* Re: tg3 timeouts with 2.6.17-rc6
From: Juergen Kreileder @ 2006-06-16 21:20 UTC (permalink / raw)
To: Michael Chan; +Cc: linux-kernel, netdev
In-Reply-To: <1551EAE59135BE47B544934E30FC4FC041BD16@NT-IRVA-0751.brcm.ad.broadcom.com>
Michael Chan <mchan@broadcom.com> writes:
> Juergen Kreileder
>
>> I'm seeing frequent network timeouts on my PowerMac G5 Quad with
>> 2.6.17-rc6. The timeouts are easily reproducible with moderate
>> network traffic, e.g. by using bittorrent.
>>
>
> Did this use to work with an older kernel or older tg3 driver? If
> yes, what version?
Works fine with 2.6.16 and earlier.
> Please also provide the full tg3 probing output during modprobe and
> ifconfig up. Thanks.
,----
| tg3.c:v3.58 (May 22, 2006)
| eth0: Tigon3 [partno(BCM95780) rev 8003 PHY(5780)] (PCIX:133MHz:64-bit) 10/100/1000BaseT Ethernet 00:14:51:66:ff:b2
| eth0: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] Split[0] WireSpeed[1] TSOcap[1]
| eth0: dma_rwctrl[76144000] dma_mask[40-bit]
| eth1: Tigon3 [partno(BCM95780) rev 8003 PHY(5780)] (PCIX:133MHz:64-bit) 10/100/1000BaseT Ethernet 00:14:51:66:ff:b3
| eth1: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] Split[0] WireSpeed[1] TSOcap[1]
| eth1: dma_rwctrl[76144000] dma_mask[40-bit]
| [...]
| tg3: lan0: Link is up at 1000 Mbps, full duplex.
| tg3: lan0: Flow control is on for TX and on for RX.
`----
eth0 and eth1 get renamed to lan0 and lan1 by udev. eth0/lan0 is
connected to a Gigabit switch eth1/lan1 is not connected.
Both interface are controlled by ifplugd (using SIOCETHTOOL for link
beat detection).
Let me know if you need more debugging output.
Juergen
--
Juergen Kreileder, Blackdown Java-Linux Team
http://blog.blackdown.de/
^ permalink raw reply
* Re: tg3 timeouts with 2.6.17-rc6
From: Michael Chan @ 2006-06-16 21:42 UTC (permalink / raw)
To: Juergen Kreileder; +Cc: linux-kernel, netdev
In-Reply-To: <87k67glrvl.fsf@blackdown.de>
On Fri, 2006-06-16 at 23:20 +0200, Juergen Kreileder wrote:
> Michael Chan <mchan@broadcom.com> writes:
>
> >
> > Did this use to work with an older kernel or older tg3 driver? If
> > yes, what version?
>
> Works fine with 2.6.16 and earlier.
>
> > Please also provide the full tg3 probing output during modprobe and
> > ifconfig up. Thanks.
>
Looking at the patch history since 2.6.16, the only patch that could
have an impact is the one that enables TSO by default.
Please try turning TSO off to see if it makes a difference:
ethtool -K eth0 tso off
^ permalink raw reply
* Re: tg3 timeouts with 2.6.17-rc6
From: Juergen Kreileder @ 2006-06-17 0:05 UTC (permalink / raw)
To: Michael Chan; +Cc: linux-kernel, netdev
In-Reply-To: <1150494161.26368.39.camel@rh4>
Michael Chan <mchan@broadcom.com> writes:
> On Fri, 2006-06-16 at 23:20 +0200, Juergen Kreileder wrote:
>> Michael Chan <mchan@broadcom.com> writes:
>>
>>>
>>> Did this use to work with an older kernel or older tg3 driver? If
>>> yes, what version?
>>
>> Works fine with 2.6.16 and earlier.
>>
>>> Please also provide the full tg3 probing output during modprobe
>>> and ifconfig up. Thanks.
>>
> Looking at the patch history since 2.6.16, the only patch that could
> have an impact is the one that enables TSO by default.
>
> Please try turning TSO off to see if it makes a difference:
>
> ethtool -K eth0 tso off
Seems to work fine with TSO disabled.
Juergen
--
Juergen Kreileder, Blackdown Java-Linux Team
http://blog.blackdown.de/
^ permalink raw reply
* Re: tg3 timeouts with 2.6.17-rc6
From: Michael Chan @ 2006-06-17 1:27 UTC (permalink / raw)
To: Juergen Kreileder; +Cc: linux-kernel, netdev
In-Reply-To: <87bqsslk9e.fsf@blackdown.de>
On Sat, 2006-06-17 at 02:05 +0200, Juergen Kreileder wrote:
> Michael Chan <mchan@broadcom.com> writes:
> > Please try turning TSO off to see if it makes a difference:
> >
> > ethtool -K eth0 tso off
>
> Seems to work fine with TSO disabled.
>
Thanks for the information. We'll look into it. If possible, please send
me (using private email) an ethereal trace of a sample of the packets
sent by the tg3 chip right before the NETDEV_WATCHDOG timeout.
In the meantime, I wonder if we should disable TSO by default on the
5780 chip for 2.6.17.
^ permalink raw reply
* Re: tg3 timeouts with 2.6.17-rc6
From: David Miller @ 2006-06-17 2:51 UTC (permalink / raw)
To: mchan; +Cc: jk, linux-kernel, netdev
In-Reply-To: <1150507652.26368.46.camel@rh4>
From: "Michael Chan" <mchan@broadcom.com>
Date: Fri, 16 Jun 2006 18:27:32 -0700
> In the meantime, I wonder if we should disable TSO by default on the
> 5780 chip for 2.6.17.
Sounds reasonable. Would we disable it for all chips that set
TG3_FLG2_5780_CLASS or a specific variant?
^ permalink raw reply
* Re: [PATCH 1/6][SCTP]: Limit association max_retrans setting in setsockopt
From: David Miller @ 2006-06-17 2:53 UTC (permalink / raw)
To: sri; +Cc: netdev, lksctp-developers
In-Reply-To: <1150411572.6028.31.camel@w-sridhar2.beaverton.ibm.com>
From: Sridhar Samudrala <sri@us.ibm.com>
Date: Thu, 15 Jun 2006 15:46:12 -0700
> Please apply the following 6 SCTP patches to 2.6 tree.
Thanks Sridhar, I'll try to apply these soon.
^ permalink raw reply
* Re: [ETHTOOL]: Fix UFO typo
From: David Miller @ 2006-06-17 2:53 UTC (permalink / raw)
To: herbert; +Cc: netdev
In-Reply-To: <20060616063905.GA32669@gondor.apana.org.au>
From: Herbert Xu <herbert@gondor.apana.org.au>
Date: Fri, 16 Jun 2006 16:39:05 +1000
> [ETHTOOL]: Fix UFO typo
>
> The function ethtool_get_ufo was referring to ETHTOOL_GTSO instead of
> ETHTOOL_GUFO.
>
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Good catch Herbert, I'll apply this as soon as I get a chance.
^ permalink raw reply
* Re: tg3 timeouts with 2.6.17-rc6
From: Michael Chan @ 2006-06-17 3:37 UTC (permalink / raw)
To: David Miller; +Cc: jk, linux-kernel, netdev
David Miller wrote:
> From: "Michael Chan" <mchan@broadcom.com>
> Date: Fri, 16 Jun 2006 18:27:32 -0700
>
> > In the meantime, I wonder if we should disable TSO by default on the
> > 5780 chip for 2.6.17.
>
> Sounds reasonable. Would we disable it for all chips that set
> TG3_FLG2_5780_CLASS or a specific variant?
>
Yes, let's disable it for all TG3_FLG2_5780_CLASS chips for now
until we figure out what's going on.
^ permalink raw reply
* updated [Patch 1/1] AF_UNIX Datagram getpeersec
From: Catherine Zhang @ 2006-06-17 3:54 UTC (permalink / raw)
To: netdev, davem, jmorris, chrisw, herbert, sds, tjaeger, akpm
Cc: latten, sergeh, gcwilson, czhang.us
Hi,
Enclosed please find the updated AF_UNIX patch, incorporating comments from
James, Stephen, Dave, Chris, Andrew and others.
The patch is now built upon the newly added SELinux functions exported in
selinux/exports.c, which are also used by the auditing subsystem. One
function, selinux_get_socket_sid(), is added to the list of exported
functions by this patch.
I have tested it with three configurations: with SELinux, w/o SELinux,
and allmodconfig.
As always, comments are appreciated!
thanks,
Catherine
---
From: cxzhang@watson.ibm.com
This patch implements an API whereby an application can determine the
label of its peer's Unix datagram sockets via the auxiliary data mechanism of
recvmsg.
Patch purpose:
This patch enables a security-aware application to retrieve the
security context of the peer of a Unix datagram socket. The application
can then use this security context to determine the security context for
processing on behalf of the peer who sent the packet.
Patch design and implementation:
The design and implementation is very similar to the UDP case for INET
sockets. Basically we build upon the existing Unix domain socket API for
retrieving user credentials. Linux offers the API for obtaining user
credentials via ancillary messages (i.e., out of band/control messages
that are bundled together with a normal message). To retrieve the security
context, the application first indicates to the kernel such desire by
setting the SO_PASSSEC option via getsockopt. Then the application
retrieves the security context using the auxiliary data mechanism.
An example server application for Unix datagram socket should look like this:
toggle = 1;
toggle_len = sizeof(toggle);
setsockopt(sockfd, SOL_SOCKET, SO_PASSSEC, &toggle, &toggle_len);
recvmsg(sockfd, &msg_hdr, 0);
if (msg_hdr.msg_controllen > sizeof(struct cmsghdr)) {
cmsg_hdr = CMSG_FIRSTHDR(&msg_hdr);
if (cmsg_hdr->cmsg_len <= CMSG_LEN(sizeof(scontext)) &&
cmsg_hdr->cmsg_level == SOL_SOCKET &&
cmsg_hdr->cmsg_type == SCM_SECURITY) {
memcpy(&scontext, CMSG_DATA(cmsg_hdr), sizeof(scontext));
}
}
sock_setsockopt is enhanced with a new socket option SOCK_PASSSEC to allow
a server socket to receive security context of the peer.
Testing:
We have tested the patch by setting up Unix datagram client and server
applications. We verified that the server can retrieve the security context
using the auxiliary data mechanism of recvmsg.
---
include/asm-alpha/socket.h | 1 +
include/asm-arm/socket.h | 1 +
include/asm-arm26/socket.h | 1 +
include/asm-cris/socket.h | 1 +
include/asm-frv/socket.h | 1 +
include/asm-h8300/socket.h | 1 +
include/asm-i386/socket.h | 1 +
include/asm-ia64/socket.h | 1 +
include/asm-m32r/socket.h | 1 +
include/asm-m68k/socket.h | 1 +
include/asm-mips/socket.h | 1 +
include/asm-parisc/socket.h | 1 +
include/asm-powerpc/socket.h | 1 +
include/asm-s390/socket.h | 1 +
include/asm-sh/socket.h | 1 +
include/asm-sparc/socket.h | 1 +
include/asm-sparc64/socket.h | 1 +
include/asm-v850/socket.h | 1 +
include/asm-x86_64/socket.h | 1 +
include/asm-xtensa/socket.h | 1 +
include/linux/net.h | 1 +
include/net/af_unix.h | 2 ++
include/net/scm.h | 13 +++++++++++++
net/core/sock.c | 11 +++++++++++
net/unix/af_unix.c | 2 ++
security/selinux/exports.c | 11 +++++++++++
security/selinux/include/security.h | 2 +-
security/selinux/ss/services.c | 4 ++--
28 files changed, 63 insertions(+), 3 deletions(-)
diff -puN include/asm-i386/socket.h~lsm-secpeer-unix include/asm-i386/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-i386/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-i386/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-x86_64/socket.h~lsm-secpeer-unix include/asm-x86_64/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-x86_64/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-x86_64/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-ia64/socket.h~lsm-secpeer-unix include/asm-ia64/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-ia64/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-ia64/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -57,5 +57,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_IA64_SOCKET_H */
diff -puN include/asm-powerpc/socket.h~lsm-secpeer-unix include/asm-powerpc/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-powerpc/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-powerpc/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -55,5 +55,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_POWERPC_SOCKET_H */
diff -puN include/asm-h8300/socket.h~lsm-secpeer-unix include/asm-h8300/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-h8300/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-h8300/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-s390/socket.h~lsm-secpeer-unix include/asm-s390/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-s390/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-s390/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -56,5 +56,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-mips/socket.h~lsm-secpeer-unix include/asm-mips/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-mips/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-mips/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -69,6 +69,7 @@ To add: #define SO_REUSEPORT 0x0200 /* A
#define SO_PEERSEC 30
#define SO_SNDBUFFORCE 31
#define SO_RCVBUFFORCE 33
+#define SO_PASSSEC 34
#ifdef __KERNEL__
diff -puN include/asm-alpha/socket.h~lsm-secpeer-unix include/asm-alpha/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-alpha/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-alpha/socket.h 2006-06-16 16:02:55.000000000 -0400
@@ -51,6 +51,7 @@
#define SCM_TIMESTAMP SO_TIMESTAMP
#define SO_PEERSEC 30
+#define SO_PASSSEC 34
/* Security levels - as per NRL IPv6 - don't actually do anything */
#define SO_SECURITY_AUTHENTICATION 19
diff -puN include/asm-v850/socket.h~lsm-secpeer-unix include/asm-v850/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-v850/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-v850/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* __V850_SOCKET_H__ */
diff -puN include/asm-sh/socket.h~lsm-secpeer-unix include/asm-sh/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-sh/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-sh/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* __ASM_SH_SOCKET_H */
diff -puN include/asm-m68k/socket.h~lsm-secpeer-unix include/asm-m68k/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-m68k/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-m68k/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-arm/socket.h~lsm-secpeer-unix include/asm-arm/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-arm/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-arm/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-cris/socket.h~lsm-secpeer-unix include/asm-cris/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-cris/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-cris/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -50,6 +50,7 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-arm26/socket.h~lsm-secpeer-unix include/asm-arm26/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-arm26/socket.h~lsm-secpeer-unix 2006-06-12 17:56:06.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-arm26/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-frv/socket.h~lsm-secpeer-unix include/asm-frv/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-frv/socket.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-frv/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,6 +48,7 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-xtensa/socket.h~lsm-secpeer-unix include/asm-xtensa/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-xtensa/socket.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-xtensa/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -59,5 +59,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _XTENSA_SOCKET_H */
diff -puN include/asm-m32r/socket.h~lsm-secpeer-unix include/asm-m32r/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-m32r/socket.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-m32r/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 30
#define SO_PEERSEC 31
+#define SO_PASSSEC 34
#endif /* _ASM_M32R_SOCKET_H */
diff -puN include/asm-parisc/socket.h~lsm-secpeer-unix include/asm-parisc/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-parisc/socket.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-parisc/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,5 +48,6 @@
#define SO_ACCEPTCONN 0x401c
#define SO_PEERSEC 0x401d
+#define SO_PASSSEC 0x401e
#endif /* _ASM_SOCKET_H */
diff -puN include/asm-sparc/socket.h~lsm-secpeer-unix include/asm-sparc/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-sparc/socket.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-sparc/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,6 +48,7 @@
#define SCM_TIMESTAMP SO_TIMESTAMP
