Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 25 Mar 2005 09:35:59 +0800
From:      wsk <wsk@gddsn.org.cn>
To:        John Baldwin <jhb@FreeBSD.org>, current@freebsd.org
Subject:   Re: <SOLEVD> Rocketport uPCI ioaddr mapping failed under FreeBSD-5.3&CURRENT
Message-ID:  <42436AFF.2020501@gddsn.org.cn>
In-Reply-To: <200503241045.42715.jhb@FreeBSD.org>
References:  <424291B5.1060204@gddsn.org.cn> <200503241045.42715.jhb@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help
John Baldwin wrote:

>On Thursday 24 March 2005 05:08 am, wsk wrote:
>  
>
>>>Can you try this patch below and verify it works?  If so, I'd like to
>>>commit it.  Thanks!
>>>
>>>Index: rp_pci.c
>>>===================================================================
>>>RCS file: /usr/cvs/src/sys/dev/rp/rp_pci.c,v
>>>retrieving revision 1.9
>>>diff -u -r1.9 rp_pci.c
>>>--- rp_pci.c	21 Jun 2004 13:02:25 -0000	1.9
>>>+++ rp_pci.c	23 Mar 2005 19:37:43 -0000
>>>@@ -67,6 +67,7 @@
>>>#define RP_DEVICE_ID_4J		0x0007
>>>#define RP_DEVICE_ID_6M		0x000C
>>>#define RP_DEVICE_ID_4M		0x000D
>>>+#define RP_DEVICE_ID_UPCI_8O	0x0805
>>>
>>>/************************************************************************
>>>** MUDBAC remapped for PCI
>>>@@ -177,7 +178,14 @@
>>>
>>>	ctlp->bus_ctlp = NULL;
>>>
>>>-	ctlp->io_rid[0] = 0x10;
>>>+	switch (pci_get_devid(dev)) {
>>>+	case RP_DEVICE_ID_UPCI_8O:
>>>+		ctlp->io_rid[0] = PCIR_BAR(2);
>>>+		break;
>>>+	default:
>>>+		ctlp->io_rid[0] = PCIR_BAR(0);
>>>+		break;
>>>+	}
>>>	ctlp->io[0] = bus_alloc_resource_any(dev, SYS_RES_IOPORT,
>>>		&ctlp->io_rid[0], RF_ACTIVE);
>>>	if(ctlp->io[0] == NULL) {
>>>      
>>>
>>    I've been tested and works failled after patched your diff! here is
>>my boot msgs:
>>rp0: <RocketPort PCI> port 0xdd80-0xddff,0xde00-0xdeff mem
>>0xfcedff80-0xfcedffff irq 19 at device 10.0 on pci2
>>rp0: failed: rid 0x10 is memory, requested 4
>>rp0: ioaddr mapping failed for RocketPort(PCI).
>>device_attach: rp0 attach returned 6
>>    
>>
>
>Oh, I see the bug.  Can you try this patch instead?  It fixes the bug and 
>simplifies a few other places in the code:
>
>Index: rp_pci.c
>===================================================================
>RCS file: /usr/cvs/src/sys/dev/rp/rp_pci.c,v
>retrieving revision 1.10
>diff -u -r1.10 rp_pci.c
>--- rp_pci.c	1 Mar 2005 08:58:04 -0000	1.10
>+++ rp_pci.c	24 Mar 2005 15:44:38 -0000
>@@ -67,6 +67,7 @@
> #define RP_DEVICE_ID_4J		0x0007
> #define RP_DEVICE_ID_6M		0x000C
> #define RP_DEVICE_ID_4M		0x000D
>+#define RP_DEVICE_ID_UPCI_8O	0x0805
> 
> /**************************************************************************
>   MUDBAC remapped for PCI
>@@ -129,7 +130,7 @@
> 	char *s;
> 
> 	s = NULL;
>-	if ((pci_get_devid(dev) & 0xffff) == RP_VENDOR_ID)
>+	if (pci_get_vendor(dev) == RP_VENDOR_ID)
> 		s = "RocketPort PCI";
> 
> 	if (s != NULL) {
>@@ -177,7 +178,14 @@
> 
> 	ctlp->bus_ctlp = NULL;
> 
>-	ctlp->io_rid[0] = 0x10;
>+	switch (pci_get_device(dev)) {
>+	case RP_DEVICE_ID_UPCI_8O:
>+		ctlp->io_rid[0] = PCIR_BAR(2);
>+		break;
>+	default:
>+		ctlp->io_rid[0] = PCIR_BAR(0);
>+		break;
>+	}
> 	ctlp->io[0] = bus_alloc_resource_any(dev, SYS_RES_IOPORT,
> 		&ctlp->io_rid[0], RF_ACTIVE);
> 	if(ctlp->io[0] == NULL) {
>@@ -188,7 +196,7 @@
> 
> 	num_aiops = sPCIInitController(ctlp,
> 				       MAX_AIOPS_PER_BOARD, 0,
>-				       FREQ_DIS, 0, (pci_get_devid(dev) >> 16) & 0xffff);
>+				       FREQ_DIS, 0, pci_get_device(dev));
> 
> 	num_ports = 0;
> 	for(aiop=0; aiop < num_aiops; aiop++) {
>
>
>  
>
yep.it works excellent and hopes commit to the src pls. thanx



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?42436AFF.2020501>