFO_THRESH; cnt++) > + pch_can_int_clr(priv, cnt+1); ^^^^^ nitpick: there should be spaces around the "+". > + } > +} > + > +static int pch_can_rx_msg_lost(struct net_device *ndev, int obj_id) > +{ > + struct pch_can_priv *priv =3D netdev_priv(ndev); > + struct net_device_stats *stats =3D &(priv->ndev->stats); > + struct sk_buff *skb; > + struct can_frame *cf; > + > + netdev_dbg(priv->ndev, "Msg Obj is overwritten.\n"); > + pch_can_bit_clear(&priv->regs->ifregs[0].mcont, > + PCH_IF_MCONT_MSGLOST); > + iowrite32(PCH_CMASK_RDWR | PCH_CMASK_CTRL, > + &priv->regs->ifregs[0].cmask); > + pch_can_check_if_busy(&priv->regs->ifregs[0].creq, obj_id); > + > + skb =3D alloc_can_err_skb(ndev, &cf); > + if (!skb) > + return -ENOMEM; > + > + cf->can_id |=3D CAN_ERR_CRTL; > + cf->data[1] =3D CAN_ERR_CRTL_RX_OVERFLOW; > + stats->rx_over_errors++; > + stats->rx_errors++; > + > + netif_receive_skb(skb); > + > + return 0; > +} > + > +static int pch_can_rx_normal(struct net_device *ndev, u32 obj_num, int= > quota) > { > u32 reg; > canid_t id; > - u32 ide; > - u32 rtr; > - int i, j, k; > int rcv_pkts =3D 0; > + int rtn; > + int next_flag =3D 0; > struct sk_buff *skb; > struct can_frame *cf; > struct pch_can_priv *priv =3D netdev_priv(ndev); > struct net_device_stats *stats =3D &(priv->ndev->stats); > + int i; > + u32 id2; > + u16 data_reg; >=20 > - /* Reading the messsage object from the Message RAM */ > - iowrite32(PCH_CMASK_RX_TX_GET, &priv->regs->ifregs[0].cmask); > - pch_can_check_if_busy(&priv->regs->ifregs[0].creq, int_stat); > + do { > + /* Reading the messsage object from the Message RAM */ > + iowrite32(PCH_CMASK_RX_TX_GET, &priv->regs->ifregs[0].cmask); > + pch_can_check_if_busy(&priv->regs->ifregs[0].creq, obj_num); >=20 > - /* Reading the MCONT register. */ > - reg =3D ioread32(&priv->regs->ifregs[0].mcont); > - reg &=3D 0xffff; > + /* Reading the MCONT register. */ > + reg =3D ioread32(&priv->regs->ifregs[0].mcont); > + > + if (reg & PCH_IF_MCONT_EOB) > + break; >=20 > - for (k =3D int_stat; !(reg & PCH_IF_MCONT_EOB); k++) { > /* If MsgLost bit set. */ > if (reg & PCH_IF_MCONT_MSGLOST) { > - dev_err(&priv->ndev->dev, "Msg Obj is overwritten.\n"); > - pch_can_bit_clear(&priv->regs->ifregs[0].mcont, > - PCH_IF_MCONT_MSGLOST); > - iowrite32(PCH_CMASK_RDWR | PCH_CMASK_CTRL, > - &priv->regs->ifregs[0].cmask); > - pch_can_check_if_busy(&priv->regs->ifregs[0].creq, k); > - > - skb =3D alloc_can_err_skb(ndev, &cf); > + rtn =3D pch_can_rx_msg_lost(ndev, obj_num); > + if (!rtn) > + return rtn; The function return "0" on success or -ENOMEM. The logic seems inverted. IMHO just keep going if you had an OOM situation. > + rcv_pkts++; > + quota--; > + next_flag =3D 1; > + } else if (!