#define SO_PEERSEC 0x001e
+#define SO_PASSSEC 0x001f
/* Security levels - as per NRL IPv6 - don't actually do anything */
#define SO_SECURITY_AUTHENTICATION 0x5001
diff -puN include/asm-sparc64/socket.h~lsm-secpeer-unix include/asm-sparc64/socket.h
--- linux-2.6.17-rc6-mm2/include/asm-sparc64/socket.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/asm-sparc64/socket.h 2006-06-13 15:45:34.000000000 -0400
@@ -48,6 +48,7 @@
#define SCM_TIMESTAMP SO_TIMESTAMP
#define SO_PEERSEC 0x001e
+#define SO_PASSSEC 0x001f
/* Security levels - as per NRL IPv6 - don't actually do anything */
#define SO_SECURITY_AUTHENTICATION 0x5001
diff -puN include/linux/net.h~lsm-secpeer-unix include/linux/net.h
--- linux-2.6.17-rc6-mm2/include/linux/net.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/linux/net.h 2006-06-13 15:45:34.000000000 -0400
@@ -61,6 +61,7 @@ typedef enum {
#define SOCK_ASYNC_WAITDATA 1
#define SOCK_NOSPACE 2
#define SOCK_PASSCRED 3
+#define SOCK_PASSSEC 4
#ifndef ARCH_HAS_SOCKET_TYPES
/**
diff -puN include/net/af_unix.h~lsm-secpeer-unix include/net/af_unix.h
--- linux-2.6.17-rc6-mm2/include/net/af_unix.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/net/af_unix.h 2006-06-13 15:45:34.000000000 -0400
@@ -53,10 +53,12 @@ struct unix_address {
struct unix_skb_parms {
struct ucred creds; /* Skb credentials */
struct scm_fp_list *fp; /* Passed files */
+ u32 sid; /* Security ID */
};
#define UNIXCB(skb) (*(struct unix_skb_parms*)&((skb)->cb))
#define UNIXCREDS(skb) (&UNIXCB((skb)).creds)
+#define UNIXSID(skb) (&UNIXCB((skb)).sid)
#define unix_state_rlock(s) spin_lock(&unix_sk(s)->lock)
#define unix_state_runlock(s) spin_unlock(&unix_sk(s)->lock)
diff -puN include/net/scm.h~lsm-secpeer-unix include/net/scm.h
--- linux-2.6.17-rc6-mm2/include/net/scm.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/include/net/scm.h 2006-06-16 13:58:16.000000000 -0400
@@ -3,6 +3,8 @@
#include <linux/limits.h>
#include <linux/net.h>
+#include <linux/security.h>
+#include <linux/selinux.h>
/* Well, we should have at least one descriptor open
* to accept passed FDs 8)
@@ -19,6 +21,7 @@ struct scm_cookie
{
struct ucred creds; /* Skb credentials */
struct scm_fp_list *fp; /* Passed files */
+ u32 sid; /* Passed security ID */
unsigned long seq; /* Connection seqno */
};
@@ -42,6 +45,7 @@ static __inline__ int scm_send(struct so
scm->creds.gid = p->gid;
scm->creds.pid = p->tgid;
scm->fp = NULL;
+ selinux_get_sock_sid(sock, &scm->sid);
scm->seq = 0;
if (msg->msg_controllen <= 0)
return 0;
@@ -51,6 +55,9 @@ static __inline__ int scm_send(struct so
static __inline__ void scm_recv(struct socket *sock, struct msghdr *msg,
struct scm_cookie *scm, int flags)
{
+ char *scontext;
+ int scontext_len, err;
+
if (!msg->msg_control)
{
if (test_bit(SOCK_PASSCRED, &sock->flags) || scm->fp)
@@ -62,6 +69,12 @@ static __inline__ void scm_recv(struct s
if (test_bit(SOCK_PASSCRED, &sock->flags))
put_cmsg(msg, SOL_SOCKET, SCM_CREDENTIALS, sizeof(scm->creds), &scm->creds);
+ if (test_bit(SOCK_PASSSEC, &sock->flags)) {
+ err = selinux_ctxid_to_string(scm->sid, &scontext, &scontext_len);
+ if (!err)
+ put_cmsg(msg, SOL_SOCKET, SCM_SECURITY, scontext_len, scontext);
+ }
+
if (!scm->fp)
return;
diff -puN net/core/sock.c~lsm-secpeer-unix net/core/sock.c
--- linux-2.6.17-rc6-mm2/net/core/sock.c~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/net/core/sock.c 2006-06-13 15:45:34.000000000 -0400
@@ -565,6 +565,13 @@ set_rcvbuf:
ret = -ENONET;
break;
+ case SO_PASSSEC:
+ if (valbool)
+ set_bit(SOCK_PASSSEC, &sock->flags);
+ else
+ clear_bit(SOCK_PASSSEC, &sock->flags);
+ break;
+
/* We implement the SO_SNDLOWAT etc to
not be settable (1003.1g 5.3) */
default:
@@ -723,6 +730,10 @@ int sock_getsockopt(struct socket *sock,
v.val = sk->sk_state == TCP_LISTEN;
break;
+ case SO_PASSSEC:
+ v.val = test_bit(SOCK_PASSSEC, &sock->flags) ? 1 : 0;
+ break;
+
case SO_PEERSEC:
return security_socket_getpeersec_stream(sock, optval, optlen, len);
diff -puN net/unix/af_unix.c~lsm-secpeer-unix net/unix/af_unix.c
--- linux-2.6.17-rc6-mm2/net/unix/af_unix.c~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/net/unix/af_unix.c 2006-06-13 15:45:34.000000000 -0400
@@ -1290,6 +1290,7 @@ static int unix_dgram_sendmsg(struct kio
memcpy(UNIXCREDS(skb), &siocb->scm->creds, sizeof(struct ucred));
if (siocb->scm->fp)
unix_attach_fds(siocb->scm, skb);
+ memcpy(UNIXSID(skb), &siocb->scm->sid, sizeof(u32));
skb->h.raw = skb->data;
err = memcpy_fromiovec(skb_put(skb,len), msg->msg_iov, len);
@@ -1570,6 +1571,7 @@ static int unix_dgram_recvmsg(struct kio
memset(&tmp_scm, 0, sizeof(tmp_scm));
}
siocb->scm->creds = *UNIXCREDS(skb);
+ siocb->scm->sid = *UNIXSID(skb);
if (!(flags & MSG_PEEK))
{
diff -puN security/selinux/include/security.h~lsm-secpeer-unix security/selinux/include/security.h
--- linux-2.6.17-rc6-mm2/security/selinux/include/security.h~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/security/selinux/include/security.h 2006-06-13 15:45:34.000000000 -0400
@@ -58,7 +58,7 @@ int security_change_sid(u32 ssid, u32 ts
int security_sid_to_context(u32 sid, char **scontext,
u32 *scontext_len);
-int security_context_to_sid(char *scontext, u32 scontext_len,
+int security_context_to_sid(const char *scontext, u32 scontext_len,
u32 *out_sid);
int security_context_to_sid_default(char *scontext, u32 scontext_len, u32 *out_sid, u32 def_sid);
diff -puN security/selinux/ss/services.c~lsm-secpeer-unix security/selinux/ss/services.c
--- linux-2.6.17-rc6-mm2/security/selinux/ss/services.c~lsm-secpeer-unix 2006-06-12 17:56:07.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/security/selinux/ss/services.c 2006-06-13 15:45:34.000000000 -0400
@@ -623,7 +623,7 @@ out:
}
-static int security_context_to_sid_core(char *scontext, u32 scontext_len, u32 *sid, u32 def_sid)
+static int security_context_to_sid_core(const char *scontext, u32 scontext_len, u32 *sid, u32 def_sid)
{
char *scontext2;
struct context context;
@@ -749,7 +749,7 @@ out:
* Returns -%EINVAL if the context is invalid, -%ENOMEM if insufficient
* memory is available, or 0 on success.
*/
-int security_context_to_sid(char *scontext, u32 scontext_len, u32 *sid)
+int security_context_to_sid(const char *scontext, u32 scontext_len, u32 *sid)
{
return security_context_to_sid_core(scontext, scontext_len,
sid, SECSID_NULL);
diff -puN security/selinux/exports.c~lsm-secpeer-unix security/selinux/exports.c
--- linux-2.6.17-rc6-mm2/security/selinux/exports.c~lsm-secpeer-unix 2006-06-16 16:18:48.000000000 -0400
+++ linux-2.6.17-rc6-mm2-cxzhang/security/selinux/exports.c 2006-06-16 16:18:25.000000000 -0400
@@ -17,6 +17,7 @@
#include <linux/selinux.h>
#include <linux/fs.h>
#include <linux/ipc.h>
+#include <net/sock.h>
#include "security.h"
#include "objsec.h"
@@ -72,6 +73,16 @@ void selinux_get_task_sid(struct task_st
*sid = 0;
}
+void selinux_get_sock_sid(const struct socket *sock, u32 *sid)
+{
+ if (selinux_enabled) {
+ const struct inode *inode = SOCK_INODE(sock);
+ selinux_get_inode_sid(inode, sid);
+ return;
+ }
+ *sid = 0;
+}
+
int selinux_string_to_sid(char *str, u32 *sid)
{
if (selinux_enabled)
_
^ permalink raw reply
* Re: [PATCH v2 4/7] AMSO1100 Memory Management.
From: Nick Piggin @ 2006-06-17 3:59 UTC (permalink / raw)
To: Tom Tucker
Cc: Andrew Morton, Steve Wise, rdreier, mshefty, linux-kernel, netdev,
openib-general
In-Reply-To: <1150128349.22704.20.camel@trinity.ogc.int>
Tom Tucker wrote:
> On Thu, 2006-06-08 at 01:17 -0700, Andrew Morton wrote:
>
>>On Wed, 07 Jun 2006 15:06:55 -0500
>>Steve Wise <swise@opengridcomputing.com> wrote:
>>
>>
>>>+void c2_free(struct c2_alloc *alloc, u32 obj)
>>>+{
>>>+ spin_lock(&alloc->lock);
>>>+ clear_bit(obj, alloc->table);
>>>+ spin_unlock(&alloc->lock);
>>>+}
>>
>>The spinlock is unneeded here.
>
>
> Good point.
Really?
clear_bit does not give you any memory ordering, so you can have
the situation where another CPU sees the bit cleared, but this
CPU still has stores pending to whatever it is being freed. Or
any number of other nasty memory ordering badness.
I'd just use the spinlocks, and prepend the clear_bit with a
double underscore (so you get the non-atomic version), if that
is appropriate.
The spinlocks nicely handle all the memory ordering issues, and
serve to document concurrency issues. If you need every last bit
of performance and scalability, that's OK, but you need comments
and I suspect you'd need more memory barriers.
--
SUSE Labs, Novell Inc.
Send instant messages to your online friends http://au.messenger.yahoo.com
^ permalink raw reply
* Re: updated [Patch 1/1] AF_UNIX Datagram getpeersec
From: Stephen Hemminger @ 2006-06-17 4:16 UTC (permalink / raw)
To: Catherine Zhang
Cc: netdev, davem, jmorris, chrisw, herbert, sds, tjaeger, akpm,
latten, sergeh, gcwilson, czhang.us
In-Reply-To: <20060617035440.GA28169@jiayuguan.watson.ibm.com>
O
>
> +void selinux_get_sock_sid(const struct socket *sock, u32 *sid)
> +{
> + if (selinux_enabled) {
> + const struct inode *inode = SOCK_INODE(sock);
> + selinux_get_inode_sid(inode, sid);
> + return;
> + }
> + *sid = 0;
> +}
> +
This is so short, it would make sense to put it in scm.h
and why not have it return the value instead of call by reference?
Same goes for selinux_get_inode_sid
static inline void selinux_get_sock_sid(const struct socket *sock, u32 *sid)
{
if (selinux_enabled)
selinux_get_inode_sid(SOCK_INODE(sock), sid);
else
*sid = 0;
}
or
static inline u32 selinux_get_sock_sid(const struct socket *sock)
{
return selinux_enabled ? selinux_get_inode_sid(SOCK_INODE(sid)) : 0;
}
^ permalink raw reply
* Re: [Ubuntu PATCH] Make tulip driver not handle Davicom NICs
From: Jamie Lenehan @ 2006-06-17 5:43 UTC (permalink / raw)
To: Randy Dunlap; +Cc: netdev, akpm, val_henson
In-Reply-To: <44909A78.4@oracle.com>
On Wed, Jun 14, 2006 at 04:23:36PM -0700, Randy Dunlap wrote:
> Make tulip driver not handle Davicom NICs, let dmfe take over"
>
> Reference: https://launchpad.net/bugs/48287
> Source URL of Patch:
> http://www.kernel.org/git/?p=linux/kernel/git/bcollins/ubuntu-dapper.git;a=commitdiff;h=1804482911a71bee9114cae1c2079507a38e9e7f
The dmfe driver doesn't work on sparc64. My sparc64 Sun Netra X1's have
Davicom NICs and they work fine with the tulip driver. It'd be nice to have
only one driver handling these things that works for both. I'll see if I
fugure out why dmfe doesn't work on sparc64.
aviendha:~# /sbin/modprobe dmfe
dmfe: Davicom DM9xxx net driver, version 1.36.4 (2002-01-17)
eth0: Davicom DM9102 at pci0000:00:05.0, 00:00:00:00:00:00, irq 7098496.
eth1: Davicom DM9102 at pci0000:00:0c.0, 00:00:00:00:00:00, irq 7097792.
aviendha:~# /sbin/ifup eth0
SABRE0: PCI SERR signal asserted.
SABRE0: PCI bus error, PCI_STATUS[caa0]
eth0: Tx timeout - resetting
SABRE0: PCI SERR signal asserted.
SABRE0: PCI bus error, PCI_STATUS[caa0]
eth0: Tx timeout - resetting
SABRE0: PCI SERR signal asserted.
SABRE0: PCI bus error, PCI_STATUS[caa0]
eth0: Tx timeout - resetting
aviendha:~# /sbin/ifdown eth0
aviendha:~# /sbin/rmmod dmfe
aviendha:~# /sbin/modprobe tulip
Linux Tulip driver version 1.1.13 (May 11, 2002)
tulip0: Old style EEPROM with no media selection information.
tulip0: MII transceiver #1 config 3100 status 7829 advertising 01e1.
eth0: Davicom DM9102/DM9102A rev 49 at 000001fe02010100, EEPROM not present, 00:03:BA:06:30:E1, IRQ 7098496.
tulip1: Old style EEPROM with no media selection information.
tulip1: MII transceiver #1 config 1000 status 7809 advertising 01e1.
eth1: Davicom DM9102/DM9102A rev 49 at 000001fe02010000, EEPROM not present, 00:03:BA:06:30:E0, IRQ 7097792.
aviendha:~# /sbin/ifup eth0
aviendha:~# ping -c 1 172.20.0.1
PING 172.20.0.1 (172.20.0.1) 56(84) bytes of data.
64 bytes from 172.20.0.1: icmp_seq=1 ttl=64 time=3.02 ms
--- 172.20.0.1 ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 3.021/3.021/3.021/0.000 ms
aviendha:~# uname -a
Linux aviendha.twibble.org 2.6.15-rc4 #12 Sat Dec 3 21:44:21 EST 2005 sparc64 GNU/Linux
aviendha:~# lspci -vxns 00:0c.0
0000:00:0c.0 0200: 1282:9102 (rev 31)
Flags: bus master, medium devsel, latency 160, IRQ 006c4dc0
I/O ports at 0000000002010000 [size=256]
[virtual] Memory at 000001ff00000000 (32-bit, non-prefetchable) [size=256]
Expansion ROM at 0000000000040000 [disabled] [size=256K]
Capabilities: [50] Power Management version 2
00: 82 12 02 91 47 01 10 02 31 00 00 02 00 a0 00 00
10: 01 00 01 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
30: 00 00 04 00 50 00 00 00 00 00 00 00 00 01 14 28
aviendha:~#
...
--
Jamie Lenehan <lenehan@twibble.org>
^ permalink raw reply
* Re: updated [Patch 1/1] AF_UNIX Datagram getpeersec
From: Xiaolan Zhang @ 2006-06-17 6:47 UTC (permalink / raw)
To: Stephen Hemminger
Cc: akpm, chrisw, Catherine Zhang, czhang.us, davem, George Wilson,
herbert, jmorris, latten, netdev, sds, Serge E Hallyn, tjaeger
In-Reply-To: <20060616211603.14547a1d@localhost.localdomain>
Hi, Stephen,
It appears that selinux_enabled is defined inside selinux module and not
visible to the rest of the kernel...
thanks,
Catherine
Stephen Hemminger <shemminger@osdl.org> wrote on 06/17/2006 12:16:03 AM:
> O
> >
> > +void selinux_get_sock_sid(const struct socket *sock, u32 *sid)
> > +{
> > + if (selinux_enabled) {
> > + const struct inode *inode = SOCK_INODE(sock);
> > + selinux_get_inode_sid(inode, sid);
> > + return;
> > + }
> > + *sid = 0;
> > +}
> > +
>
> This is so short, it would make sense to put it in scm.h
> and why not have it return the value instead of call by reference?
> Same goes for selinux_get_inode_sid
>
> static inline void selinux_get_sock_sid(const struct socket *sock, u32
*sid)
> {
> if (selinux_enabled)
> selinux_get_inode_sid(SOCK_INODE(sock), sid);
> else
> *sid = 0;
> }
>
> or
> static inline u32 selinux_get_sock_sid(const struct socket *sock)
> {
> return selinux_enabled ? selinux_get_inode_sid(SOCK_INODE(sid)) : 0;
> }
^ permalink raw reply
* New Now you have chance to do it Relax with
From: Patricia @ 2006-06-17 7:37 UTC (permalink / raw)
To: newbie
Hello,
Most quality products for anyone who wants to become a champion in bed
Order now and benefit from lowest costs and convenient shipment
World famous brands which keep men happy all over the world
Enter here: http://www.unpraiselk.com
We thank you for your interest!
^ permalink raw reply
* Re: Linux 2.6 Vs linux 2.4 routing
From: Krzysztof Oledzki @ 2006-06-17 10:33 UTC (permalink / raw)
To: Mayuresh Chitale; +Cc: netdev, linux-net
In-Reply-To: <d096a3ee0606161141n3c481ed9kfcbf70af7393affa@mail.gmail.com>
[-- Attachment #1: Type: TEXT/PLAIN, Size: 955 bytes --]
On Sat, 17 Jun 2006, Mayuresh Chitale wrote:
> Hi All,
Hi,
> I am doing a routing test using linux 2.4 and linux 2.6 kernel on our
> board. I am using 2.4.31 and 2.6.16.17 for comparison.
>
> I have ported the bsp and drivers from 2.4 to 2.6 and am using a
> smartbits tester which pumps traffic at high rates and at varying
> packet sizes.
>
> On 2.4, the test goes fine ie I get meaningful values. But on 2.6
> kernel, routing begins and continues for sometime before it completely
> stops. The smartbits configuration is identical in both the cases.
>
> I have noticed same problem with dlink and 3com cards which is causing
> me to suspect that there is some probem with 2.6 routing.
>
> Has anyone seen a similar problem ?
> Any inputs on this would be appreciated.
Did you enable IRQBALANCE in 2.6?
$ zcat /proc/config.gz |grep IRQBALANCE
If so, please try to disable it.
