Patchwork [U-Boot] powerpc/85xx: fix null pointer dereference when initializing the SGMII TBI PHY

login
register
mail settings
Submitter Timur Tabi
Date Oct. 4, 2011, 9:44 p.m.
Message ID <1317764683-29426-1-git-send-email-timur@freescale.com>
Download mbox | patch
Permalink /patch/117703/
State Accepted
Commit 30381716284b688cb1b4e315aa6b8ef7422fa172
Delegated to: Kumar Gala
Headers show

Comments

Timur Tabi - Oct. 4, 2011, 9:44 p.m.
Function dtsec_configure_serdes() needs to know where the TBI PHY registers
are in order to configure SGMII for proper SerDes operation.

During SGMII initialzation, fm_eth_init_mac() passing NULL for 'phyregs' when
it called init_dtsec(), because it was believed that phyregs was not used.
In fact, it is used by dtsec_configure_serdes() to configure the TBI PHY
registers.

We also need to define the PHY registers in struct fm_mdio.

Signed-off-by: Timur Tabi <timur@freescale.com>
---

Note: I'm not sure of my change to struct fm_mdio.  It works, but it doesn't
smell right.

 arch/powerpc/include/asm/fsl_fman.h |    9 ++++++++-
 drivers/net/fm/dtsec.c              |    2 +-
 drivers/net/fm/eth.c                |    3 ++-
 3 files changed, 11 insertions(+), 3 deletions(-)
Kumar Gala - Oct. 4, 2011, 9:50 p.m.
On Oct 4, 2011, at 4:44 PM, Timur Tabi wrote:

> Function dtsec_configure_serdes() needs to know where the TBI PHY registers
> are in order to configure SGMII for proper SerDes operation.
> 
> During SGMII initialzation, fm_eth_init_mac() passing NULL for 'phyregs' when
> it called init_dtsec(), because it was believed that phyregs was not used.
> In fact, it is used by dtsec_configure_serdes() to configure the TBI PHY
> registers.
> 
> We also need to define the PHY registers in struct fm_mdio.
> 
> Signed-off-by: Timur Tabi <timur@freescale.com>
> ---
> 
> Note: I'm not sure of my change to struct fm_mdio.  It works, but it doesn't
> smell right.

What exactly is the concern?

- k
Tabi Timur-B04825 - Oct. 4, 2011, 9:52 p.m.
Kumar Gala wrote:
>> >  Note: I'm not sure of my change to struct fm_mdio.  It works, but it doesn't
>> >  smell right.

> What exactly is the concern?

Nothing specific, just a gut feeling.
Kumar Gala - Oct. 7, 2011, 3:06 p.m.
On Oct 4, 2011, at 4:44 PM, Timur Tabi wrote:

> Function dtsec_configure_serdes() needs to know where the TBI PHY registers
> are in order to configure SGMII for proper SerDes operation.
> 
> During SGMII initialzation, fm_eth_init_mac() passing NULL for 'phyregs' when
> it called init_dtsec(), because it was believed that phyregs was not used.
> In fact, it is used by dtsec_configure_serdes() to configure the TBI PHY
> registers.
> 
> We also need to define the PHY registers in struct fm_mdio.
> 
> Signed-off-by: Timur Tabi <timur@freescale.com>
> ---
> 
> Note: I'm not sure of my change to struct fm_mdio.  It works, but it doesn't
> smell right.
> 
> arch/powerpc/include/asm/fsl_fman.h |    9 ++++++++-
> drivers/net/fm/dtsec.c              |    2 +-
> drivers/net/fm/eth.c                |    3 ++-
> 3 files changed, 11 insertions(+), 3 deletions(-)

applied to 85xx

- k

Patch

diff --git a/arch/powerpc/include/asm/fsl_fman.h b/arch/powerpc/include/asm/fsl_fman.h
index fddc0cc..2c0c9bc 100644
--- a/arch/powerpc/include/asm/fsl_fman.h
+++ b/arch/powerpc/include/asm/fsl_fman.h
@@ -405,7 +405,14 @@  typedef struct fm_dtesc {
 } fm_dtsec_t;
 
 typedef struct fm_mdio {
-	u8	res[4*1024];
+	u8	res0[0x120];
+	u32	miimcfg;	/* MII management configuration reg */
+	u32	miimcom;	/* MII management command reg */
+	u32	miimadd;	/* MII management address reg */
+	u32	miimcon;	/* MII management control reg */
+	u32	miimstat;	/* MII management status reg  */
+	u32	miimind;	/* MII management indication reg */
+	u8	res1[0x1000 - 0x138];
 } fm_mdio_t;
 
 typedef struct fm_10gec {
diff --git a/drivers/net/fm/dtsec.c b/drivers/net/fm/dtsec.c
index a77ee20..7dd78f2 100644
--- a/drivers/net/fm/dtsec.c
+++ b/drivers/net/fm/dtsec.c
@@ -171,7 +171,7 @@  void init_dtsec(struct fsl_enet_mac *mac, void *base,
 		void *phyregs, int max_rx_len)
 {
 	mac->base = base;
-	mac->phyregs = NULL;
+	mac->phyregs = phyregs;
 	mac->max_rx_len = max_rx_len;
 	mac->init_mac = dtsec_init_mac;
 	mac->enable_mac = dtsec_enable_mac;
diff --git a/drivers/net/fm/eth.c b/drivers/net/fm/eth.c
index 308d610..f7ed850 100644
--- a/drivers/net/fm/eth.c
+++ b/drivers/net/fm/eth.c
@@ -537,6 +537,7 @@  static int fm_eth_init_mac(struct fm_eth *fm_eth, struct ccsr_fman *reg)
 	/* Get the mac registers base address */
 	if (fm_eth->type == FM_ETH_1G_E) {
 		base = &reg->mac_1g[num].fm_dtesc;
+		phyregs = &reg->mac_1g[num].fm_mdio.miimcfg;
 	} else {
 		base = &reg->mac_10g[num].fm_10gec;
 		phyregs = &reg->mac_10g[num].fm_10gec_mdio;
@@ -552,7 +553,7 @@  static int fm_eth_init_mac(struct fm_eth *fm_eth, struct ccsr_fman *reg)
 	fm_eth->mac = mac;
 
 	if (fm_eth->type == FM_ETH_1G_E)
-		init_dtsec(mac, base, NULL, MAX_RXBUF_LEN);
+		init_dtsec(mac, base, phyregs, MAX_RXBUF_LEN);
 	else
 		init_tgec(mac, base, phyregs, MAX_RXBUF_LEN);