(reg & PCH_IF_MCONT_NEWDAT)) > + next_flag =3D 1; Please get rid of the next_flag, rearrange your code and use "continue". > + if (!next_flag) { > + skb =3D alloc_can_skb(priv->ndev, &cf); > if (!skb) > return -ENOMEM; >=20 > - priv->can.can_stats.error_passive++; > - priv->can.state =3D CAN_STATE_ERROR_PASSIVE; > - cf->can_id |=3D CAN_ERR_CRTL; > - cf->data[1] |=3D CAN_ERR_CRTL_RX_OVERFLOW; > - cf->data[2] |=3D CAN_ERR_PROT_OVERLOAD; > - stats->rx_packets++; > - stats->rx_bytes +=3D cf->can_dlc; > + /* Get Received data */ > + id2 =3D ioread32(&priv->regs->ifregs[0].id2); > + if (id2 & PCH_ID2_XTD) { > + id =3D (ioread32(&priv->regs->ifregs[0].id1) & > + 0xffff); > + id |=3D (((id2) & 0x1fff) << 16); > + cf->can_id =3D id | CAN_EFF_FLAG; > + } else { > + id =3D ((id2 & (CAN_SFF_MASK << 2)) >> 2); > + cf->can_id =3D id; > + } > + > + if (id2 & PCH_ID2_DIR) > + cf->can_id |=3D CAN_RTR_FLAG; > + > + cf->can_dlc =3D get_can_dlc((ioread32(&priv->regs-> > + ifregs[0].mcont)) & 0xF); > + > + for (i =3D 0; i < cf->can_dlc; i +=3D 2) { > + data_reg =3D ioread16(&priv->regs->ifregs[0]. > + data[i / 2]); > + cf->data[i] =3D data_reg & 0xff; > + cf->data[i + 1] =3D (data_reg >> 8) & 0xff; the & 0xff are not needed, cf->data[] is 8bit only. > + } >=20 > netif_receive_skb(skb); > rcv_pkts++; > - goto RX_NEXT; > - } > - if (!(reg & PCH_IF_MCONT_NEWDAT)) > - goto RX_NEXT; > - > - skb =3D alloc_can_skb(priv->ndev, &cf); > - if (!skb) > - return -ENOMEM; > - > - /* Get Received data */ > - ide =3D ((ioread32(&priv->regs->ifregs[0].id2)) & PCH_ID2_XTD) >> > - 14; > - if (ide) { > - id =3D (ioread32(&priv->regs->ifregs[0].id1) & 0xffff); > - id |=3D (((ioread32(&priv->regs->ifregs[0].id2)) & > - 0x1fff) << 16); > - cf->can_id =3D (id & CAN_EFF_MASK) | CAN_EFF_FLAG; > - } else { > - id =3D (((ioread32(&priv->regs->ifregs[0].id2)) & > - (CAN_SFF_MASK << 2)) >> 2); > - cf->can_id =3D (id & CAN_SFF_MASK); > - } > + stats->rx_packets++; > + quota--; > + stats->rx_bytes +=3D cf->can_dlc; >=20 > - rtr =3D (ioread32(&priv->regs->ifregs[0].id2) & PCH_ID2_DIR); > - if (rtr) { > - cf->can_dlc =3D 0; > - cf->can_id |=3D CAN_RTR_FLAG; > - } else { > - cf->can_dlc =3D > - ((ioread32(&priv->regs->ifregs[0].mcont)) & 0x0f); > + pch_fifo_thresh(priv, obj_num); > } > + obj_num++; > + next_flag =3D 0; > + } while (quota > 0); >=20 > - for (i =3D 0, j =3D 0; i < cf->can_dlc; j++) { > - reg =3D ioread32(&priv->regs->ifregs[0].dataa1 + j*4); > - cf->data[i++] =3D cpu_to_le32(reg & 0xff); > - if (i =3D=3D cf->can_dlc) > - break; > - cf->data[i++] =3D cpu_to_le32((reg >> 8) & 0xff); > - } > + return rcv_pkts; > +} >=20 > - netif_receive_skb(skb); > - rcv_pkts++; > - stats->rx_packets++; > - stats->rx_bytes +=3D cf->can_dlc; > - > - if (k < PCH_FIFO_THRESH) { > - iowrite32(PCH_CMASK_RDWR | PCH_CMASK_CTRL | > - PCH_CMASK_ARB, &priv->regs->ifregs[0].cmask); > - > - /* Clearing the Dir bit. */ > - pch_can_bit_clear(&priv->regs->ifregs[0].id2, > - PCH_ID2_DIR); > - > - /* Clearing NewDat & IntPnd */ > - pch_can_bit_clear(&priv->regs->ifregs[0].mcont, > - PCH_IF_MCONT_INTPND); > - pch_can_check_if_busy(&priv->regs->ifregs[0].creq, k); > - } else if (k > PCH_FIFO_THRESH) { > - pch_can_int_clr(priv, k); > - } else if (k =3D=3D PCH_FIFO_THRESH) { > - int cnt; > - for (cnt =3D 0; cnt < PCH_FIFO_THRESH; cnt++) > - pch_can_int_clr(priv, cnt+1); > - } > -RX_NEXT: > - /* Reading the messsage object from the Message RAM */ > - iowrite32(PCH_CMASK_RX_TX_GET, &priv->regs->ifregs[0].cmask); > - pch_can_check_if_busy(&priv->regs->ifregs[0].creq, k); > - reg =3D ioread32(&priv->regs->ifregs[0].mcont); > - } > +static void pch_can_tx_complete(struct net_device *ndev, u32 int_stat)= > +{ > + struct pch_can_priv *priv =3D netdev_priv(ndev); > + struct net_device_stats *stats =3D &(priv->ndev->stats); > + u32 dlc; >=20 > - return rcv_pkts; > + can_get_echo_skb(ndev, int_stat - PCH_RX_OBJ_END - 1); > + iowrite32(PCH_CMASK_RX_TX_GET | PCH_CMASK_CLRINTPND, > + &priv->regs->ifregs[1].cmask); > + dlc =3D get_can_dlc(ioread32(&priv->regs->ifregs[1].mcont) & > + PCH_IF_MCONT_DLC); > + pch_can_check_if_busy(&priv->regs->ifregs[1].creq, int_stat); This is the TX-Complete interrupt handler, right? What does the pch_can_check_if_busy trigger here? > + stats->tx_bytes +=3D dlc; > + stats->tx_packets++; > + if (int_stat =3D=3D PCH_TX_OBJ_END) > + netif_wake_queue(ndev); > } > + > static int pch_can_rx_poll(struct napi_struct *napi, int quota) > { > struct net_device *ndev =3D napi->dev; > struct pch_can_priv *priv =3D netdev_priv(ndev); > - struct net_device_stats *stats =3D &(priv->ndev->stats); > - u32 dlc; > u32 int_stat; > int rcv_pkts =3D 0; > u32 reg_stat; >=20 > int_stat =3D pch_can_int_pending(priv); > if (!int_stat) > - return 0; > + goto end; >=20 > -INT_STAT: > - if (int_stat =3D=3D PCH_STATUS_INT) { > + if ((int_stat =3D=3D PCH_STATUS_INT) && (quota > 0)) { > reg_stat =3D ioread32(&priv->regs->stat); > if (reg_stat & (PCH_BUS_OFF | PCH_LEC_ALL)) { > - if ((reg_stat & PCH_LEC_ALL) !=3D PCH_LEC_ALL) > + if (reg_stat & PCH_BUS_OFF || > + (reg_stat & PCH_LEC_ALL) !=3D PCH_LEC_ALL) { > pch_can_error(ndev, reg_stat); > + quota--; > + } > } >=20 > - if (reg_stat & PCH_TX_OK) { > - iowrite32(PCH_CMASK_RX_TX_GET, > - &priv->regs->ifregs[1].cmask); > - pch_can_check_if_busy(&priv->regs->ifregs[1].creq, > - ioread32(&priv->regs->intr)); > + if (reg_stat & PCH_TX_OK) > pch_can_bit_clear(&priv->regs->stat, PCH_TX_OK); > - } >=20 > if (reg_stat & PCH_RX_OK) > pch_can_bit_clear(&priv->regs->stat, PCH_RX_OK); >=20 > int_stat =3D pch_can_int_pending(priv); > - if (int_stat =3D=3D PCH_STATUS_INT) > - goto INT_STAT; > } >=20 > -MSG_OBJ: > + if (quota =3D=3D 0) > + goto end; > + > if ((int_stat >=3D PCH_RX_OBJ_START) && (int_stat <=3D PCH_RX_OBJ_END= )) { > - rcv_pkts =3D pch_can_rx_normal(ndev, int_stat); > - if (rcv_pkts < 0) > - return 0; > + rcv_pkts +=3D pch_can_rx_normal(ndev, int_stat, quota); > + quota -=3D rcv_pkts; > + if (quota < 0) > + goto end; > } else if ((int_stat >=3D PCH_TX_OBJ_START) && > (int_stat <=3D PCH_TX_OBJ_END)) { > /* Handle transmission interrupt */ > - can_get_echo_skb(ndev, int_stat - PCH_RX_OBJ_END - 1); > - iowrite32(PCH_CMASK_RX_TX_GET | PCH_CMASK_CLRINTPND, > - &priv->regs->ifregs[1].cmask); > - dlc =3D ioread32(&priv->regs->ifregs[1].mcont) & > - PCH_IF_MCONT_DLC; > - pch_can_check_if_busy(&priv->regs->ifregs[1].creq, int_stat); > - if (dlc > 8) > - dlc =3D 8; > - stats->tx_bytes +=3D dlc; > - stats->tx_packets++; > - if (int_stat =3D=3D PCH_TX_OBJ_END) > - netif_wake_queue(ndev); > + pch_can_tx_complete(ndev, int_stat); > } >=20 > - int_stat =3D pch_can_int_pending(priv); > - if (int_stat =3D=3D PCH_STATUS_INT) > - goto INT_STAT; > - else if (int_stat >=3D 1 && int_stat <=3D 32) > - goto MSG_OBJ; > - > +end: > napi_complete(napi); > pch_can_set_int_enables(priv, PCH_CAN_ALL); >=20 > @@ -1017,10 +1026,10 @@ static int pch_close(struct net_device *ndev) >=20 > static netdev_tx_t pch_xmit(struct sk_buff *skb, struct net_device *nd= ev) > { > - int i, j; > struct pch_can_priv *priv =3D netdev_priv(ndev); > struct can_frame *cf =3D (struct can_frame *)skb->data; > int tx_obj_no =3D 0; > + int i; >=20 > if (can_dropped_invalid_skb(ndev, skb)) > return NETDEV_TX_OK; > @@ -1061,13 +1070,10 @@ static netdev_tx_t pch_xmit(struct sk_buff *skb= , > struct net_device *ndev) > if (cf->can_id & CAN_RTR_FLAG) > pch_can_bit_clear(&priv->regs->ifregs[1].id2, PCH_ID2_DIR); >=20 > - for (i =3D 0, j =3D 0; i < cf->can_dlc; j++) { > - iowrite32(le32_to_cpu(cf->data[i++]), > - (&priv->regs->ifregs[1].dataa1) + j*4); > - if (i =3D=3D cf->can_dlc) > - break; > - iowrite32(le32_to_cpu(cf->data[i++] << 8), > - (&priv->regs->ifregs[1].dataa1) + j*4); > + /* Copy data to register */ > + for (i =3D 0; i < cf->can_dlc; i +=3D 2) { > + iowrite16(cf->data[i] | (cf->data[i + 1] << 8), > + &priv->regs->ifregs[1].data[i / 2]); > } >=20 > can_put_echo_skb(skb, ndev, tx_obj_no - PCH_RX_OBJ_END - 1); cheers, Marc --=20 Pengutronix e.K. | Marc Kleine-Budde | Industrial Linux Solutions | Phone: +49-231-2826-924 | Vertretung West/Dortmund | Fax: +49-5121-206917-5555 | Amtsgericht Hildesheim, HRA 2686 | http://www.pengutronix.de | --------------enigF36C65A1CE5F6BAE3857FD71 Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkztE7gACgkQjTAFq1RaXHOBlQCfRJvevx8DwGk+fARuR0ag3i53 jjQAoJOHCUEB/hhqejFX0n99y2lXTUjw =Ydo1 -----END PGP SIGNATURE----- --------------enigF36C65A1CE5F6BAE3857FD71-- -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/