Best regards,
Krzysztof Olędzki
^ permalink raw reply
* [PATCH 2.6.17 fix] bcm43xx: workaround init_board vs. IRQ race
From: Michael Buesch @ 2006-06-17 13:19 UTC (permalink / raw)
To: linville; +Cc: bcm43xx-dev, netdev, linux-kernel
Hi John,
Please try to push this for 2.6.17, as it is a crash fix.
--
Place the Init-vs-IRQ workaround before any card register
access, because we might not have the wireless core mapped
at all times in init. So this will result in a Machine Check
caused by a bus error.
Signed-off-by: Michael Buesch <mb@bu3sch.de>
Index: wireless-2.6/drivers/net/wireless/bcm43xx/bcm43xx_main.c
===================================================================
--- wireless-2.6.orig/drivers/net/wireless/bcm43xx/bcm43xx_main.c 2006-06-17 15:06:38.000000000 +0200
+++ wireless-2.6/drivers/net/wireless/bcm43xx/bcm43xx_main.c 2006-06-17 15:17:49.000000000 +0200
@@ -1885,6 +1885,15 @@
spin_lock(&bcm->irq_lock);
+ /* Only accept IRQs, if we are initialized properly.
+ * This avoids an RX race while initializing.
+ * We should probably not enable IRQs before we are initialized
+ * completely, but some careful work is needed to fix this. I think it
+ * is best to stay with this cheap workaround for now... .
+ */
+ if (unlikely(bcm43xx_status(bcm) != BCM43xx_STAT_INITIALIZED))
+ goto out;
+
reason = bcm43xx_read32(bcm, BCM43xx_MMIO_GEN_IRQ_REASON);
if (reason == 0xffffffff) {
/* irq not for us (shared irq) */
@@ -1906,19 +1915,11 @@
bcm43xx_interrupt_ack(bcm, reason);
- /* Only accept IRQs, if we are initialized properly.
- * This avoids an RX race while initializing.
- * We should probably not enable IRQs before we are initialized
- * completely, but some careful work is needed to fix this. I think it
- * is best to stay with this cheap workaround for now... .
- */
- if (likely(bcm43xx_status(bcm) == BCM43xx_STAT_INITIALIZED)) {
- /* disable all IRQs. They are enabled again in the bottom half. */
- bcm->irq_savedstate = bcm43xx_interrupt_disable(bcm, BCM43xx_IRQ_ALL);
- /* save the reason code and call our bottom half. */
- bcm->irq_reason = reason;
- tasklet_schedule(&bcm->isr_tasklet);
- }
+ /* disable all IRQs. They are enabled again in the bottom half. */
+ bcm->irq_savedstate = bcm43xx_interrupt_disable(bcm, BCM43xx_IRQ_ALL);
+ /* save the reason code and call our bottom half. */
+ bcm->irq_reason = reason;
+ tasklet_schedule(&bcm->isr_tasklet);
out:
mmiowb();
^ permalink raw reply
* Fwd: Your letter
From: Benny Higgins @ 2006-06-17 15:45 UTC (permalink / raw)
To: netdev
-S''ensationall r'evoolution in m''edicine!
-E'n'l'a'r'g'e your p''e'n'i's up to 10 cm or up to 4 i'nches!
-It's h'e'r'b'a'l solution what hasn't side e'ffect, but has 100% g'uaranted results!
-Don't lose your chance and but know wihtout d'oubts, you will be i'mpressed with results!
Clisk here: http://posmonsters.info
liturgy simon funeral chaparral slug cutlet jot strike alcoholism ritual
syntax aerobacter electorate error cruel airway codify micronesia shire
daly bristle condemnatory betrothal dyke fibrin kleenex condensate paginate ambiguity phone
liquid nicotinamide circulate the curvature heckle squawbush enunciable
cowherd copernicus axle numinous run copy moist paradigm
shimmy copter barnett wolcott drier mode scandinavia caliper
^ permalink raw reply
* Re: [RFC PATCH 1/2] Hardware button support for Wireless cards: radiobtn
From: Ivo van Doorn @ 2006-06-17 15:05 UTC (permalink / raw)
To: Stefan Rompf; +Cc: Francois Romieu, netdev, rt2400-devel
In-Reply-To: <200606041214.45999.stefan@loplof.de>
[-- Attachment #1: Type: text/plain, Size: 14016 bytes --]
> > Except for the bluetooth radio key (which should be supported by the
> > radiobtn interface as well) the other buttons have support through already
> > excisting input devices if I am correct.
>
> You are wrong for quite a bunch of laptop models. That's why I pointed you to
> the wistron_btns driver. Alternatively, look at the acerhk driver
> (http://www2.informatik.hu-berlin.de/~tauber/acerhk/). Many systems have a
> number of additional buttons that need to be handled by a special driver, all
> sent to userspace, and just one of them to trigger the wireless card. Other
> models just handle the button in ACPI and generate an additional ACPI event.
>
> Looking at the RT2400-driver, I see what you want to accomplish: Take the view
> of the WLAN card on the hardware controlled button enable/disable and
> generate events on it. However, in many cases it is another driver (see
> above) that sets or clears this state, and this should be the instance to
> send the input event.
>
> Note that I do not have objections against the driver being included in the
> kernel - it just does not qualify as generic radiobutton support, but I know
> it's hard to find a good name ;-)
>
> Looking at the code only: There should be an additional non-polling interface
> for drivers that can generate events on the own.
Sorry for the late reply, but I have been quite busy with other things lately.
I have created an updated patch, that is more generic than the previous.
I have also renamed it to "laptopkey".
With this approach more buttons can be registered,
it includes the optional field to report an update of the key status
to the driver that registered it, and it supports for non-polling keys.
By default all events will go over the default input device, but a driver
can choose to create additional input devices.
Signed-off-by Ivo van Doorn <IvDoorn@gmail.com>
---
diff --git a/drivers/input/misc/Kconfig b/drivers/input/misc/Kconfig
index 4bad588..097ae8c 100644
--- a/drivers/input/misc/Kconfig
+++ b/drivers/input/misc/Kconfig
@@ -79,4 +79,12 @@ config HP_SDC_RTC
Say Y here if you want to support the built-in real time clock
of the HP SDC controller.
+config LAPTOPKEY
+ tristate "Laptop keys support"
+ help
+ Say Y here if you have laptop with additional keys like hardware
+ wireless or bluetooth radio control, or other (programmable) keys.
+ This driver will create an input device to which other hardware
+ drivers can register their keys to.
+
endif
diff --git a/drivers/input/misc/Makefile b/drivers/input/misc/Makefile
index 415c491..e691fab 100644
--- a/drivers/input/misc/Makefile
+++ b/drivers/input/misc/Makefile
@@ -11,3 +11,4 @@ obj-$(CONFIG_INPUT_UINPUT) += uinput.o
obj-$(CONFIG_INPUT_WISTRON_BTNS) += wistron_btns.o
obj-$(CONFIG_HP_SDC_RTC) += hp_sdc_rtc.o
obj-$(CONFIG_INPUT_IXP4XX_BEEPER) += ixp4xx-beeper.o
+obj-$(CONFIG_LAPTOPKEY) += laptopkey.o
diff --git a/drivers/input/misc/laptopkey.c b/drivers/input/misc/laptopkey.c
new file mode 100644
index 0000000..7731554
--- /dev/null
+++ b/drivers/input/misc/laptopkey.c
@@ -0,0 +1,231 @@
+/*
+ Copyright (C) 2006 Ivo van Doorn
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the
+ Free Software Foundation, Inc.,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/laptopkey.h>
+
+MODULE_AUTHOR("Ivo van Doorn <IvDoorn@gmail.com>");
+MODULE_VERSION("1.0");
+MODULE_DESCRIPTION("Laptop keys support");
+MODULE_LICENSE("GPL");
+
+/*
+ * Default laptop_key_device structure.
+ */
+static struct laptop_key_device default_keydev = {
+ .dev_name = "laptopbtn",
+ .delay = 100,
+};
+
+void laptopkey_key_event(struct laptop_key *key, int status)
+{
+ /*
+ * Check if the hardware needs to be informed of
+ * the status change.
+ */
+ if (key->key_event)
+ key->key_event(key->data, status);
+
+ /*
+ * Send input event to userspace.
+ */
+ input_report_key(key->keydev->input_dev, key->key, status);
+ input_sync(key->keydev->input_dev);
+
+}
+
+static void laptopkey_poll(unsigned long data)
+{
+ struct laptop_key_device *keydev = (struct laptop_key_device *)data;
+ struct laptop_key *key;
+ struct list_head *entry;
+ int status;
+
+ list_for_each(entry, &keydev->list){
+ key = list_entry(entry, struct laptop_key, entry);
+
+ /*
+ * Check if this key requires polling.
+ * Otherwise just skip this key,
+ * and move to the next one.
+ */
+ if (!key->key_poll)
+ continue;
+
+ /*
+ * Poll the status, and report a key event
+ * when the status has changed.
+ */
+ status = !!key->key_poll(key->data);
+ if (status != key->current_status) {
+ laptopkey_key_event(key, status);
+ key->current_status = status;
+ }
+ }
+
+ /*
+ * Check if the polling can continue.
+ */
+ if (keydev->delay == 0)
+ return;
+
+ mod_timer(&keydev->timer, jiffies + msecs_to_jiffies(keydev->delay));
+}
+
+void laptopkey_dev_add_key(struct laptop_key_device *keydev,
+ struct laptop_key *key)
+{
+ /*
+ * If polling is enabled, obtain the initial
+ * status from the polling functions, otherwise
+ * set it to 0.
+ */
+ if (key->key_poll)
+ key->current_status = key->key_poll(key->data);
+ else
+ key->current_status = 0;
+
+ /*
+ * Add key to key device.
+ */
+ key->keydev = keydev;
+ list_add(&key->entry, &keydev->list);
+
+ /*
+ * Add key to the input device.
+ */
+ set_bit(key->key, keydev->input_dev->keybit);
+}
+
+void laptopkey_dev_del_key(struct laptop_key_device *keydev,
+ struct laptop_key *key)
+{
+ /*
+ * Remove list from device.
+ */
+ list_del(&key->entry);
+
+ /*
+ * Remove key to the input device.
+ */
+ clear_bit(key->key, keydev->input_dev->keybit);
+}
+
+void laptopkey_add_key(struct laptop_key *key)
+{
+ laptopkey_dev_add_key(&default_keydev, key);
+}
+
+void laptopkey_del_key(struct laptop_key *key)
+{
+ laptopkey_dev_del_key(&default_keydev, key);
+}
+
+int laptopkey_register_device(struct laptop_key_device *keydev)
+{
+ int status;
+
+ /*
+ * Create the input device
+ */
+ keydev->input_dev = input_allocate_device();
+ if (!keydev->input_dev) {
+ printk(KERN_ERR "Failed to allocate input device %s.\n",
+ keydev->dev_name);
+ return -ENOMEM;
+ }
+
+ /*
+ * Initialize the input device
+ */
+ keydev->input_dev->name = "Laptop keys";
+ keydev->input_dev->phys = keydev->dev_name;
+ keydev->input_dev->id.bustype = BUS_HOST;
+
+ /*
+ * Register the input device
+ */
+ status = input_register_device(keydev->input_dev);
+ if (status) {
+ printk(KERN_ERR "Failed to register input device %s.\n",
+ keydev->dev_name);
+ input_free_device(keydev->input_dev);
+ return status;
+ }
+
+ /*
+ * Initialize list head.
+ */
+ INIT_LIST_HEAD(&keydev->list);
+
+ /*
+ * Initialize timer.
+ */
+ init_timer(&keydev->timer);
+ keydev->timer.function = laptopkey_poll;
+ keydev->timer.data = (unsigned long)keydev;
+ keydev->timer.expires = jiffies + msecs_to_jiffies(keydev->delay);
+ add_timer(&keydev->timer);
+
+ printk(KERN_INFO "Created new %s: %s.\n",
+ keydev->input_dev->name, keydev->input_dev->phys);
+
+ return 0;
+}
+
+void laptopkey_unregister_device(struct laptop_key_device *keydev)
+{
+ /*
+ * Clear the delay field before deleting the timer,
+ * to prevent the scheduled work to rearm itself.
+ */
+ keydev->delay = 0;
+ del_timer_sync(&keydev->timer);
+
+ /*
+ * Remove input device.
+ */
+ input_unregister_device(keydev->input_dev);
+ input_free_device(keydev->input_dev);
+}
+
+static int __init radiobtn_init(void)
+{
+ printk(KERN_INFO "Loading radio button driver.\n");
+ return laptopkey_register_device(&default_keydev);
+}
+
+static void __exit radiobtn_exit(void)
+{
+ laptopkey_unregister_device(&default_keydev);
+ printk(KERN_INFO "Unloading radio button driver.\n");
+}
+
+EXPORT_SYMBOL(laptopkey_add_key);
+EXPORT_SYMBOL(laptopkey_del_key);
+EXPORT_SYMBOL(laptopkey_register_device);
+EXPORT_SYMBOL(laptopkey_unregister_device);
+EXPORT_SYMBOL(laptopkey_dev_add_key);
+EXPORT_SYMBOL(laptopkey_dev_del_key);
+EXPORT_SYMBOL(laptopkey_key_event);
+
+module_init(radiobtn_init);
+module_exit(radiobtn_exit);
diff --git a/include/linux/laptopkey.h b/include/linux/laptopkey.h
new file mode 100644
index 0000000..063ecf5
--- /dev/null
+++ b/include/linux/laptopkey.h
@@ -0,0 +1,131 @@
+/*
+ Copyright (C) 2006 Ivo van Doorn
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the
+ Free Software Foundation, Inc.,
+ 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ */
+
+/*
+ Laptop keys support
+
+ INTRODUCTION
+ Laptops are quite often equiped with additional hardware keys.
+ This includes the radio keys for wireless and bluetooth devices,
+ but also other (programmable) keys.
+ This driver will create a generic layer between hardware and
+ userspace. Drivers can add keys to the input device for
+ frequent polling and reporting the status to userspace.
+
+ INPUT DEVICES
+ When loading this driver the main input device is directly created.
+ When desired drivers can also register a new laptop_key_device
+ which creates a new input device to which the laptop keys can be
+ registered.
+
+ POLLING
+ In the laptop_key_device structure the delay is set, this delay
+ is used to determine the polling interval in which each key will be
+ polled for its status. When the key_poll field is not set, the
+ key will not be polled.
+
+ STATUS REPORTING
+ When the status has changed, an input event will be send to userspace.
+ The key field will be used as reported key in that input event.
+ When the status of the key has changed, this event will also be send to
+ the driver that registered the key when the key_event field has been
+ set. It is also possible for keys to skip the polling and inform
+ this driver of the event directly.
+ */
+
+#ifndef LAPTOPKEY_H
+#define LAPTOPKEY_H
+
+#include <linux/input.h>
+#include <linux/list.h>
+
+/**
+ * struct laptop_key_device - laptop key device structure
+ * @dev_name: Name of the input device.
+ * @delay: When polling is required for any of the keys,
+ * this polling delay (in ms) will be used.
+ * @list: All laptop keys are part of a linked list.
+ * This field should not be touched by any driver.
+ * @input_dev: This is the input device created for this laptop_key_device.
+ * This field should not be touched by any driver.
+ * @timer: The timer structure for the frequent polling of the keys.
+ * This field should not be touched by any driver.
+ */
+struct laptop_key_device {
+ const char *dev_name;
+
+ unsigned int delay;
+
+ struct list_head list;
+
+ struct input_dev *input_dev;
+
+ struct timer_list timer;
+};
+
+/**
+ * struct laptop_key - laptop key structure
+ * @key: Key type that is send to userspace.
+ * should be any of the KEY_* defines from <linux/input.h>
+ * @data: private data that is send as argument with the
+ * functions key_poll and key_event.
+ * @current_status: Current status of the button is stored here.
+ * @key_poll: This function is used to determine the current state
+ * of the key. This field only needs to be set when polling is required.
+ * @key_event: This function is used when the key status has changed, it will
+ * then send the new key status as argument to inform the hardware
+ * of the change. This field is only required when the hardware needs to
+ * be informed of the change.
+ * @keydev: laptop_key_device structure this key is attached to.
+ * This field should not be touched by any driver.
+ * @entry: All laptop keys are part of a linked list. This field should not
+ * be touched by any driver.
+ */
+struct laptop_key {
+ unsigned int key;
+ unsigned long data;
+ int current_status;
+
+ int (*key_poll)(unsigned long data);
+ void (*key_event)(unsigned long data, int status);
+
+ struct laptop_key_device *keydev;
+ struct list_head entry;
+};
+
+/*
+ * Functions when using the main input device.
+ */
+void laptopkey_add_key(struct laptop_key *);
+void laptopkey_del_key(struct laptop_key *);
+
+/*
+ * Functions when using a seperate input device.
+ */
+int laptopkey_register_device(struct laptop_key_device *);
+void laptopkey_unregister_device(struct laptop_key_device *);
+void laptopkey_dev_add_key(struct laptop_key_device *, struct laptop_key *);
+void laptopkey_dev_del_key(struct laptop_key_device *, struct laptop_key *);
+
+/*
+ * Manually report a key event.
