Skip to contentSkip to conversation

rtl8169: fix cache coherency problem

PATCH New Yoshihiro Shimoda 4 participants 6 messages Feb 25, 2009 – Apr 27, 2009 Network drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/2 rtl8169: fix cache coherency problem

Fix the problem that cannot access actual data when CPU data cache enabled.

---
 drivers/net/rtl8169.c |   11 ++++++++++-
 1 files changed, 10 insertions(+), 1 deletions(-)
drivers/net/rtl8169.c +10 -1
index e9f6391..b441ec4 100644
@@ -420,6 +420,8 @@ static int rtl_recv(struct eth_device *dev)
 	ioaddr = dev->iobase;
 
 	cur_rx = tpc->cur_rx;
+	flush_cache((unsigned long)&tpc->RxDescArray[cur_rx],
+			sizeof(struct RxDesc));
 	if ((le32_to_cpu(tpc->RxDescArray[cur_rx].status) & OWNbit) == 0) {
 		if (!(le32_to_cpu(tpc->RxDescArray[cur_rx].status) & RxRES)) {
 			unsigned char rxdata[RX_BUF_LEN];
@@ -437,6 +439,8 @@ static int rtl_recv(struct eth_device *dev)
 					cpu_to_le32(OWNbit + RX_BUF_SIZE);
 			tpc->RxDescArray[cur_rx].buf_addr =
 				cpu_to_le32((unsigned long)tpc->RxBufferRing[cur_rx]);
+			flush_cache((unsigned long)tpc->RxBufferRing[cur_rx],
+					RX_BUF_SIZE);
 		} else {
 			puts("Error Rx");
 		}
@@ -478,6 +482,7 @@ static int rtl_send(struct eth_device *dev, volatile void *packet, int length)
 	/* point to the current txb incase multiple tx_rings are used */
 	ptxb = tpc->Tx_skbuff[entry * MAX_ETH_FRAME_SIZE];
 	memcpy(ptxb, (char *)packet, (int)length);
+	flush_cache((unsigned long)ptxb, length);
 
 	while (len < ETH_ZLEN)
 		ptxb[len++] = '\0';
@@ -497,7 +502,10 @@ static int rtl_send(struct eth_device *dev, volatile void *packet, int length)
 
 	tpc->cur_tx++;
 	to = currticks() + TX_TIMEOUT;
-	while ((le32_to_cpu(tpc->TxDescArray[entry].status) & OWNbit)
+	do {
+		flush_cache((unsigned long)&tpc->TxDescArray[entry],
+				sizeof(struct TxDesc));
+	} while ((le32_to_cpu(tpc->TxDescArray[entry].status) & OWNbit)
 				&& (currticks() < to));	/* wait */
 
 	if (currticks() >= to) {
@@ -639,6 +647,7 @@ static void rtl8169_init_ring(struct eth_device *dev)
 		tpc->RxBufferRing[i] = &rxb[i * RX_BUF_SIZE];
 		tpc->RxDescArray[i].buf_addr =
 			cpu_to_le32((unsigned long)tpc->RxBufferRing[i]);
+		flush_cache((unsigned long)tpc->RxBufferRing[i], RX_BUF_SIZE);
 	}
 
 #ifdef DEBUG_RTL8169
Signature · 1 line
1.5.5
#1 lore raw

PATCH 1/2 rtl8169: fix cache coherency problem

Dear Ben,

Quoted 6 lines · In message <[email protected]> Yoshihiro Shimoda wrote:

Fix the problem that cannot access actual data when CPU data cache enabled.

---
 drivers/net/rtl8169.c |   11 ++++++++++-
 1 files changed, 10 insertions(+), 1 deletions(-)

I did not see any comments on this?

Ditto for the other patch:

http://article.gmane.org/gmane.comp.boot-loaders.u-boot/55286

?

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Do not follow where the path may lead....go instead where there is no
path and leave a trail.
#2 lore raw

PATCH 1/2 rtl8169: fix cache coherency problem

Quoted 10 lines · Wolfgang Denk

Dear Ben,

Quoted 6 lines · In message <[email protected]> Yoshihiro Shimoda wrote:

Fix the problem that cannot access actual data when CPU data cache enabled.

---
 drivers/net/rtl8169.c |   11 ++++++++++-
 1 files changed, 10 insertions(+), 1 deletions(-)
Quoted 6 lines

I did not see any comments on this?

Ditto for the other patch:

http://article.gmane.org/gmane.comp.boot-loaders.u-boot/55286

I checked and tested these. Work file.

Best regards,
Nobuhiro

#3 lore raw

PATCH 1/2 rtl8169: fix cache coherency problem

In reply to Wolfgang Denk

Dear Wolfgang Denk,

Quoted 10 lines · In message <[email protected]> I wrote:

Dear Ben,

Quoted 6 lines · In message <[email protected]> Yoshihiro Shimoda wrote:

Fix the problem that cannot access actual data when CPU data cache enabled.

---
 drivers/net/rtl8169.c |   11 ++++++++++-
 1 files changed, 10 insertions(+), 1 deletions(-)
Quoted 7 lines

I did not see any comments on this?

Ditto for the other patch:

http://article.gmane.org/gmane.comp.boot-loaders.u-boot/55286

?

Any comments?

If you don't have time, I can apply this directly.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Well, the way I see it, logic is only a way of being ignorant by num-
bers.                                 - Terry Pratchett, _Small Gods_
#4 lore raw

PATCH 1/2 rtl8169: fix cache coherency problem

Hi Wolfgang,

Quoted 24 lines · Wolfgang Denk

Dear Wolfgang Denk,

Quoted 11 lines · In message <[email protected]> I wrote:

Dear Ben,

Quoted 7 lines · In message <[email protected]> Yoshihiro Shimoda wrote:

Fix the problem that cannot access actual data when CPU data cache enabled.

---
 drivers/net/rtl8169.c |   11 ++++++++++-
 1 files changed, 10 insertions(+), 1 deletions(-)
Quoted 8 lines

I did not see any comments on this?

Ditto for the other patch:

http://article.gmane.org/gmane.comp.boot-loaders.u-boot/55286

?

Quoted 9 lines

Any comments?

If you don't have time, I can apply this directly.

Best regards,

Wolfgang Denk

This one also looks fine. Please apply directly.

thanks,
Ben

#5 lore raw

PATCH 1/2 rtl8169: fix cache coherency problem

In reply to Yoshihiro Shimoda

Dear Yoshihiro Shimoda,

Quoted 6 lines · In message <[email protected]> you wrote:

Fix the problem that cannot access actual data when CPU data cache enabled.

---
 drivers/net/rtl8169.c |   11 ++++++++++-
 1 files changed, 10 insertions(+), 1 deletions(-)

Applied, thanks.

Best regards,

Wolfgang Denk

Signature · 5 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
I express preference for a chronological  sequence  of  events  which
precludes a violence.   - Terry Pratchett, _The Dark Side of the Sun_