+ */
+void laptopkey_key_event(struct laptop_key *, int status);
+
+#endif /* LAPTOPKEY_H */
[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]
^ permalink raw reply related
* Re: updated [Patch 1/1] AF_UNIX Datagram getpeersec
From: James Morris @ 2006-06-17 15:46 UTC (permalink / raw)
To: Xiaolan Zhang
Cc: Stephen Hemminger, akpm, chrisw, Catherine Zhang, czhang.us,
davem, George Wilson, herbert, jmorris, latten, netdev, sds,
Serge E Hallyn, tjaeger
In-Reply-To: <OF42C27083.DD6A539B-ON85257190.00252DFB-85257190.0025456C@us.ibm.com>
On Sat, 17 Jun 2006, Xiaolan Zhang wrote:
> Hi, Stephen,
>
> It appears that selinux_enabled is defined inside selinux module and not
> visible to the rest of the kernel...
Yes, these are SELinux API functions, they belong in
security/selinux/exports.c
But what about Stephens question on the return value?
>
> thanks,
> Catherine
>
> Stephen Hemminger <shemminger@osdl.org> wrote on 06/17/2006 12:16:03 AM:
>
> > O
> > >
> > > +void selinux_get_sock_sid(const struct socket *sock, u32 *sid)
> > > +{
> > > + if (selinux_enabled) {
> > > + const struct inode *inode = SOCK_INODE(sock);
> > > + selinux_get_inode_sid(inode, sid);
> > > + return;
> > > + }
> > > + *sid = 0;
> > > +}
> > > +
> >
> > This is so short, it would make sense to put it in scm.h
> > and why not have it return the value instead of call by reference?
> > Same goes for selinux_get_inode_sid
> >
> > static inline void selinux_get_sock_sid(const struct socket *sock, u32
> *sid)
> > {
> > if (selinux_enabled)
> > selinux_get_inode_sid(SOCK_INODE(sock), sid);
> > else
> > *sid = 0;
> > }
> >
> > or
> > static inline u32 selinux_get_sock_sid(const struct socket *sock)
> > {
> > return selinux_enabled ? selinux_get_inode_sid(SOCK_INODE(sid)) : 0;
> > }
>
> -
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
--
James Morris
<jmorris@namei.org>
^ permalink raw reply
* Re: updated [Patch 1/1] AF_UNIX Datagram getpeersec
From: James Morris @ 2006-06-17 15:55 UTC (permalink / raw)
To: Stephen Hemminger
Cc: Catherine Zhang, netdev, davem, jmorris, chrisw, herbert, sds,
tjaeger, akpm, latten, sergeh, gcwilson, czhang.us
In-Reply-To: <20060616211603.14547a1d@localhost.localdomain>
On Fri, 16 Jun 2006, Stephen Hemminger wrote:
> This is so short, it would make sense to put it in scm.h
> and why not have it return the value instead of call by reference?
> Same goes for selinux_get_inode_sid
Actually, all of the SELinux API functions are like this, and I'm not sure
why for void return methods (it's a good idea when they return errnos).
Once this area has settled down (post 2.6.18), I'm planning to do some
cleanups for this API anyway, and can fix these all at the same time.
- James
--
James Morris
<jmorris@namei.org>
^ permalink raw reply
* [PATCH, RFT] bcm43xx: Busting the 1G limit
From: Michael Buesch @ 2006-06-17 17:28 UTC (permalink / raw)
To: bcm43xx-dev, netdev
Hi,
This patch adds full 32-bit and 64-bit DMA support
to the bcm43xx driver. Well, it _should_ do this. I can
not test it, as I don't have a machine to trigger the 1G
limit.
The 1G limit should be exploitable on an AMD64 machine
with more than 1G RAM.
Please test and report, if it works or not. In the
case of "works not", please provide full dmesg log.
Note that I am not sure which cards actually support
full 32-bit or even 64-bit mode. Older cards might still
only support 30-bit DMA.
Index: wireless-dev/drivers/net/wireless/bcm43xx/bcm43xx.h
===================================================================
--- wireless-dev.orig/drivers/net/wireless/bcm43xx/bcm43xx.h 2006-06-17 17:48:26.000000000 +0200
+++ wireless-dev/drivers/net/wireless/bcm43xx/bcm43xx.h 2006-06-17 18:01:34.000000000 +0200
@@ -32,14 +32,18 @@
#define BCM43xx_PCICFG_ICR 0x94
/* MMIO offsets */
-#define BCM43xx_MMIO_DMA1_REASON 0x20
-#define BCM43xx_MMIO_DMA1_IRQ_MASK 0x24
-#define BCM43xx_MMIO_DMA2_REASON 0x28
-#define BCM43xx_MMIO_DMA2_IRQ_MASK 0x2C
-#define BCM43xx_MMIO_DMA3_REASON 0x30
-#define BCM43xx_MMIO_DMA3_IRQ_MASK 0x34
-#define BCM43xx_MMIO_DMA4_REASON 0x38
-#define BCM43xx_MMIO_DMA4_IRQ_MASK 0x3C
+#define BCM43xx_MMIO_DMA0_REASON 0x20
+#define BCM43xx_MMIO_DMA0_IRQ_MASK 0x24
+#define BCM43xx_MMIO_DMA1_REASON 0x28
+#define BCM43xx_MMIO_DMA1_IRQ_MASK 0x2C
+#define BCM43xx_MMIO_DMA2_REASON 0x30
+#define BCM43xx_MMIO_DMA2_IRQ_MASK 0x34
+#define BCM43xx_MMIO_DMA3_REASON 0x38
+#define BCM43xx_MMIO_DMA3_IRQ_MASK 0x3C
+#define BCM43xx_MMIO_DMA4_REASON 0x40
+#define BCM43xx_MMIO_DMA4_IRQ_MASK 0x44
+#define BCM43xx_MMIO_DMA5_REASON 0x48
+#define BCM43xx_MMIO_DMA5_IRQ_MASK 0x4C
#define BCM43xx_MMIO_STATUS_BITFIELD 0x120
#define BCM43xx_MMIO_STATUS2_BITFIELD 0x124
#define BCM43xx_MMIO_GEN_IRQ_REASON 0x128
@@ -55,14 +59,27 @@
#define BCM43xx_MMIO_XMITSTAT_1 0x174
#define BCM43xx_MMIO_REV3PLUS_TSF_LOW 0x180 /* core rev >= 3 only */
#define BCM43xx_MMIO_REV3PLUS_TSF_HIGH 0x184 /* core rev >= 3 only */
-#define BCM43xx_MMIO_DMA1_BASE 0x200
-#define BCM43xx_MMIO_DMA2_BASE 0x220
-#define BCM43xx_MMIO_DMA3_BASE 0x240
-#define BCM43xx_MMIO_DMA4_BASE 0x260
+
+/* 32-bit DMA */
+#define BCM43xx_MMIO_DMA32_BASE0 0x200
+#define BCM43xx_MMIO_DMA32_BASE1 0x220
+#define BCM43xx_MMIO_DMA32_BASE2 0x240
+#define BCM43xx_MMIO_DMA32_BASE3 0x260
+#define BCM43xx_MMIO_DMA32_BASE4 0x280
+#define BCM43xx_MMIO_DMA32_BASE5 0x2A0
+/* 64-bit DMA */
+#define BCM43xx_MMIO_DMA64_BASE0 0x200
+#define BCM43xx_MMIO_DMA64_BASE1 0x240
+#define BCM43xx_MMIO_DMA64_BASE2 0x280
+#define BCM43xx_MMIO_DMA64_BASE3 0x2C0
+#define BCM43xx_MMIO_DMA64_BASE4 0x300
+#define BCM43xx_MMIO_DMA64_BASE5 0x340
+/* PIO */
#define BCM43xx_MMIO_PIO1_BASE 0x300
#define BCM43xx_MMIO_PIO2_BASE 0x310
#define BCM43xx_MMIO_PIO3_BASE 0x320
#define BCM43xx_MMIO_PIO4_BASE 0x330
+
#define BCM43xx_MMIO_PHY_VER 0x3E0
#define BCM43xx_MMIO_PHY_RADIO 0x3E2
#define BCM43xx_MMIO_ANTENNA 0x3E8
@@ -231,8 +248,14 @@
#define BCM43xx_SBTMSTATELOW_FORCE_GATE_CLOCK 0x20000
/* sbtmstatehigh state flags */
-#define BCM43xx_SBTMSTATEHIGH_SERROR 0x1
-#define BCM43xx_SBTMSTATEHIGH_BUSY 0x4
+#define BCM43xx_SBTMSTATEHIGH_SERROR 0x00000001
+#define BCM43xx_SBTMSTATEHIGH_BUSY 0x00000004
+#define BCM43xx_SBTMSTATEHIGH_TIMEOUT 0x00000020
+#define BCM43xx_SBTMSTATEHIGH_COREFLAGS 0x1FFF0000
+#define BCM43xx_SBTMSTATEHIGH_DMA64BIT 0x10000000
+#define BCM43xx_SBTMSTATEHIGH_GATEDCLK 0x20000000
+#define BCM43xx_SBTMSTATEHIGH_BISTFAILED 0x40000000
+#define BCM43xx_SBTMSTATEHIGH_BISTCOMPLETE 0x80000000
/* sbimstate flags */
#define BCM43xx_SBIMSTATE_IB_ERROR 0x20000
@@ -566,8 +589,11 @@
struct bcm43xx_dmaring *tx_ring1;
struct bcm43xx_dmaring *tx_ring2;
struct bcm43xx_dmaring *tx_ring3;
+ struct bcm43xx_dmaring *tx_ring4;
+ struct bcm43xx_dmaring *tx_ring5;
+
struct bcm43xx_dmaring *rx_ring0;
- struct bcm43xx_dmaring *rx_ring1; /* only available on core.rev < 5 */
+ struct bcm43xx_dmaring *rx_ring3; /* only available on core.rev < 5 */
};
/* Data structures for PIO transmission, per 80211 core. */
@@ -708,7 +734,7 @@
/* Reason code of the last interrupt. */
u32 irq_reason;
- u32 dma_reason[4];
+ u32 dma_reason[6];
/* saved irq enable/disable state bitfield. */
u32 irq_savedstate;
/* Link Quality calculation context. */
Index: wireless-dev/drivers/net/wireless/bcm43xx/bcm43xx_dma.c
===================================================================
--- wireless-dev.orig/drivers/net/wireless/bcm43xx/bcm43xx_dma.c 2006-06-17 17:48:26.000000000 +0200
+++ wireless-dev/drivers/net/wireless/bcm43xx/bcm43xx_dma.c 2006-06-17 19:18:41.000000000 +0200
@@ -4,7 +4,7 @@
DMA ringbuffer and descriptor allocation/management
- Copyright (c) 2005 Michael Buesch <mbuesch@freenet.de>
+ Copyright (c) 2005, 2006 Michael Buesch <mbuesch@freenet.de>
Some code in this file is derived from the b44.c driver
Copyright (C) 2002 David S. Miller
@@ -109,6 +109,35 @@
}
}
+u16 bcm43xx_dmacontroller_base(int dma64bit, int controller_idx)
+{
+ static const u16 map64[] = {
+ BCM43xx_MMIO_DMA64_BASE0,
+ BCM43xx_MMIO_DMA64_BASE1,
+ BCM43xx_MMIO_DMA64_BASE2,
+ BCM43xx_MMIO_DMA64_BASE3,
+ BCM43xx_MMIO_DMA64_BASE4,
+ BCM43xx_MMIO_DMA64_BASE5,
+ };
+ static const u16 map32[] = {
+ BCM43xx_MMIO_DMA32_BASE0,
+ BCM43xx_MMIO_DMA32_BASE1,
+ BCM43xx_MMIO_DMA32_BASE2,
+ BCM43xx_MMIO_DMA32_BASE3,
+ BCM43xx_MMIO_DMA32_BASE4,
+ BCM43xx_MMIO_DMA32_BASE5,
+ };
+
+ if (dma64bit) {
+ assert(controller_idx >= 0 ||
+ controller_idx < ARRAY_SIZE(map64));
+ return map64[controller_idx];
+ }
+ assert(controller_idx >= 0 ||
+ controller_idx < ARRAY_SIZE(map32));
+ return map32[controller_idx];
+}
+
static inline
dma_addr_t map_descbuffer(struct bcm43xx_dmaring *ring,
unsigned char *buf,
@@ -172,7 +201,6 @@
/* Unmap and free a descriptor buffer. */
static inline
void free_descriptor_buffer(struct bcm43xx_dmaring *ring,
- struct bcm43xx_dmadesc *desc,
struct bcm43xx_dmadesc_meta *meta,
int irq_context)
{
@@ -188,23 +216,13 @@
{
struct device *dev = &(ring->bcm->pci_dev->dev);
- ring->vbase = dma_alloc_coherent(dev, BCM43xx_DMA_RINGMEMSIZE,
- &(ring->dmabase), GFP_KERNEL);
- if (!ring->vbase) {
+ ring->descbase = dma_alloc_coherent(dev, BCM43xx_DMA_RINGMEMSIZE,
+ &(ring->dmabase), GFP_KERNEL);
+ if (!ring->descbase) {
printk(KERN_ERR PFX "DMA ringmemory allocation failed\n");
return -ENOMEM;
}
- if (ring->dmabase + BCM43xx_DMA_RINGMEMSIZE > BCM43xx_DMA_BUSADDRMAX) {
- printk(KERN_ERR PFX ">>>FATAL ERROR<<< DMA RINGMEMORY >1G "
- "(0x%llx, len: %lu)\n",
- (unsigned long long)ring->dmabase,
- BCM43xx_DMA_RINGMEMSIZE);
- dma_free_coherent(dev, BCM43xx_DMA_RINGMEMSIZE,
- ring->vbase, ring->dmabase);
- return -ENOMEM;
- }
- assert(!(ring->dmabase & 0x000003FF));
- memset(ring->vbase, 0, BCM43xx_DMA_RINGMEMSIZE);
+ memset(ring->descbase, 0, BCM43xx_DMA_RINGMEMSIZE);
return 0;
}
@@ -214,26 +232,34 @@
struct device *dev = &(ring->bcm->pci_dev->dev);
dma_free_coherent(dev, BCM43xx_DMA_RINGMEMSIZE,
- ring->vbase, ring->dmabase);
+ ring->descbase, ring->dmabase);
}
/* Reset the RX DMA channel */
int bcm43xx_dmacontroller_rx_reset(struct bcm43xx_private *bcm,
- u16 mmio_base)
+ u16 mmio_base, int dma64)
{
int i;
u32 value;
+ u16 offset;
- bcm43xx_write32(bcm,
- mmio_base + BCM43xx_DMA_RX_CONTROL,
- 0x00000000);
+ offset = dma64 ? BCM43xx_DMA64_RXCTL : BCM43xx_DMA32_RXCTL;
+ bcm43xx_write32(bcm, mmio_base + offset, 0);
for (i = 0; i < 1000; i++) {
- value = bcm43xx_read32(bcm,
- mmio_base + BCM43xx_DMA_RX_STATUS);
- value &= BCM43xx_DMA_RXSTAT_STAT_MASK;
- if (value == BCM43xx_DMA_RXSTAT_STAT_DISABLED) {
- i = -1;
- break;
+ offset = dma64 ? BCM43xx_DMA64_RXSTATUS : BCM43xx_DMA32_RXSTATUS;
+ value = bcm43xx_read32(bcm, mmio_base + offset);
+ if (dma64) {
+ value &= BCM43xx_DMA64_RXSTAT;
+ if (value == BCM43xx_DMA64_RXSTAT_DISABLED) {
+ i = -1;
+ break;
+ }
+ } else {
+ value &= BCM43xx_DMA32_RXSTATE;
+ if (value == BCM43xx_DMA32_RXSTAT_DISABLED) {
+ i = -1;
+ break;
+ }
}
udelay(10);
}
@@ -247,31 +273,47 @@
/* Reset the RX DMA channel */
int bcm43xx_dmacontroller_tx_reset(struct bcm43xx_private *bcm,
- u16 mmio_base)
+ u16 mmio_base, int dma64)
{
int i;
u32 value;
+ u16 offset;
for (i = 0; i < 1000; i++) {
- value = bcm43xx_read32(bcm,
- mmio_base + BCM43xx_DMA_TX_STATUS);
- value &= BCM43xx_DMA_TXSTAT_STAT_MASK;
- if (value == BCM43xx_DMA_TXSTAT_STAT_DISABLED ||
- value == BCM43xx_DMA_TXSTAT_STAT_IDLEWAIT ||
- value == BCM43xx_DMA_TXSTAT_STAT_STOPPED)
- break;
+ offset = dma64 ? BCM43xx_DMA64_TXSTATUS : BCM43xx_DMA32_TXSTATUS;
+ value = bcm43xx_read32(bcm, mmio_base + offset);
+ if (dma64) {
+ value &= BCM43xx_DMA64_TXSTAT;
+ if (value == BCM43xx_DMA64_TXSTAT_DISABLED ||
+ value == BCM43xx_DMA64_TXSTAT_IDLEWAIT ||
+ value == BCM43xx_DMA64_TXSTAT_STOPPED)
+ break;
+ } else {
+ value &= BCM43xx_DMA32_TXSTATE;
+ if (value == BCM43xx_DMA32_TXSTAT_DISABLED ||
+ value == BCM43xx_DMA32_TXSTAT_IDLEWAIT ||
+ value == BCM43xx_DMA32_TXSTAT_STOPPED)
+ break;
+ }
udelay(10);
}
- bcm43xx_write32(bcm,
- mmio_base + BCM43xx_DMA_TX_CONTROL,
- 0x00000000);
+ offset = dma64 ? BCM43xx_DMA64_TXCTL : BCM43xx_DMA32_TXCTL;
+ bcm43xx_write32(bcm, mmio_base + offset, 0);
for (i = 0; i < 1000; i++) {
- value = bcm43xx_read32(bcm,
- mmio_base + BCM43xx_DMA_TX_STATUS);
- value &= BCM43xx_DMA_TXSTAT_STAT_MASK;
- if (value == BCM43xx_DMA_TXSTAT_STAT_DISABLED) {
- i = -1;
- break;
+ offset = dma64 ? BCM43xx_DMA64_TXSTATUS : BCM43xx_DMA32_TXSTATUS;
+ value = bcm43xx_read32(bcm, mmio_base + offset);
+ if (dma64) {
+ value &= BCM43xx_DMA64_TXSTAT;
+ if (value == BCM43xx_DMA64_TXSTAT_DISABLED) {
+ i = -1;
+ break;
+ }
+ } else {
+ value &= BCM43xx_DMA32_TXSTATE;
+ if (value == BCM43xx_DMA32_TXSTAT_DISABLED) {
+ i = -1;
+ break;
+ }
}
udelay(10);
}
@@ -285,47 +327,98 @@
return 0;
}
+static void fill_descriptor(struct bcm43xx_dmaring *ring,
+ struct bcm43xx_dmadesc_generic *desc,
+ dma_addr_t dmaaddr,
+ u16 bufsize,
+ int start, int end, int irq)
+{
+ int slot;
+
+ slot = bcm43xx_dma_desc2idx(ring, desc);
+ assert(slot >= 0 && slot < ring->nr_slots);
+
+ if (ring->dma64) {
+ u32 ctl0 = 0, ctl1 = 0;
+ u32 addrlo, addrhi;
+ u32 addrext;
+
+ addrlo = (u32)(dmaaddr & 0xFFFFFFFF);
+ addrhi = (((u64)dmaaddr >> 32) & ~BCM43xx_DMA64_ROUTING);
+ addrext = (((u64)dmaaddr >> 32) >> BCM43xx_DMA64_ROUTING_SHIFT);
+ addrhi |= ring->routing;
+ if (slot == ring->nr_slots - 1)
+ ctl0 |= BCM43xx_DMA64_DCTL0_DTABLEEND;
+ if (start)
+ ctl0 |= BCM43xx_DMA64_DCTL0_FRAMESTART;
+ if (end)
+ ctl0 |= BCM43xx_DMA64_DCTL0_FRAMEEND;
+ if (irq)
+ ctl0 |= BCM43xx_DMA64_DCTL0_IRQ;
+ ctl1 |= (bufsize - ring->frameoffset)
+ & BCM43xx_DMA64_DCTL1_BYTECNT;
+ ctl1 |= (addrext << BCM43xx_DMA64_DCTL1_ADDREXT_SHIFT)
+ & BCM43xx_DMA64_DCTL1_ADDREXT_MASK;
+
+ desc->dma64.control0 = cpu_to_le32(ctl0);
+ desc->dma64.control1 = cpu_to_le32(ctl1);
+ desc->dma64.address_low = cpu_to_le32(addrlo);
+ desc->dma64.address_high = cpu_to_le32(addrhi);
+ } else {
+ u32 ctl;
+ u32 addr;
+ u32 addrext;
+
+ addr = (u32)(dmaaddr & ~BCM43xx_DMA32_ROUTING);
+ addrext = (u32)(dmaaddr & BCM43xx_DMA32_ROUTING)
+ >> BCM43xx_DMA32_ROUTING_SHIFT;
+ addr |= ring->routing;
+ ctl = (bufsize - ring->frameoffset)
+ & BCM43xx_DMA32_DCTL_BYTECNT;
+ if (slot == ring->nr_slots - 1)
+ ctl |= BCM43xx_DMA32_DCTL_DTABLEEND;
+ if (start)
+ ctl |= BCM43xx_DMA32_DCTL_FRAMESTART;
+ if (end)
+ ctl |= BCM43xx_DMA32_DCTL_FRAMEEND;
+ if (irq)
+ ctl |= BCM43xx_DMA32_DCTL_IRQ;
+ ctl |= (addrext << BCM43xx_DMA32_DCTL_ADDREXT_SHIFT)
+ & BCM43xx_DMA32_DCTL_ADDREXT_MASK;
+
+ desc->dma32.control = cpu_to_le32(ctl);
+ desc->dma32.address = cpu_to_le32(addr);
+ }
+}
+
static int setup_rx_descbuffer(struct bcm43xx_dmaring *ring,
- struct bcm43xx_dmadesc *desc,
+ struct bcm43xx_dmadesc_generic *desc,
struct bcm43xx_dmadesc_meta *meta,
gfp_t gfp_flags)
{
struct bcm43xx_rxhdr *rxhdr;
+ struct bcm43xx_hwxmitstatus *xmitstat;
dma_addr_t dmaaddr;
- u32 desc_addr;
- u32 desc_ctl;
- const int slot = (int)(desc - ring->vbase);
struct sk_buff *skb;
- assert(slot >= 0 && slot < ring->nr_slots);
assert(!ring->tx);
skb = __dev_alloc_skb(ring->rx_buffersize, gfp_flags);
if (unlikely(!skb))
return -ENOMEM;
dmaaddr = map_descbuffer(ring, skb->data, ring->rx_buffersize, 0);
- if (unlikely(dmaaddr + ring->rx_buffersize > BCM43xx_DMA_BUSADDRMAX)) {
- unmap_descbuffer(ring, dmaaddr, ring->rx_buffersize, 0);
- dev_kfree_skb_any(skb);
- printk(KERN_ERR PFX ">>>FATAL ERROR<<< DMA RX SKB >1G "
- "(0x%llx, len: %u)\n",
- (unsigned long long)dmaaddr, ring->rx_buffersize);
- return -ENOMEM;
- }
meta->skb = skb;
meta->dmaaddr = dmaaddr;
skb->dev = ring->bcm->net_dev;
- desc_addr = (u32)(dmaaddr + ring->memoffset);
- desc_ctl = (BCM43xx_DMADTOR_BYTECNT_MASK &
- (u32)(ring->rx_buffersize - ring->frameoffset));
- if (slot == ring->nr_slots - 1)
- desc_ctl |= BCM43xx_DMADTOR_DTABLEEND;
- set_desc_addr(desc, desc_addr);
- set_desc_ctl(desc, desc_ctl);
+
+ fill_descriptor(ring, desc, dmaaddr,
+ ring->rx_buffersize, 0, 0, 0);
rxhdr = (struct bcm43xx_rxhdr *)(skb->data);
rxhdr->frame_length = 0;
rxhdr->flags1 = 0;
+ xmitstat = (struct bcm43xx_hwxmitstatus *)(skb->data);
+ xmitstat->cookie = 0;
return 0;
}
@@ -336,17 +429,17 @@
static int alloc_initial_descbuffers(struct bcm43xx_dmaring *ring)
{
int i, err = -ENOMEM;
- struct bcm43xx_dmadesc *desc;
+ struct bcm43xx_dmadesc_generic *desc;
struct bcm43xx_dmadesc_meta *meta;
for (i = 0; i < ring->nr_slots; i++) {
- desc = ring->vbase + i;
- meta = ring->meta + i;
+ desc = bcm43xx_dma_idx2desc(ring, i, &meta);
err = setup_rx_descbuffer(ring, desc, meta, GFP_KERNEL);
if (err)
goto err_unwind;
}
+ mb();
ring->used_slots = ring->nr_slots;
err = 0;
out:
@@ -354,8 +447,7 @@
err_unwind:
for (i--; i >= 0; i--) {
- desc = ring->vbase + i;
- meta = ring->meta + i;
+ desc = bcm43xx_dma_idx2desc(ring, i, &meta);
unmap_descbuffer(ring, meta->dmaaddr, ring->rx_buffersize, 0);
dev_kfree_skb(meta->skb);
@@ -371,27 +463,67 @@
{
int err = 0;
u32 value;
+ u32 addrext;
if (ring->tx) {
- /* Set Transmit Control register to "transmit enable" */
- bcm43xx_dma_write(ring, BCM43xx_DMA_TX_CONTROL,
- BCM43xx_DMA_TXCTRL_ENABLE);
- /* Set Transmit Descriptor ring address. */
- bcm43xx_dma_write(ring, BCM43xx_DMA_TX_DESC_RING,
- ring->dmabase + ring->memoffset);
+ if (ring->dma64) {
+ u64 ringbase = (u64)(ring->dmabase);
+
+ addrext = ((ringbase >> 32) >> BCM43xx_DMA64_ROUTING_SHIFT);
+ value = BCM43xx_DMA64_TXENABLE;
+ value |= (addrext << BCM43xx_DMA64_TXADDREXT_SHIFT)
+ & BCM43xx_DMA64_TXADDREXT_MASK;
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_TXCTL, value);
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_TXRINGLO,
+ (ringbase & 0xFFFFFFFF));
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_TXRINGHI,
+ ((ringbase >> 32) & ~BCM43xx_DMA64_ROUTING)
+ | ring->routing);
+ } else {
+ u32 ringbase = (u32)(ring->dmabase);
+
+ addrext = (ringbase >> BCM43xx_DMA32_ROUTING_SHIFT);
+ value = BCM43xx_DMA32_TXENABLE;
+ value |= (addrext << BCM43xx_DMA32_TXADDREXT_SHIFT)
+ & BCM43xx_DMA32_TXADDREXT_MASK;
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_TXCTL, value);
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_TXRING,
+ (ringbase & ~BCM43xx_DMA32_ROUTING)
+ | ring->routing);
+ }
} else {
err = alloc_initial_descbuffers(ring);
if (err)
goto out;
- /* Set Receive Control "receive enable" and frame offset */
- value = (ring->frameoffset << BCM43xx_DMA_RXCTRL_FRAMEOFF_SHIFT);
- value |= BCM43xx_DMA_RXCTRL_ENABLE;
- bcm43xx_dma_write(ring, BCM43xx_DMA_RX_CONTROL, value);
- /* Set Receive Descriptor ring address. */
- bcm43xx_dma_write(ring, BCM43xx_DMA_RX_DESC_RING,
- ring->dmabase + ring->memoffset);
- /* Init the descriptor pointer. */
- bcm43xx_dma_write(ring, BCM43xx_DMA_RX_DESC_INDEX, 200);
+ if (ring->dma64) {
+ u64 ringbase = (u64)(ring->dmabase);
+
+ addrext = ((ringbase >> 32) >> BCM43xx_DMA64_ROUTING_SHIFT);
+ value = (ring->frameoffset << BCM43xx_DMA64_RXFROFF_SHIFT);
+ value |= BCM43xx_DMA64_RXENABLE;
+ value |= (addrext << BCM43xx_DMA64_RXADDREXT_SHIFT)
+ & BCM43xx_DMA64_RXADDREXT_MASK;
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_RXCTL, value);
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_RXRINGLO,
+ (ringbase & 0xFFFFFFFF));
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_RXRINGHI,
+ ((ringbase >> 32) & ~BCM43xx_DMA64_ROUTING)
+ | ring->routing);
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_RXINDEX, 200);
+ } else {
+ u32 ringbase = (u32)(ring->dmabase);
+
+ addrext = (ringbase >> BCM43xx_DMA32_ROUTING_SHIFT);
+ value = (ring->frameoffset << BCM43xx_DMA32_RXFROFF_SHIFT);
+ value |= BCM43xx_DMA32_RXENABLE;
+ value |= (addrext << BCM43xx_DMA32_RXADDREXT_SHIFT)
+ & BCM43xx_DMA32_RXADDREXT_MASK;
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_RXCTL, value);
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_RXRING,
+ (ringbase & ~BCM43xx_DMA32_ROUTING)
+ | ring->routing);
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_RXINDEX, 200);
+ }
}
out:
@@ -402,27 +534,32 @@
static void dmacontroller_cleanup(struct bcm43xx_dmaring *ring)
{
if (ring->tx) {
- bcm43xx_dmacontroller_tx_reset(ring->bcm, ring->mmio_base);
- /* Zero out Transmit Descriptor ring address. */
- bcm43xx_dma_write(ring, BCM43xx_DMA_TX_DESC_RING, 0);
+ bcm43xx_dmacontroller_tx_reset(ring->bcm, ring->mmio_base, ring->dma64);
+ if (ring->dma64) {
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_TXRINGLO, 0);
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_TXRINGHI, 0);
+ } else
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_TXRING, 0);
} else {
- bcm43xx_dmacontroller_rx_reset(ring->bcm, ring->mmio_base);
- /* Zero out Receive Descriptor ring address. */
- bcm43xx_dma_write(ring, BCM43xx_DMA_RX_DESC_RING, 0);
+ bcm43xx_dmacontroller_rx_reset(ring->bcm, ring->mmio_base, ring->dma64);
+ if (ring->dma64) {
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_RXRINGLO, 0);
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_RXRINGHI, 0);
+ } else
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_RXRING, 0);
}
}
static void free_all_descbuffers(struct bcm43xx_dmaring *ring)
{
- struct bcm43xx_dmadesc *desc;
+ struct bcm43xx_dmadesc_generic *desc;
struct bcm43xx_dmadesc_meta *meta;
int i;
if (!ring->used_slots)
return;
for (i = 0; i < ring->nr_slots; i++) {
- desc = ring->vbase + i;
- meta = ring->meta + i;
+ desc = bcm43xx_dma_idx2desc(ring, i, &meta);
if (!meta->skb) {
assert(ring->tx);
@@ -430,62 +567,67 @@
}
if (ring->tx) {
unmap_descbuffer(ring, meta->dmaaddr,
- meta->skb->len, 1);
+ meta->skb->len, 1);
} else {
unmap_descbuffer(ring, meta->dmaaddr,
- ring->rx_buffersize, 0);
+ ring->rx_buffersize, 0);
}
- free_descriptor_buffer(ring, desc, meta, 0);
+ free_descriptor_buffer(ring, meta, 0);
}
}
/* Main initialization function. */
static
struct bcm43xx_dmaring * bcm43xx_setup_dmaring(struct bcm43xx_private *bcm,
- u16 dma_controller_base,
- int nr_descriptor_slots,
- int tx)
+ int controller_index,
+ int for_tx,
+ int dma64)
{
struct bcm43xx_dmaring *ring;
int err;
+ int nr_slots;
ring = kzalloc(sizeof(*ring), GFP_KERNEL);
if (!ring)
goto out;
- ring->meta = kzalloc(sizeof(*ring->meta) * nr_descriptor_slots,
- GFP_KERNEL);
+ nr_slots = BCM43xx_RXRING_SLOTS;
+ if (for_tx)
+ nr_slots = BCM43xx_TXRING_SLOTS;
+
+ ring->meta = kzalloc(sizeof(*ring->meta) * nr_slots,
+ GFP_KERNEL);
if (!ring->meta)
goto err_kfree_ring;
- ring->memoffset = BCM43xx_DMA_DMABUSADDROFFSET;
+ ring->routing = BCM43xx_DMA32_CLIENTTRANS;
+ if (dma64)
+ ring->routing = BCM43xx_DMA64_CLIENTTRANS;
#ifdef CONFIG_BCM947XX
if (bcm->pci_dev->bus->number == 0)
- ring->memoffset = 0;
+ ring->routing = dma64 ? BCM43xx_DMA64_NOTRANS : BCM43xx_DMA32_NOTRANS;
#endif
ring->bcm = bcm;
- ring->nr_slots = nr_descriptor_slots;
+ ring->nr_slots = nr_slots;
ring->suspend_mark = ring->nr_slots * BCM43xx_TXSUSPEND_PERCENT / 100;
ring->resume_mark = ring->nr_slots * BCM43xx_TXRESUME_PERCENT / 100;
assert(ring->suspend_mark < ring->resume_mark);
- ring->mmio_base = dma_controller_base;
- if (tx) {
+ ring->mmio_base = bcm43xx_dmacontroller_base(dma64, controller_index);
+ ring->index = controller_index;
+ ring->dma64 = !!dma64;
+ if (for_tx) {
ring->tx = 1;
ring->current_slot = -1;
} else {
- switch (dma_controller_base) {
- case BCM43xx_MMIO_DMA1_BASE:
- ring->rx_buffersize = BCM43xx_DMA1_RXBUFFERSIZE;
- ring->frameoffset = BCM43xx_DMA1_RX_FRAMEOFFSET;
- break;
- case BCM43xx_MMIO_DMA4_BASE:
- ring->rx_buffersize = BCM43xx_DMA4_RXBUFFERSIZE;
- ring->frameoffset = BCM43xx_DMA4_RX_FRAMEOFFSET;
- break;
- default:
+ if (ring->index == 0) {
+ ring->rx_buffersize = BCM43xx_DMA0_RX_BUFFERSIZE;
+ ring->frameoffset = BCM43xx_DMA0_RX_FRAMEOFFSET;
+ } else if (ring->index == 3) {
+ ring->rx_buffersize = BCM43xx_DMA3_RX_BUFFERSIZE;
+ ring->frameoffset = BCM43xx_DMA3_RX_FRAMEOFFSET;
+ } else
assert(0);
- }
}
err = alloc_ringmemory(ring);
@@ -514,7 +656,8 @@
if (!ring)
return;
- dprintk(KERN_INFO PFX "DMA 0x%04x (%s) max used slots: %d/%d\n",
+ dprintk(KERN_INFO PFX "DMA-%s 0x%04X (%s) max used slots: %d/%d\n",
+ (ring->dma64) ? "64" : "32",
ring->mmio_base,
(ring->tx) ? "TX" : "RX",
ring->max_used_slots, ring->nr_slots);
@@ -537,10 +680,15 @@
return;
dma = bcm43xx_current_dma(bcm);
- bcm43xx_destroy_dmaring(dma->rx_ring1);
- dma->rx_ring1 = NULL;
+ bcm43xx_destroy_dmaring(dma->rx_ring3);
+ dma->rx_ring3 = NULL;
bcm43xx_destroy_dmaring(dma->rx_ring0);
dma->rx_ring0 = NULL;
+
+ bcm43xx_destroy_dmaring(dma->tx_ring5);
+ dma->tx_ring5 = NULL;
+ bcm43xx_destroy_dmaring(dma->tx_ring4);
+ dma->tx_ring4 = NULL;
bcm43xx_destroy_dmaring(dma->tx_ring3);
dma->tx_ring3 = NULL;
bcm43xx_destroy_dmaring(dma->tx_ring2);
@@ -556,48 +704,59 @@
struct bcm43xx_dma *dma = bcm43xx_current_dma(bcm);
struct bcm43xx_dmaring *ring;
int err = -ENOMEM;
+ int dma64 = 0;
+ u32 sbtmstatehi;
+
+ sbtmstatehi = bcm43xx_read32(bcm, BCM43xx_CIR_SBTMSTATEHIGH);
+ if (sbtmstatehi & BCM43xx_SBTMSTATEHIGH_DMA64BIT)
+ dma64 = 1;
/* setup TX DMA channels. */
- ring = bcm43xx_setup_dmaring(bcm, BCM43xx_MMIO_DMA1_BASE,
- BCM43xx_TXRING_SLOTS, 1);
+ ring = bcm43xx_setup_dmaring(bcm, 0, 1, dma64);
if (!ring)
goto out;
dma->tx_ring0 = ring;
- ring = bcm43xx_setup_dmaring(bcm, BCM43xx_MMIO_DMA2_BASE,
- BCM43xx_TXRING_SLOTS, 1);
+ ring = bcm43xx_setup_dmaring(bcm, 1, 1, dma64);
if (!ring)
goto err_destroy_tx0;
dma->tx_ring1 = ring;
- ring = bcm43xx_setup_dmaring(bcm, BCM43xx_MMIO_DMA3_BASE,
- BCM43xx_TXRING_SLOTS, 1);
+ ring = bcm43xx_setup_dmaring(bcm, 2, 1, dma64);
if (!ring)
goto err_destroy_tx1;
dma->tx_ring2 = ring;
- ring = bcm43xx_setup_dmaring(bcm, BCM43xx_MMIO_DMA4_BASE,
- BCM43xx_TXRING_SLOTS, 1);
+ ring = bcm43xx_setup_dmaring(bcm, 3, 1, dma64);
if (!ring)
goto err_destroy_tx2;
dma->tx_ring3 = ring;
- /* setup RX DMA channels. */
- ring = bcm43xx_setup_dmaring(bcm, BCM43xx_MMIO_DMA1_BASE,
- BCM43xx_RXRING_SLOTS, 0);
+ ring = bcm43xx_setup_dmaring(bcm, 4, 1, dma64);
if (!ring)
goto err_destroy_tx3;
+ dma->tx_ring4 = ring;
+
+ ring = bcm43xx_setup_dmaring(bcm, 5, 1, dma64);
+ if (!ring)
+ goto err_destroy_tx4;
+ dma->tx_ring5 = ring;
+
+ /* setup RX DMA channels. */
+ ring = bcm43xx_setup_dmaring(bcm, 0, 0, dma64);
+ if (!ring)
+ goto err_destroy_tx5;
dma->rx_ring0 = ring;
if (bcm->current_core->rev < 5) {
- ring = bcm43xx_setup_dmaring(bcm, BCM43xx_MMIO_DMA4_BASE,
- BCM43xx_RXRING_SLOTS, 0);
+ ring = bcm43xx_setup_dmaring(bcm, 3, 0, dma64);
if (!ring)
goto err_destroy_rx0;
- dma->rx_ring1 = ring;
+ dma->rx_ring3 = ring;
}
- dprintk(KERN_INFO PFX "DMA initialized\n");
+ dprintk(KERN_INFO PFX "%s DMA initialized\n",
+ dma64 ? "64-bit" : "32-bit");
err = 0;
out:
return err;
@@ -605,6 +764,12 @@
err_destroy_rx0:
bcm43xx_destroy_dmaring(dma->rx_ring0);
dma->rx_ring0 = NULL;
+err_destroy_tx5:
+ bcm43xx_destroy_dmaring(dma->tx_ring5);
+ dma->tx_ring5 = NULL;
+err_destroy_tx4:
+ bcm43xx_destroy_dmaring(dma->tx_ring4);
+ dma->tx_ring4 = NULL;
err_destroy_tx3:
bcm43xx_destroy_dmaring(dma->tx_ring3);
dma->tx_ring3 = NULL;
@@ -624,7 +789,7 @@
static u16 generate_cookie(struct bcm43xx_dmaring *ring,
int slot)
{
- u16 cookie = 0xF000;
+ u16 cookie = 0x1000;
/* Use the upper 4 bits of the cookie as
* DMA controller ID and store the slot number
@@ -632,21 +797,25 @@
* Note that the cookie must never be 0, as this
* is a special value used in RX path.
*/
- switch (ring->mmio_base) {
- default:
- assert(0);
- case BCM43xx_MMIO_DMA1_BASE:
+ switch (ring->index) {
+ case 0:
cookie = 0xA000;
break;
- case BCM43xx_MMIO_DMA2_BASE:
+ case 1:
cookie = 0xB000;
break;
- case BCM43xx_MMIO_DMA3_BASE:
+ case 2:
cookie = 0xC000;
break;
- case BCM43xx_MMIO_DMA4_BASE:
+ case 3:
cookie = 0xD000;
break;
+ case 4:
+ cookie = 0xE000;
+ break;
+ case 5:
+ cookie = 0xF000;
+ break;
}
assert(((u16)slot & 0xF000) == 0x0000);
cookie |= (u16)slot;
@@ -675,6 +844,12 @@
case 0xD000:
ring = dma->tx_ring3;
break;
+ case 0xE000:
+ ring = dma->tx_ring4;
+ break;
+ case 0xF000:
+ ring = dma->tx_ring5;
+ break;
default:
assert(0);
}
@@ -687,6 +862,9 @@
static void dmacontroller_poke_tx(struct bcm43xx_dmaring *ring,
int slot)
{
+ u16 offset;
+ int descsize;
+
/* Everything is ready to start. Buffers are DMA mapped and
* associated with slots.
* "slot" is the last slot of the new frame we want to transmit.
@@ -694,25 +872,26 @@
*/
wmb();
slot = next_slot(ring, slot);
- bcm43xx_dma_write(ring, BCM43xx_DMA_TX_DESC_INDEX,
- (u32)(slot * sizeof(struct bcm43xx_dmadesc)));
+ offset = (ring->dma64) ? BCM43xx_DMA64_TXINDEX : BCM43xx_DMA32_TXINDEX;
+ descsize = (ring->dma64) ? sizeof(struct bcm43xx_dmadesc64)
+ : sizeof(struct bcm43xx_dmadesc32);
+ bcm43xx_dma_write(ring, offset,
+ (u32)(slot * descsize));
}
-static int dma_tx_fragment(struct bcm43xx_dmaring *ring,
- struct sk_buff *skb,
- u8 cur_frag)
+static void dma_tx_fragment(struct bcm43xx_dmaring *ring,
+ struct sk_buff *skb,
+ u8 cur_frag)
{
int slot;
- struct bcm43xx_dmadesc *desc;
+ struct bcm43xx_dmadesc_generic *desc;
struct bcm43xx_dmadesc_meta *meta;
- u32 desc_ctl;
- u32 desc_addr;
+ dma_addr_t dmaaddr;
assert(skb_shinfo(skb)->nr_frags == 0);
slot = request_slot(ring);
- desc = ring->vbase + slot;
- meta = ring->meta + slot;
+ desc = bcm43xx_dma_idx2desc(ring, slot, &meta);
/* Add a device specific TX header. */
assert(skb_headroom(skb) >= sizeof(struct bcm43xx_txhdr));
@@ -729,29 +908,14 @@
generate_cookie(ring, slot));
meta->skb = skb;
- meta->dmaaddr = map_descbuffer(ring, skb->data, skb->len, 1);
- if (unlikely(meta->dmaaddr + skb->len > BCM43xx_DMA_BUSADDRMAX)) {
- return_slot(ring, slot);
- printk(KERN_ERR PFX ">>>FATAL ERROR<<< DMA TX SKB >1G "
- "(0x%llx, len: %u)\n",
- (unsigned long long)meta->dmaaddr, skb->len);
- return -ENOMEM;
- }
+ dmaaddr = map_descbuffer(ring, skb->data, skb->len, 1);
+ meta->dmaaddr = dmaaddr;
- desc_addr = (u32)(meta->dmaaddr + ring->memoffset);
- desc_ctl = BCM43xx_DMADTOR_FRAMESTART | BCM43xx_DMADTOR_FRAMEEND;
- desc_ctl |= BCM43xx_DMADTOR_COMPIRQ;
- desc_ctl |= (BCM43xx_DMADTOR_BYTECNT_MASK &
- (u32)(meta->skb->len - ring->frameoffset));
- if (slot == ring->nr_slots - 1)
- desc_ctl |= BCM43xx_DMADTOR_DTABLEEND;
+ fill_descriptor(ring, desc, dmaaddr,
+ skb->len, 1, 1, 1);
- set_desc_ctl(desc, desc_ctl);
- set_desc_addr(desc, desc_addr);
/* Now transfer the whole frame. */
dmacontroller_poke_tx(ring, slot);
-
- return 0;
}
int bcm43xx_dma_tx(struct bcm43xx_private *bcm,
@@ -781,7 +945,6 @@
/* Take skb from ieee80211_txb_free */
txb->fragments[i] = NULL;
dma_tx_fragment(ring, skb, i);
- //TODO: handle failure of dma_tx_fragment
}
ieee80211_txb_free(txb);
@@ -792,23 +955,28 @@
struct bcm43xx_xmitstatus *status)
{
struct bcm43xx_dmaring *ring;
- struct bcm43xx_dmadesc *desc;
+ struct bcm43xx_dmadesc_generic *desc;
struct bcm43xx_dmadesc_meta *meta;
int is_last_fragment;
int slot;
+ u32 tmp;
ring = parse_cookie(bcm, status->cookie, &slot);
assert(ring);
assert(ring->tx);
- assert(get_desc_ctl(ring->vbase + slot) & BCM43xx_DMADTOR_FRAMESTART);
while (1) {
assert(slot >= 0 && slot < ring->nr_slots);
- desc = ring->vbase + slot;
- meta = ring->meta + slot;
+ desc = bcm43xx_dma_idx2desc(ring, slot, &meta);
- is_last_fragment = !!(get_desc_ctl(desc) & BCM43xx_DMADTOR_FRAMEEND);
+ if (ring->dma64) {
+ tmp = le32_to_cpu(desc->dma64.control0);
+ is_last_fragment = !!(tmp & BCM43xx_DMA64_DCTL0_FRAMEEND);
+ } else {
+ tmp = le32_to_cpu(desc->dma32.control);
+ is_last_fragment = !!(tmp & BCM43xx_DMA32_DCTL_FRAMEEND);
+ }
unmap_descbuffer(ring, meta->dmaaddr, meta->skb->len, 1);
- free_descriptor_buffer(ring, desc, meta, 1);
+ free_descriptor_buffer(ring, meta, 1);
/* Everything belonging to the slot is unmapped
* and freed, so we can return it.
*/
@@ -824,7 +992,7 @@
static void dma_rx(struct bcm43xx_dmaring *ring,
int *slot)
{
- struct bcm43xx_dmadesc *desc;
+ struct bcm43xx_dmadesc_generic *desc;
struct bcm43xx_dmadesc_meta *meta;
struct bcm43xx_rxhdr *rxhdr;
struct sk_buff *skb;
@@ -832,13 +1000,12 @@
int err;
dma_addr_t dmaaddr;
- desc = ring->vbase + *slot;
- meta = ring->meta + *slot;
+ desc = bcm43xx_dma_idx2desc(ring, *slot, &meta);
sync_descbuffer_for_cpu(ring, meta->dmaaddr, ring->rx_buffersize);
skb = meta->skb;
- if (ring->mmio_base == BCM43xx_MMIO_DMA4_BASE) {
+ if (ring->index == 3) {
/* We received an xmit status. */
struct bcm43xx_hwxmitstatus *hw = (struct bcm43xx_hwxmitstatus *)skb->data;
struct bcm43xx_xmitstatus stat;
@@ -894,8 +1061,7 @@
s32 tmp = len;
while (1) {
- desc = ring->vbase + *slot;
- meta = ring->meta + *slot;
+ desc = bcm43xx_dma_idx2desc(ring, *slot, &meta);
/* recycle the descriptor buffer. */
sync_descbuffer_for_device(ring, meta->dmaaddr,
ring->rx_buffersize);
@@ -906,8 +1072,8 @@
break;
}
printkl(KERN_ERR PFX "DMA RX buffer too small "
- "(len: %u, buffer: %u, nr-dropped: %d)\n",
- len, ring->rx_buffersize, cnt);
+ "(len: %u, buffer: %u, nr-dropped: %d)\n",
+ len, ring->rx_buffersize, cnt);
goto drop;
}
len -= IEEE80211_FCS_LEN;
@@ -945,9 +1111,15 @@
#endif
assert(!ring->tx);
- status = bcm43xx_dma_read(ring, BCM43xx_DMA_RX_STATUS);
- descptr = (status & BCM43xx_DMA_RXSTAT_DPTR_MASK);
- current_slot = descptr / sizeof(struct bcm43xx_dmadesc);
+ if (ring->dma64) {
+ status = bcm43xx_dma_read(ring, BCM43xx_DMA64_RXSTATUS);
+ descptr = (status & BCM43xx_DMA64_RXSTATDPTR);
+ current_slot = descptr / sizeof(struct bcm43xx_dmadesc64);
+ } else {
+ status = bcm43xx_dma_read(ring, BCM43xx_DMA32_RXSTATUS);
+ descptr = (status & BCM43xx_DMA32_RXDPTR);
+ current_slot = descptr / sizeof(struct bcm43xx_dmadesc32);
+ }
assert(current_slot >= 0 && current_slot < ring->nr_slots);
slot = ring->current_slot;
@@ -958,8 +1130,13 @@
ring->max_used_slots = used_slots;
#endif
}
- bcm43xx_dma_write(ring, BCM43xx_DMA_RX_DESC_INDEX,
- (u32)(slot * sizeof(struct bcm43xx_dmadesc)));
+ if (ring->dma64) {
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_RXINDEX,
+ (u32)(slot * sizeof(struct bcm43xx_dmadesc64)));
+ } else {
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_RXINDEX,
+ (u32)(slot * sizeof(struct bcm43xx_dmadesc32)));
+ }
ring->current_slot = slot;
}
@@ -967,16 +1144,28 @@
{
assert(ring->tx);
bcm43xx_power_saving_ctl_bits(ring->bcm, -1, 1);
- bcm43xx_dma_write(ring, BCM43xx_DMA_TX_CONTROL,
- bcm43xx_dma_read(ring, BCM43xx_DMA_TX_CONTROL)
- | BCM43xx_DMA_TXCTRL_SUSPEND);
+ if (ring->dma64) {
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_TXCTL,
+ bcm43xx_dma_read(ring, BCM43xx_DMA64_TXCTL)
+ | BCM43xx_DMA64_TXSUSPEND);
+ } else {
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_TXCTL,
+ bcm43xx_dma_read(ring, BCM43xx_DMA32_TXCTL)
+ | BCM43xx_DMA32_TXSUSPEND);
+ }
}
void bcm43xx_dma_tx_resume(struct bcm43xx_dmaring *ring)
{
assert(ring->tx);
- bcm43xx_dma_write(ring, BCM43xx_DMA_TX_CONTROL,
- bcm43xx_dma_read(ring, BCM43xx_DMA_TX_CONTROL)
- & ~BCM43xx_DMA_TXCTRL_SUSPEND);
+ if (ring->dma64) {
+ bcm43xx_dma_write(ring, BCM43xx_DMA64_TXCTL,
+ bcm43xx_dma_read(ring, BCM43xx_DMA64_TXCTL)
+ & ~BCM43xx_DMA64_TXSUSPEND);
+ } else {
+ bcm43xx_dma_write(ring, BCM43xx_DMA32_TXCTL,
+ bcm43xx_dma_read(ring, BCM43xx_DMA32_TXCTL)
+ & ~BCM43xx_DMA32_TXSUSPEND);
+ }
bcm43xx_power_saving_ctl_bits(ring->bcm, -1, -1);
}
Index: wireless-dev/drivers/net/wireless/bcm43xx/bcm43xx_dma.h
===================================================================
--- wireless-dev.orig/drivers/net/wireless/bcm43xx/bcm43xx_dma.h 2006-06-17 17:48:26.000000000 +0200
+++ wireless-dev/drivers/net/wireless/bcm43xx/bcm43xx_dma.h 2006-06-17 19:09:40.000000000 +0200
@@ -14,63 +14,179 @@
#define BCM43xx_DMAIRQ_NONFATALMASK (1 << 13)
#define BCM43xx_DMAIRQ_RX_DONE (1 << 16)
-/* DMA controller register offsets. (relative to BCM43xx_DMA#_BASE) */
-#define BCM43xx_DMA_TX_CONTROL 0x00
-#define BCM43xx_DMA_TX_DESC_RING 0x04
-#define BCM43xx_DMA_TX_DESC_INDEX 0x08
-#define BCM43xx_DMA_TX_STATUS 0x0c
-#define BCM43xx_DMA_RX_CONTROL 0x10
-#define BCM43xx_DMA_RX_DESC_RING 0x14
-#define BCM43xx_DMA_RX_DESC_INDEX 0x18
-#define BCM43xx_DMA_RX_STATUS 0x1c
-
-/* DMA controller channel control word values. */
-#define BCM43xx_DMA_TXCTRL_ENABLE (1 << 0)
-#define BCM43xx_DMA_TXCTRL_SUSPEND (1 << 1)
-#define BCM43xx_DMA_TXCTRL_LOOPBACK (1 << 2)
-#define BCM43xx_DMA_TXCTRL_FLUSH (1 << 4)
-#define BCM43xx_DMA_RXCTRL_ENABLE (1 << 0)
-#define BCM43xx_DMA_RXCTRL_FRAMEOFF_MASK 0x000000fe
-#define BCM43xx_DMA_RXCTRL_FRAMEOFF_SHIFT 1
-#define BCM43xx_DMA_RXCTRL_PIO (1 << 8)
-/* DMA controller channel status word values. */
-#define BCM43xx_DMA_TXSTAT_DPTR_MASK 0x00000fff
-#define BCM43xx_DMA_TXSTAT_STAT_MASK 0x0000f000
-#define BCM43xx_DMA_TXSTAT_STAT_DISABLED 0x00000000
-#define BCM43xx_DMA_TXSTAT_STAT_ACTIVE 0x00001000
-#define BCM43xx_DMA_TXSTAT_STAT_IDLEWAIT 0x00002000
-#define BCM43xx_DMA_TXSTAT_STAT_STOPPED 0x00003000
-#define BCM43xx_DMA_TXSTAT_STAT_SUSP 0x00004000
-#define BCM43xx_DMA_TXSTAT_ERROR_MASK 0x000f0000
-#define BCM43xx_DMA_TXSTAT_FLUSHED (1 << 20)
-#define BCM43xx_DMA_RXSTAT_DPTR_MASK 0x00000fff
-#define BCM43xx_DMA_RXSTAT_STAT_MASK 0x0000f000
-#define BCM43xx_DMA_RXSTAT_STAT_DISABLED 0x00000000
-#define BCM43xx_DMA_RXSTAT_STAT_ACTIVE 0x00001000
-#define BCM43xx_DMA_RXSTAT_STAT_IDLEWAIT 0x00002000
-#define BCM43xx_DMA_RXSTAT_STAT_RESERVED 0x00003000
-#define BCM43xx_DMA_RXSTAT_STAT_ERRORS 0x00004000
-#define BCM43xx_DMA_RXSTAT_ERROR_MASK 0x000f0000
-
-/* DMA descriptor control field values. */
-#define BCM43xx_DMADTOR_BYTECNT_MASK 0x00001fff
-#define BCM43xx_DMADTOR_DTABLEEND (1 << 28) /* End of descriptor table */
-#define BCM43xx_DMADTOR_COMPIRQ (1 << 29) /* IRQ on completion request */
-#define BCM43xx_DMADTOR_FRAMEEND (1 << 30)
-#define BCM43xx_DMADTOR_FRAMESTART (1 << 31)
+
+/*** 32-bit DMA Engine. ***/
+
+/* 32-bit DMA controller registers. */
+#define BCM43xx_DMA32_TXCTL 0x00
+#define BCM43xx_DMA32_TXENABLE 0x00000001
+#define BCM43xx_DMA32_TXSUSPEND 0x00000002
+#define BCM43xx_DMA32_TXLOOPBACK 0x00000004
+#define BCM43xx_DMA32_TXFLUSH 0x00000010
+#define BCM43xx_DMA32_TXADDREXT_MASK 0x00030000
+#define BCM43xx_DMA32_TXADDREXT_SHIFT 16
+#define BCM43xx_DMA32_TXRING 0x04
+#define BCM43xx_DMA32_TXINDEX 0x08
+#define BCM43xx_DMA32_TXSTATUS 0x0C
+#define BCM43xx_DMA32_TXDPTR 0x00000FFF
+#define BCM43xx_DMA32_TXSTATE 0x0000F000
+#define BCM43xx_DMA32_TXSTAT_DISABLED 0x00000000
+#define BCM43xx_DMA32_TXSTAT_ACTIVE 0x00001000
+#define BCM43xx_DMA32_TXSTAT_IDLEWAIT 0x00002000
+#define BCM43xx_DMA32_TXSTAT_STOPPED 0x00003000
+#define BCM43xx_DMA32_TXSTAT_SUSP 0x00004000
+#define BCM43xx_DMA32_TXERROR 0x000F0000
+#define BCM43xx_DMA32_TXERR_NOERR 0x00000000
+#define BCM43xx_DMA32_TXERR_PROT 0x00010000
+#define BCM43xx_DMA32_TXERR_UNDERRUN 0x00020000
+#define BCM43xx_DMA32_TXERR_BUFREAD 0x00030000
+#define BCM43xx_DMA32_TXERR_DESCREAD 0x00040000
+#define BCM43xx_DMA32_TXACTIVE 0xFFF00000
+#define BCM43xx_DMA32_RXCTL 0x10
+#define BCM43xx_DMA32_RXENABLE 0x00000001
+#define BCM43xx_DMA32_RXFROFF_MASK 0x000000FE
+#define BCM43xx_DMA32_RXFROFF_SHIFT 1
+#define BCM43xx_DMA32_RXDIRECTFIFO 0x00000100
+#define BCM43xx_DMA32_RXADDREXT_MASK 0x00030000
+#define BCM43xx_DMA32_RXADDREXT_SHIFT 16
+#define BCM43xx_DMA32_RXRING 0x14
+#define BCM43xx_DMA32_RXINDEX 0x18
+#define BCM43xx_DMA32_RXSTATUS 0x1C
+#define BCM43xx_DMA32_RXDPTR 0x00000FFF
+#define BCM43xx_DMA32_RXSTATE 0x0000F000
+#define BCM43xx_DMA32_RXSTAT_DISABLED 0x00000000
+#define BCM43xx_DMA32_RXSTAT_ACTIVE 0x00001000
+#define BCM43xx_DMA32_RXSTAT_IDLEWAIT 0x00002000
+#define BCM43xx_DMA32_RXSTAT_STOPPED 0x00003000
+#define BCM43xx_DMA32_RXERROR 0x000F0000
+#define BCM43xx_DMA32_RXERR_NOERR 0x00000000
+#define BCM43xx_DMA32_RXERR_PROT 0x00010000
+#define BCM43xx_DMA32_RXERR_OVERFLOW 0x00020000
+#define BCM43xx_DMA32_RXERR_BUFWRITE 0x00030000
+#define BCM43xx_DMA32_RXERR_DESCREAD 0x00040000
+#define BCM43xx_DMA32_RXACTIVE 0xFFF00000
+
+/* 32-bit DMA descriptor. */
+struct bcm43xx_dmadesc32 {
+ __le32 control;
+ __le32 address;
+} __attribute__((__packed__));
+#define BCM43xx_DMA32_DCTL_BYTECNT 0x00001FFF
+#define BCM43xx_DMA32_DCTL_ADDREXT_MASK 0x00030000
+#define BCM43xx_DMA32_DCTL_ADDREXT_SHIFT 16
+#define BCM43xx_DMA32_DCTL_DTABLEEND 0x10000000
+#define BCM43xx_DMA32_DCTL_IRQ 0x20000000
+#define BCM43xx_DMA32_DCTL_FRAMEEND 0x40000000
+#define BCM43xx_DMA32_DCTL_FRAMESTART 0x80000000
+
+/* Address field Routing value. */
+#define BCM43xx_DMA32_ROUTING 0xC0000000
+#define BCM43xx_DMA32_ROUTING_SHIFT 30
+#define BCM43xx_DMA32_NOTRANS 0x00000000
+#define BCM43xx_DMA32_CLIENTTRANS 0x40000000
+
+
+
+/*** 64-bit DMA Engine. ***/
+
+/* 64-bit DMA controller registers. */
+#define BCM43xx_DMA64_TXCTL 0x00
+#define BCM43xx_DMA64_TXENABLE 0x00000001
+#define BCM43xx_DMA64_TXSUSPEND 0x00000002
+#define BCM43xx_DMA64_TXLOOPBACK 0x00000004
+#define BCM43xx_DMA64_TXFLUSH 0x00000010
+#define BCM43xx_DMA64_TXADDREXT_MASK 0x00030000
+#define BCM43xx_DMA64_TXADDREXT_SHIFT 16
+#define BCM43xx_DMA64_TXINDEX 0x04
+#define BCM43xx_DMA64_TXRINGLO 0x08
+#define BCM43xx_DMA64_TXRINGHI 0x0C
+#define BCM43xx_DMA64_TXSTATUS 0x10
+#define BCM43xx_DMA64_TXSTATDPTR 0x00001FFF
+#define BCM43xx_DMA64_TXSTAT 0xF0000000
+#define BCM43xx_DMA64_TXSTAT_DISABLED 0x00000000
+#define BCM43xx_DMA64_TXSTAT_ACTIVE 0x10000000
+#define BCM43xx_DMA64_TXSTAT_IDLEWAIT 0x20000000
+#define BCM43xx_DMA64_TXSTAT_STOPPED 0x30000000
+#define BCM43xx_DMA64_TXSTAT_SUSP 0x40000000
+#define BCM43xx_DMA64_TXERROR 0x14
+#define BCM43xx_DMA64_TXERRDPTR 0x0001FFFF
+#define BCM43xx_DMA64_TXERR 0xF0000000
+#define BCM43xx_DMA64_TXERR_NOERR 0x00000000
+#define BCM43xx_DMA64_TXERR_PROT 0x10000000
+#define BCM43xx_DMA64_TXERR_UNDERRUN 0x20000000
+#define BCM43xx_DMA64_TXERR_TRANSFER 0x30000000
+#define BCM43xx_DMA64_TXERR_DESCREAD 0x40000000
+#define BCM43xx_DMA64_TXERR_CORE 0x50000000
+#define BCM43xx_DMA64_RXCTL 0x20
+#define BCM43xx_DMA64_RXENABLE 0x00000001
+#define BCM43xx_DMA64_RXFROFF_MASK 0x000000FE
+#define BCM43xx_DMA64_RXFROFF_SHIFT 1
+#define BCM43xx_DMA64_RXDIRECTFIFO 0x00000100
+#define BCM43xx_DMA64_RXADDREXT_MASK 0x00030000
+#define BCM43xx_DMA64_RXADDREXT_SHIFT 16
+#define BCM43xx_DMA64_RXINDEX 0x24
+#define BCM43xx_DMA64_RXRINGLO 0x28
+#define BCM43xx_DMA64_RXRINGHI 0x2C
+#define BCM43xx_DMA64_RXSTATUS 0x30
+#define BCM43xx_DMA64_RXSTATDPTR 0x00001FFF
+#define BCM43xx_DMA64_RXSTAT 0xF0000000
+#define BCM43xx_DMA64_RXSTAT_DISABLED 0x00000000
+#define BCM43xx_DMA64_RXSTAT_ACTIVE 0x10000000
+#define BCM43xx_DMA64_RXSTAT_IDLEWAIT 0x20000000
+#define BCM43xx_DMA64_RXSTAT_STOPPED 0x30000000
+#define BCM43xx_DMA64_RXSTAT_SUSP 0x40000000
+#define BCM43xx_DMA64_RXERROR 0x34
+#define BCM43xx_DMA64_RXERRDPTR 0x0001FFFF
+#define BCM43xx_DMA64_RXERR 0xF0000000
+#define BCM43xx_DMA64_RXERR_NOERR 0x00000000
+#define BCM43xx_DMA64_RXERR_PROT 0x10000000
+#define BCM43xx_DMA64_RXERR_UNDERRUN 0x20000000
+#define BCM43xx_DMA64_RXERR_TRANSFER 0x30000000
+#define BCM43xx_DMA64_RXERR_DESCREAD 0x40000000
+#define BCM43xx_DMA64_RXERR_CORE 0x50000000
+
+/* 64-bit DMA descriptor. */
+struct bcm43xx_dmadesc64 {
+ __le32 control0;
+ __le32 control1;
+ __le32 address_low;
+ __le32 address_high;
+} __attribute__((__packed__));
+#define BCM43xx_DMA64_DCTL0_DTABLEEND 0x10000000
+#define BCM43xx_DMA64_DCTL0_IRQ 0x20000000
+#define BCM43xx_DMA64_DCTL0_FRAMEEND 0x40000000
+#define BCM43xx_DMA64_DCTL0_FRAMESTART 0x80000000
+#define BCM43xx_DMA64_DCTL1_BYTECNT 0x00001FFF
+#define BCM43xx_DMA64_DCTL1_ADDREXT_MASK 0x00030000
+#define BCM43xx_DMA64_DCTL1_ADDREXT_SHIFT 16
+
+/* Address field Routing value. */
+#define BCM43xx_DMA64_ROUTING 0xC0000000
+#define BCM43xx_DMA64_ROUTING_SHIFT 30
+#define BCM43xx_DMA64_NOTRANS 0x00000000
+#define BCM43xx_DMA64_CLIENTTRANS 0x80000000
+
+
+
+struct bcm43xx_dmadesc_generic {
+ union {
+ struct bcm43xx_dmadesc32 dma32;
+ struct bcm43xx_dmadesc64 dma64;
+ } __attribute__((__packed__));
+} __attribute__((__packed__));
+
/* Misc DMA constants */
#define BCM43xx_DMA_RINGMEMSIZE PAGE_SIZE
-#define BCM43xx_DMA_BUSADDRMAX 0x3FFFFFFF
-#define BCM43xx_DMA_DMABUSADDROFFSET (1 << 30)
-#define BCM43xx_DMA1_RX_FRAMEOFFSET 30
-#define BCM43xx_DMA4_RX_FRAMEOFFSET 0
+#define BCM43xx_DMA0_RX_FRAMEOFFSET 30
+#define BCM43xx_DMA3_RX_FRAMEOFFSET 0
+
/* DMA engine tuning knobs */
#define BCM43xx_TXRING_SLOTS 512
#define BCM43xx_RXRING_SLOTS 64
-#define BCM43xx_DMA1_RXBUFFERSIZE (2304 + 100)
-#define BCM43xx_DMA4_RXBUFFERSIZE 16
+#define BCM43xx_DMA0_RX_BUFFERSIZE (2304 + 100)
+#define BCM43xx_DMA3_RX_BUFFERSIZE 16
/* Suspend the tx queue, if less than this percent slots are free. */
#define BCM43xx_TXSUSPEND_PERCENT 20
/* Resume the tx queue, if more than this percent slots are free. */
@@ -86,17 +202,6 @@
struct bcm43xx_xmitstatus;
-struct bcm43xx_dmadesc {
- __le32 _control;
- __le32 _address;
-} __attribute__((__packed__));
-
-/* Macros to access the bcm43xx_dmadesc struct */
-#define get_desc_ctl(desc) le32_to_cpu((desc)->_control)
-#define set_desc_ctl(desc, ctl) do { (desc)->_control = cpu_to_le32(ctl); } while (0)
-#define get_desc_addr(desc) le32_to_cpu((desc)->_address)
-#define set_desc_addr(desc, addr) do { (desc)->_address = cpu_to_le32(addr); } while (0)
-
struct bcm43xx_dmadesc_meta {
/* The kernel DMA-able buffer. */
struct sk_buff *skb;
@@ -105,15 +210,14 @@
};
struct bcm43xx_dmaring {
- struct bcm43xx_private *bcm;
/* Kernel virtual base address of the ring memory. */
- struct bcm43xx_dmadesc *vbase;
- /* DMA memory offset */
- dma_addr_t memoffset;
- /* (Unadjusted) DMA base bus-address of the ring memory. */
- dma_addr_t dmabase;
+ void *descbase;
/* Meta data about all descriptors. */
struct bcm43xx_dmadesc_meta *meta;
+ /* DMA Routing value. */
+ u32 routing;
+ /* (Unadjusted) DMA base bus-address of the ring memory. */
+ dma_addr_t dmabase;
/* Number of descriptor slots in the ring. */
int nr_slots;
/* Number of used descriptor slots. */
@@ -127,12 +231,14 @@
u32 frameoffset;
/* Descriptor buffer size. */
u16 rx_buffersize;
- /* The MMIO base register of the DMA controller, this
- * ring is posted to.
- */
+ /* The MMIO base register of the DMA controller. */
u16 mmio_base;
+ /* DMA controller index number (0-5). */
+ int index;
u8 tx:1, /* TRUE, if this is a TX ring. */
+ dma64:1, /* TRUE, if 64-bit DMA is enabled (FALSE if 32bit). */
suspended:1; /* TRUE, if transfers are suspended on this ring. */
+ struct bcm43xx_private *bcm;
#ifdef CONFIG_BCM43XX_DEBUG
/* Maximum number of used slots. */
int max_used_slots;
@@ -141,6 +247,34 @@
static inline
+int bcm43xx_dma_desc2idx(struct bcm43xx_dmaring *ring,
+ struct bcm43xx_dmadesc_generic *desc)
+{
+ if (ring->dma64) {
+ struct bcm43xx_dmadesc64 *dd64 = ring->descbase;
+ return (int)(&(desc->dma64) - dd64);
+ } else {
+ struct bcm43xx_dmadesc32 *dd32 = ring->descbase;
+ return (int)(&(desc->dma32) - dd32);
+ }
+}
+
+static inline
+struct bcm43xx_dmadesc_generic * bcm43xx_dma_idx2desc(struct bcm43xx_dmaring *ring,
+ int slot,
+ struct bcm43xx_dmadesc_meta **meta)
+{
+ *meta = &(ring->meta[slot]);
+ if (ring->dma64) {
+ struct bcm43xx_dmadesc64 *dd64 = ring->descbase;
+ return (struct bcm43xx_dmadesc_generic *)(&(dd64[slot]));
+ } else {
+ struct bcm43xx_dmadesc32 *dd32 = ring->descbase;
+ return (struct bcm43xx_dmadesc_generic *)(&(dd32[slot]));
+ }
+}
+
+static inline
u32 bcm43xx_dma_read(struct bcm43xx_dmaring *ring,
u16 offset)
{
@@ -159,9 +293,13 @@
void bcm43xx_dma_free(struct bcm43xx_private *bcm);
int bcm43xx_dmacontroller_rx_reset(struct bcm43xx_private *bcm,
- u16 dmacontroller_mmio_base);
+ u16 dmacontroller_mmio_base,
+ int dma64);
int bcm43xx_dmacontroller_tx_reset(struct bcm43xx_private *bcm,
- u16 dmacontroller_mmio_base);
+ u16 dmacontroller_mmio_base,
+ int dma64);
+
+u16 bcm43xx_dmacontroller_base(int dma64bit, int dmacontroller_idx);
void bcm43xx_dma_tx_suspend(struct bcm43xx_dmaring *ring);
void bcm43xx_dma_tx_resume(struct bcm43xx_dmaring *ring);
@@ -173,7 +311,6 @@
struct ieee80211_txb *txb);
void bcm43xx_dma_rx(struct bcm43xx_dmaring *ring);
-
#else /* CONFIG_BCM43XX_DMA */
@@ -188,13 +325,15 @@
}
static inline
int bcm43xx_dmacontroller_rx_reset(struct bcm43xx_private *bcm,
- u16 dmacontroller_mmio_base)
+ u16 dmacontroller_mmio_base,
+ int dma64)
{
return 0;
}
static inline
int bcm43xx_dmacontroller_tx_reset(struct bcm43xx_private *bcm,
- u16 dmacontroller_mmio_base)
+ u16 dmacontroller_mmio_base,
+ int dma64)
{
return 0;
}
Index: wireless-dev/drivers/net/wireless/bcm43xx/bcm43xx_main.c
===================================================================
--- wireless-dev.orig/drivers/net/wireless/bcm43xx/bcm43xx_main.c 2006-06-17 17:48:26.000000000 +0200
+++ wireless-dev/drivers/net/wireless/bcm43xx/bcm43xx_main.c 2006-06-17 18:15:31.000000000 +0200
@@ -1376,6 +1376,7 @@
if ((bcm43xx_core_enabled(bcm)) &&
!bcm43xx_using_pio(bcm)) {
//FIXME: Do we _really_ want #ifndef CONFIG_BCM947XX here?
+#if 0
#ifndef CONFIG_BCM947XX
/* reset all used DMA controllers. */
bcm43xx_dmacontroller_tx_reset(bcm, BCM43xx_MMIO_DMA1_BASE);
@@ -1386,6 +1387,7 @@
if (bcm->current_core->rev < 5)
bcm43xx_dmacontroller_rx_reset(bcm, BCM43xx_MMIO_DMA4_BASE);
#endif
+#endif
}
if (bcm->shutting_down) {
bcm43xx_write32(bcm, BCM43xx_MMIO_STATUS_BITFIELD,
@@ -1696,8 +1698,9 @@
static void bcm43xx_interrupt_tasklet(struct bcm43xx_private *bcm)
{
u32 reason;
- u32 dma_reason[4];
- int activity = 0;
+ u32 dma_reason[6];
+ u32 merged_dma_reason = 0;
+ int i, activity = 0;
unsigned long flags;
#ifdef CONFIG_BCM43XX_DEBUG
@@ -1709,10 +1712,10 @@
bcm43xx_lock_mmio(bcm, flags);
reason = bcm->irq_reason;
- dma_reason[0] = bcm->dma_reason[0];
- dma_reason[1] = bcm->dma_reason[1];
- dma_reason[2] = bcm->dma_reason[2];
- dma_reason[3] = bcm->dma_reason[3];
+ for (i = 5; i >= 0; i--) {
+ dma_reason[i] = bcm->dma_reason[i];
+ merged_dma_reason |= dma_reason[i];
+ }
if (unlikely(reason & BCM43xx_IRQ_XMIT_ERROR)) {
/* TX error. We get this when Template Ram is written in wrong endianess
@@ -1723,26 +1726,24 @@
printkl(KERN_ERR PFX "FATAL ERROR: BCM43xx_IRQ_XMIT_ERROR\n");
bcmirq_handled(BCM43xx_IRQ_XMIT_ERROR);
}
- if (unlikely((dma_reason[0] & BCM43xx_DMAIRQ_FATALMASK) |
- (dma_reason[1] & BCM43xx_DMAIRQ_FATALMASK) |
- (dma_reason[2] & BCM43xx_DMAIRQ_FATALMASK) |
- (dma_reason[3] & BCM43xx_DMAIRQ_FATALMASK))) {
+ if (unlikely(merged_dma_reason & BCM43xx_DMAIRQ_FATALMASK)) {
printkl(KERN_ERR PFX "FATAL ERROR: Fatal DMA error: "
- "0x%08X, 0x%08X, 0x%08X, 0x%08X\n",
+ "0x%08X, 0x%08X, 0x%08X, "
+ "0x%08X, 0x%08X, 0x%08X\n",
dma_reason[0], dma_reason[1],
- dma_reason[2], dma_reason[3]);
+ dma_reason[2], dma_reason[3],
+ dma_reason[4], dma_reason[5]);
bcm43xx_controller_restart(bcm, "DMA error");
bcm43xx_unlock_mmio(bcm, flags);
return;
}
- if (unlikely((dma_reason[0] & BCM43xx_DMAIRQ_NONFATALMASK) |
- (dma_reason[1] & BCM43xx_DMAIRQ_NONFATALMASK) |
- (dma_reason[2] & BCM43xx_DMAIRQ_NONFATALMASK) |
- (dma_reason[3] & BCM43xx_DMAIRQ_NONFATALMASK))) {
+ if (unlikely(merged_dma_reason & BCM43xx_DMAIRQ_NONFATALMASK)) {
printkl(KERN_ERR PFX "DMA error: "
- "0x%08X, 0x%08X, 0x%08X, 0x%08X\n",
+ "0x%08X, 0x%08X, 0x%08X, "
+ "0x%08X, 0x%08X, 0x%08X\n",
dma_reason[0], dma_reason[1],
- dma_reason[2], dma_reason[3]);
+ dma_reason[2], dma_reason[3],
+ dma_reason[4], dma_reason[5]);
}
if (reason & BCM43xx_IRQ_PS) {
@@ -1777,8 +1778,6 @@
}
/* Check the DMA reason registers for received data. */
- assert(!(dma_reason[1] & BCM43xx_DMAIRQ_RX_DONE));
- assert(!(dma_reason[2] & BCM43xx_DMAIRQ_RX_DONE));
if (dma_reason[0] & BCM43xx_DMAIRQ_RX_DONE) {
if (bcm43xx_using_pio(bcm))
bcm43xx_pio_rx(bcm43xx_current_pio(bcm)->queue0);
@@ -1786,13 +1785,17 @@
bcm43xx_dma_rx(bcm43xx_current_dma(bcm)->rx_ring0);
/* We intentionally don't set "activity" to 1, here. */
}
+ assert(!(dma_reason[1] & BCM43xx_DMAIRQ_RX_DONE));
+ assert(!(dma_reason[2] & BCM43xx_DMAIRQ_RX_DONE));
if (dma_reason[3] & BCM43xx_DMAIRQ_RX_DONE) {
if (bcm43xx_using_pio(bcm))
bcm43xx_pio_rx(bcm43xx_current_pio(bcm)->queue3);
else
- bcm43xx_dma_rx(bcm43xx_current_dma(bcm)->rx_ring1);
+ bcm43xx_dma_rx(bcm43xx_current_dma(bcm)->rx_ring3);
activity = 1;
}
+ assert(!(dma_reason[4] & BCM43xx_DMAIRQ_RX_DONE));
+ assert(!(dma_reason[5] & BCM43xx_DMAIRQ_RX_DONE));
bcmirq_handled(BCM43xx_IRQ_RX);
if (reason & BCM43xx_IRQ_XMIT_STATUS) {
@@ -1848,14 +1851,18 @@
bcm43xx_write32(bcm, BCM43xx_MMIO_GEN_IRQ_REASON, reason);
- bcm43xx_write32(bcm, BCM43xx_MMIO_DMA1_REASON,
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA0_REASON,
bcm->dma_reason[0]);
- bcm43xx_write32(bcm, BCM43xx_MMIO_DMA2_REASON,
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA1_REASON,
bcm->dma_reason[1]);
- bcm43xx_write32(bcm, BCM43xx_MMIO_DMA3_REASON,
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA2_REASON,
bcm->dma_reason[2]);
- bcm43xx_write32(bcm, BCM43xx_MMIO_DMA4_REASON,
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA3_REASON,
bcm->dma_reason[3]);
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA4_REASON,
+ bcm->dma_reason[4]);
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA5_REASON,
+ bcm->dma_reason[5]);
}
/* Interrupt handler top-half */
@@ -1880,14 +1887,18 @@
if (!reason)
goto out;
- bcm->dma_reason[0] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA1_REASON)
- & 0x0001dc00;
- bcm->dma_reason[1] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA2_REASON)
- & 0x0000dc00;
- bcm->dma_reason[2] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA3_REASON)
- & 0x0000dc00;
- bcm->dma_reason[3] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA4_REASON)
- & 0x0001dc00;
+ bcm->dma_reason[0] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA0_REASON)
+ & 0x0001DC00;
+ bcm->dma_reason[1] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA1_REASON)
+ & 0x0000DC00;
+ bcm->dma_reason[2] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA2_REASON)
+ & 0x0000DC00;
+ bcm->dma_reason[3] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA3_REASON)
+ & 0x0001DC00;
+ bcm->dma_reason[4] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA4_REASON)
+ & 0x0000DC00;
+ bcm->dma_reason[5] = bcm43xx_read32(bcm, BCM43xx_MMIO_DMA5_REASON)
+ & 0x0000DC00;
bcm43xx_interrupt_ack(bcm, reason);
@@ -2462,10 +2473,12 @@
bcm43xx_write32(bcm, 0x018C, 0x02000000);
}
bcm43xx_write32(bcm, BCM43xx_MMIO_GEN_IRQ_REASON, 0x00004000);
- bcm43xx_write32(bcm, BCM43xx_MMIO_DMA1_IRQ_MASK, 0x0001DC00);
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA0_IRQ_MASK, 0x0001DC00);
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA1_IRQ_MASK, 0x0000DC00);
bcm43xx_write32(bcm, BCM43xx_MMIO_DMA2_IRQ_MASK, 0x0000DC00);
- bcm43xx_write32(bcm, BCM43xx_MMIO_DMA3_IRQ_MASK, 0x0000DC00);
- bcm43xx_write32(bcm, BCM43xx_MMIO_DMA4_IRQ_MASK, 0x0001DC00);
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA3_IRQ_MASK, 0x0001DC00);
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA4_IRQ_MASK, 0x0000DC00);
+ bcm43xx_write32(bcm, BCM43xx_MMIO_DMA5_IRQ_MASK, 0x0000DC00);
value32 = bcm43xx_read32(bcm, BCM43xx_CIR_SBTMSTATELOW);
value32 |= 0x00100000;
--
Greetings Michael.
^ permalink raw reply
* [IOC3] IP27: Really set PCI64_ATTR_VIRTUAL, not PCI64_ATTR_PREC.
From: Ralf Baechle @ 2006-06-17 17:57 UTC (permalink / raw)
To: netdev, Jeff Garzik
IOC3's homegrown DMA mapping functions that are used to optimize things
a little on IP27 set the wrong bit.
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
diff --git a/drivers/net/ioc3-eth.c b/drivers/net/ioc3-eth.c
index ae71ed5..e76e6e7 100644
--- a/drivers/net/ioc3-eth.c
+++ b/drivers/net/ioc3-eth.c
@@ -145,7 +145,7 @@ static inline struct sk_buff * ioc3_allo
static inline unsigned long ioc3_map(void *ptr, unsigned long vdev)
{
#ifdef CONFIG_SGI_IP27
- vdev <<= 58; /* Shift to PCI64_ATTR_VIRTUAL */
+ vdev <<= 57; /* Shift to PCI64_ATTR_VIRTUAL */
return vdev | (0xaUL << PCI64_ATTR_TARG_SHFT) | PCI64_ATTR_PREF |
((unsigned long)ptr & TO_PHYS_MASK);
^ permalink raw reply related
* Re: tg3 timeouts with 2.6.17-rc6
From: Michael Chan @ 2006-06-17 18:23 UTC (permalink / raw)
To: David Miller; +Cc: jk, linux-kernel, netdev
In-Reply-To: <1551EAE59135BE47B544934E30FC4FC041BD1E@NT-IRVA-0751.brcm.ad.broadcom.com>
On Fri, 2006-06-16 at 20:37 -0700, Michael Chan wrote:
> David Miller wrote:
>
> > From: "Michael Chan" <mchan@broadcom.com>
> > Date: Fri, 16 Jun 2006 18:27:32 -0700
> >
> > > In the meantime, I wonder if we should disable TSO by default on the
> > > 5780 chip for 2.6.17.
> >
> > Sounds reasonable. Would we disable it for all chips that set
> > TG3_FLG2_5780_CLASS or a specific variant?
> >
> Yes, let's disable it for all TG3_FLG2_5780_CLASS chips for now
> until we figure out what's going on.
David, Here's the patch if you haven't already made one:
[TG3]: Disable TSO by default on 5780 class chips.
Disable TSO by default on 5780, 5714, and 5715 chips for now while we
investigate the reported tx timeouts by Juergen Kreileder. Thanks to
Juergen for reporting the problem.
Update version to 3.60.
Signed-off-by: Michael Chan <mchan@broadcom.com>
diff --git a/drivers/net/tg3.c b/drivers/net/tg3.c
index 862c226..607d87e 100644
--- a/drivers/net/tg3.c
+++ b/drivers/net/tg3.c
@@ -69,8 +69,8 @@
#define DRV_MODULE_NAME "tg3"
#define PFX DRV_MODULE_NAME ": "
-#define DRV_MODULE_VERSION "3.59"
-#define DRV_MODULE_RELDATE "June 8, 2006"
+#define DRV_MODULE_VERSION "3.60"
+#define DRV_MODULE_RELDATE "June 17, 2006"
#define TG3_DEF_MAC_MODE 0
#define TG3_DEF_RX_MODE 0
@@ -11385,7 +11385,11 @@ static int __devinit tg3_init_one(struct
* Firmware TSO on older chips gives lower performance, so it
* is off by default, but can be enabled using ethtool.
*/
- if (tp->tg3_flags2 & TG3_FLG2_HW_TSO)
+ /* Disable TSO by default on all 5780 class chips because
+ * of reported tx timeouts.
+ */
+ if ((tp->tg3_flags2 & TG3_FLG2_HW_TSO) &&
+ !(tp->tg3_flags2 & TG3_FLG2_5780_CLASS))
dev->features |= NETIF_F_TSO;
#endif
^ permalink raw reply related
* Dear Friend
From: Duoala Mbale @ 2006-06-17 19:03 UTC (permalink / raw)
To: netdev
My Dear Friend,
I am interested in a partnership investment programme with your
self/corporation.
There is this huge amount of Five million Seven hundred and fifthy
thousand U.S dollars($5,750,000.00)which my late father kept in a
security/financial company before he was assasinated by unknown persons,
during a political crisis in my country.
My father deposited it as a family treasure and that was before his
death.
Now I and mother left Angola to Senegal,through the help of my late
father's good friend.Right now we are in refugee camp and
wehave decided to invest these money in your country or anywhere safe enough outside my country Angola and the whole of Africa for security and political reasons.
We would want you to assist us to transfer this fund to your country
for safty and investment purposes on the followings below:
1).Telecommunication
2).Manufacturing
3).Real Estate Business
If you will be able of rendering an assistance to us we will
adequately compensate you 25% Of the total fund.
We will arrange all the necessary procedures in ensuring a smooth
process for the funds to get to you.We will also appreciate if you contact
me once you receive this mail to enable me give you more details.This
matter requires your urgent attention ,confidentiality and discretion no
matter what your decision maybe.
Thank you and God bless you.
Your's Sincerely,
Duoala Mbale
^ permalink raw reply
* recommend a network card
From: bastard operater @ 2006-06-17 21:12 UTC (permalink / raw)
To: netdev
Hello,
My 3Com 3C905B-TX is starting to malfunction and it is getting worse. So I
am in the market for a new NIC. The card has to work on Linux (natively)
and windows. I thought I would ask the experts if you have any
recommendations for a good 100MB PCI card?
Thanks,
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox