Skip to contentSkip to conversation

MIPS Malta board support

PATCH New Paul Burton 3 participants 44 messages Nov 8, 2013 – Nov 11, 2013

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

cover MIPS Malta board support

This series adds support for booting on a physical MIPS Malta board using a coreFPGA6 core card.

The first 6 patches lay some groundwork, then the next 8 genericise the existing qemu-malta board to also function on a physical Malta.

In the final patch I stake my claim upon, errm I mean step up to support :), the malta(el) board(s).

Since the original posting this series has been rebased atop current master & adapted to the KBuild changes. I also included the RTC support patch I previously posted separately, and a couple of extra patches which fix up issues with booted kernels & add environment support.

After this series Gabor Juhos' "malta: define CONFIG_MEMSIZE_IN_BYTES" patch is also required in order for a booted Linux kernel to function correctly.

Paul Burton (18):
mips32: detect L1 cache sizes if they're not defined
pcnet: code style cleanup
pcnet: s/le16_to_cpu/cpu_to_le16/ in pcnet_send
pcnet: add cache flushing & invalidation
pcnet: enable the NOUFLO feature
pci.h: allow inclusion in assembly source
qemu-malta: rename to just "malta"
malta: setup super I/O UARTs
malta: support for coreFPGA6 boards
malta: display "U-boot" on the LCD screen
malta: enable CONFIG_PCNET_79C973, PCNET_HAS_PROM, CONFIG_CMD_DHCP
malta: remove cache size definitions
malta: disable L2 caches
malta: enable RTC support
malta: store environment in flash
malta: setup PIIX4 interrupt route
malta: add script & instructions to flash U-boot
malta: add myself to maintainers

 arch/mips/cpu/mips32/cache.S            |  90 +++++++++--
 arch/mips/cpu/mips32/cpu.c              |  73 ++++++++-
 arch/mips/cpu/mips32/start.S            |   2 +-
 arch/mips/include/asm/malta.h           |  50 +++++-
 arch/mips/include/asm/mipsregs.h        |   6 +
 arch/mips/lib/bootm.c                   |  12 +-
 board/imgtec/malta/Makefile             |  10 ++
 board/imgtec/malta/flash-malta-boot.tcl |  40 +++++
 board/imgtec/malta/lowlevel_init.S      | 238 +++++++++++++++++++++++++++
 board/imgtec/malta/malta.c              | 208 ++++++++++++++++++++++++
 board/imgtec/malta/superio.c            |  63 ++++++++
 board/imgtec/malta/superio.h            |  15 ++
 board/qemu-malta/Makefile               |   9 --
 board/qemu-malta/lowlevel_init.S        |  69 --------
 board/qemu-malta/qemu-malta.c           |  47 ------
 boards.cfg                              |   4 +-
 doc/README.malta                        |  16 ++
 drivers/net/pcnet.c                     | 279 ++++++++++++++++++--------------
 drivers/pci/Makefile                    |   1 +
 drivers/pci/pci_msc01.c                 | 125 ++++++++++++++
 drivers/rtc/mc146818.c                  |   2 +-
 include/configs/malta.h                 | 122 ++++++++++++++
 include/configs/qemu-malta.h            | 113 -------------
 include/msc01.h                         | 135 ++++++++++++++++
 include/pci.h                           |   6 +-
 include/pci_ids.h                       |   3 +
 include/pci_msc01.h                     |  17 ++
 27 files changed, 1356 insertions(+), 399 deletions(-)
 create mode 100644 board/imgtec/malta/Makefile
 create mode 100644 board/imgtec/malta/flash-malta-boot.tcl
 create mode 100644 board/imgtec/malta/lowlevel_init.S
 create mode 100644 board/imgtec/malta/malta.c
 create mode 100644 board/imgtec/malta/superio.c
 create mode 100644 board/imgtec/malta/superio.h
 delete mode 100644 board/qemu-malta/Makefile
 delete mode 100644 board/qemu-malta/lowlevel_init.S
 delete mode 100644 board/qemu-malta/qemu-malta.c
 create mode 100644 doc/README.malta
 create mode 100644 drivers/pci/pci_msc01.c
 create mode 100644 include/configs/malta.h
 delete mode 100644 include/configs/qemu-malta.h
 create mode 100644 include/msc01.h
 create mode 100644 include/pci_msc01.h
Signature · 1 line
1.8.4.1
#1 lore raw

PATCH v2 1/18 mips32: detect L1 cache sizes if they're not defined

For boards such as the MIPS Malta with an FPGA core card it is desirable to be able to detect the L1 cache sizes at runtime, since they are not dependant upon the board but on the FPGA bitstream in use. This patch performs that detection when the CONFIG_SYS_[DI]CACHE_SIZE macros are not defined by the board configuration. In cases where the sizes are detected this patch also removes the restriction that the I-cache & D-cache line sizes must be the same, as this is not necessarily true.

If the cache sizes are defined by a configuration then they will be hardcoded as before, so this patch will not add overhead to such boards.

---
Changes in v2:
- rebase atop master

---
 arch/mips/cpu/mips32/cache.S     | 90 ++++++++++++++++++++++++++++++++++------
 arch/mips/cpu/mips32/cpu.c       | 73 +++++++++++++++++++++++++++++---
 arch/mips/include/asm/mipsregs.h |  6 +++
 3 files changed, 150 insertions(+), 19 deletions(-)
arch/mips/cpu/mips32/cache.S +77 -13
index 12f656c..22bd844 100644
@@ -20,15 +20,6 @@
 
 #define RA		t9
 
-/*
- * 16kB is the maximum size of instruction and data caches on MIPS 4K,
- * 64kB is on 4KE, 24K, 5K, etc. Set bigger size for convenience.
- *
- * Note that the above size is the maximum size of primary cache. U-Boot
- * doesn't have L2 cache support for now.
- */
-#define MIPS_MAX_CACHE_SIZE	0x10000
-
 #define INDEX_BASE	CKSEG0
 
 	.macro	cache_op op addr
@@ -126,12 +117,85 @@ LEAF(mips_init_dcache)
  */
 NESTED(mips_cache_reset, 0, ra)
 	move	RA, ra
-	li	t2, CONFIG_SYS_ICACHE_SIZE
-	li	t3, CONFIG_SYS_DCACHE_SIZE
+
+#if !defined(CONFIG_SYS_ICACHE_SIZE) || !defined(CONFIG_SYS_DCACHE_SIZE) || \
+    !defined(CONFIG_SYS_CACHELINE_SIZE)
+	/* read Config1 for use below */
+	mfc0	t5, CP0_CONFIG, 1
+#endif
+
+#ifdef CONFIG_SYS_CACHELINE_SIZE
+	li	t7, CONFIG_SYS_CACHELINE_SIZE
 	li	t8, CONFIG_SYS_CACHELINE_SIZE
+#else
+	/* Detect I-cache line size. */
+	srl	t8, t5, MIPS_CONF1_IL_SHIFT
+	andi	t8, t8, (MIPS_CONF1_IL >> MIPS_CONF1_IL_SHIFT)
+	beqz	t8, 1f
+	li	t6, 2
+	sllv	t8, t6, t8
 
-	li	v0, MIPS_MAX_CACHE_SIZE
+1:	/* Detect D-cache line size. */
+	srl	t7, t5, MIPS_CONF1_DL_SHIFT
+	andi	t7, t7, (MIPS_CONF1_DL >> MIPS_CONF1_DL_SHIFT)
+	beqz	t7, 1f
+	li	t6, 2
+	sllv	t7, t6, t7
+1:
+#endif
 
+#ifdef CONFIG_SYS_ICACHE_SIZE
+	li	t2, CONFIG_SYS_ICACHE_SIZE
+#else
+	/* Detect I-cache size. */
+	srl	t6, t5, MIPS_CONF1_IS_SHIFT
+	andi	t6, t6, (MIPS_CONF1_IS >> MIPS_CONF1_IS_SHIFT)
+	li	t4, 32
+	xori	t2, t6, 0x7
+	beqz	t2, 1f
+	addi	t6, t6, 1
+	sllv	t4, t4, t6
+1:	/* At this point t4 == I-cache sets. */
+	mul	t2, t4, t8
+	srl	t6, t5, MIPS_CONF1_IA_SHIFT
+	andi	t6, t6, (MIPS_CONF1_IA >> MIPS_CONF1_IA_SHIFT)
+	addi	t6, t6, 1
+	/* At this point t6 == I-cache ways. */
+	mul	t2, t2, t6
+#endif
+
+#ifdef CONFIG_SYS_DCACHE_SIZE
+	li	t3, CONFIG_SYS_DCACHE_SIZE
+#else
+	/* Detect D-cache size. */
+	srl	t6, t5, MIPS_CONF1_DS_SHIFT
+	andi	t6, t6, (MIPS_CONF1_DS >> MIPS_CONF1_DS_SHIFT)
+	li	t4, 32
+	xori	t3, t6, 0x7
+	beqz	t3, 1f
+	addi	t6, t6, 1
+	sllv	t4, t4, t6
+1:	/* At this point t4 == I-cache sets. */
+	mul	t3, t4, t7
+	srl	t6, t5, MIPS_CONF1_DA_SHIFT
+	andi	t6, t6, (MIPS_CONF1_DA >> MIPS_CONF1_DA_SHIFT)
+	addi	t6, t6, 1
+	/* At this point t6 == I-cache ways. */
+	mul	t3, t3, t6
+#endif
+
+	/* Determine the largest L1 cache size */
+#if defined(CONFIG_SYS_ICACHE_SIZE) && defined(CONFIG_SYS_DCACHE_SIZE)
+#if CONFIG_SYS_ICACHE_SIZE > CONFIG_SYS_DCACHE_SIZE
+	li	v0, CONFIG_SYS_ICACHE_SIZE
+#else
+	li	v0, CONFIG_SYS_DCACHE_SIZE
+#endif
+#else
+	move	v0, t2
+	sltu	t1, t2, t3
+	movn	v0, t3, t1
+#endif
 	/*
 	 * Now clear that much memory starting from zero.
 	 */
@@ -163,7 +227,7 @@ NESTED(mips_cache_reset, 0, ra)
 	 * then initialize D-cache.
 	 */
 	move	a1, t3
-	move	a2, t8
+	move	a2, t7
 	PTR_LA	v1, mips_init_dcache
 	jalr	v1
 
arch/mips/cpu/mips32/cpu.c +67 -6

102 lines not on this page · raw message

arch/mips/include/asm/mipsregs.h +6 -0
index be7e5c6..3571e4f 100644
@@ -494,11 +494,17 @@
 #define MIPS_CONF1_PC		(_ULCAST_(1) <<  4)
 #define MIPS_CONF1_MD		(_ULCAST_(1) <<  5)
 #define MIPS_CONF1_C2		(_ULCAST_(1) <<  6)
+#define MIPS_CONF1_DA_SHIFT	7
 #define MIPS_CONF1_DA		(_ULCAST_(7) <<  7)
+#define MIPS_CONF1_DL_SHIFT	10
 #define MIPS_CONF1_DL		(_ULCAST_(7) << 10)
+#define MIPS_CONF1_DS_SHIFT	13
 #define MIPS_CONF1_DS		(_ULCAST_(7) << 13)
+#define MIPS_CONF1_IA_SHIFT	16
 #define MIPS_CONF1_IA		(_ULCAST_(7) << 16)
+#define MIPS_CONF1_IL_SHIFT	19
 #define MIPS_CONF1_IL		(_ULCAST_(7) << 19)
+#define MIPS_CONF1_IS_SHIFT	22
 #define MIPS_CONF1_IS		(_ULCAST_(7) << 22)
 #define MIPS_CONF1_TLBS		(_ULCAST_(63)<< 25)
 
Signature · 1 line
1.8.4.1
#2 lore raw

PATCH v2 2/18 pcnet: code style cleanup

In reply to Paul Burton

Fix up the code to match Documentation/CodingStyle. This is mostly removing extraneous spaces.

No functional change is intended.

---
Changes in v2:
- rebase atop master

---
 drivers/net/pcnet.c | 248 ++++++++++++++++++++++++++--------------------------
 1 file changed, 123 insertions(+), 125 deletions(-)
drivers/net/pcnet.c +123 -125
index 283cb48..a30a0bc 100644
@@ -89,39 +89,39 @@ static pcnet_priv_t *lp;
 #define PCNET_RESET		0x14
 #define PCNET_BDP		0x16
 
-static u16 pcnet_read_csr (struct eth_device *dev, int index)
+static u16 pcnet_read_csr(struct eth_device *dev, int index)
 {
-	outw (index, dev->iobase + PCNET_RAP);
-	return inw (dev->iobase + PCNET_RDP);
+	outw(index, dev->iobase + PCNET_RAP);
+	return inw(dev->iobase + PCNET_RDP);
 }
 
-static void pcnet_write_csr (struct eth_device *dev, int index, u16 val)
+static void pcnet_write_csr(struct eth_device *dev, int index, u16 val)
 {
-	outw (index, dev->iobase + PCNET_RAP);
-	outw (val, dev->iobase + PCNET_RDP);
+	outw(index, dev->iobase + PCNET_RAP);
+	outw(val, dev->iobase + PCNET_RDP);
 }
 
-static u16 pcnet_read_bcr (struct eth_device *dev, int index)
+static u16 pcnet_read_bcr(struct eth_device *dev, int index)
 {
-	outw (index, dev->iobase + PCNET_RAP);
-	return inw (dev->iobase + PCNET_BDP);
+	outw(index, dev->iobase + PCNET_RAP);
+	return inw(dev->iobase + PCNET_BDP);
 }
 
-static void pcnet_write_bcr (struct eth_device *dev, int index, u16 val)
+static void pcnet_write_bcr(struct eth_device *dev, int index, u16 val)
 {
-	outw (index, dev->iobase + PCNET_RAP);
-	outw (val, dev->iobase + PCNET_BDP);
+	outw(index, dev->iobase + PCNET_RAP);
+	outw(val, dev->iobase + PCNET_BDP);
 }
 
-static void pcnet_reset (struct eth_device *dev)
+static void pcnet_reset(struct eth_device *dev)
 {
-	inw (dev->iobase + PCNET_RESET);
+	inw(dev->iobase + PCNET_RESET);
 }
 
-static int pcnet_check (struct eth_device *dev)
+static int pcnet_check(struct eth_device *dev)
 {
-	outw (88, dev->iobase + PCNET_RAP);
-	return (inw (dev->iobase + PCNET_RAP) == 88);
+	outw(88, dev->iobase + PCNET_RAP);
+	return inw(dev->iobase + PCNET_RAP) == 88;
 }
 
 static int pcnet_init (struct eth_device *dev, bd_t * bis);
@@ -139,63 +139,64 @@ static struct pci_device_id supported[] = {
 };
 
 
-int pcnet_initialize (bd_t * bis)
+int pcnet_initialize(bd_t *bis)
 {
 	pci_dev_t devbusfn;
 	struct eth_device *dev;
 	u16 command, status;
 	int dev_nr = 0;
 
-	PCNET_DEBUG1 ("\npcnet_initialize...\n");
+	PCNET_DEBUG1("\npcnet_initialize...\n");
 
 	for (dev_nr = 0;; dev_nr++) {
 
 		/*
 		 * Find the PCnet PCI device(s).
 		 */
-		if ((devbusfn = pci_find_devices (supported, dev_nr)) < 0) {
+		devbusfn = pci_find_devices(supported, dev_nr);
+		if (devbusfn < 0)
 			break;
-		}
 
 		/*
 		 * Allocate and pre-fill the device structure.
 		 */
-		dev = (struct eth_device *) malloc (sizeof *dev);
+		dev = (struct eth_device *)malloc(sizeof(*dev));
 		if (!dev) {
 			printf("pcnet: Can not allocate memory\n");
 			break;
 		}
 		memset(dev, 0, sizeof(*dev));
-		dev->priv = (void *) devbusfn;
-		sprintf (dev->name, "pcnet#%d", dev_nr);
+		dev->priv = (void *)devbusfn;
+		sprintf(dev->name, "pcnet#%d", dev_nr);
 
 		/*
 		 * Setup the PCI device.
 		 */
-		pci_read_config_dword (devbusfn, PCI_BASE_ADDRESS_0,
-				       (unsigned int *) &dev->iobase);
-		dev->iobase=pci_io_to_phys (devbusfn, dev->iobase);
+		pci_read_config_dword(devbusfn, PCI_BASE_ADDRESS_0,
+				      (unsigned int *)&dev->iobase);
+		dev->iobase = pci_io_to_phys(devbusfn, dev->iobase);
 		dev->iobase &= ~0xf;
 
-		PCNET_DEBUG1 ("%s: devbusfn=0x%x iobase=0x%x: ",
-			      dev->name, devbusfn, dev->iobase);
+		PCNET_DEBUG1("%s: devbusfn=0x%x iobase=0x%x: ",
+			     dev->name, devbusfn, dev->iobase);
 
 		command = PCI_COMMAND_IO | PCI_COMMAND_MASTER;
-		pci_write_config_word (devbusfn, PCI_COMMAND, command);
-		pci_read_config_word (devbusfn, PCI_COMMAND, &status);
+		pci_write_config_word(devbusfn, PCI_COMMAND, command);
+		pci_read_config_word(devbusfn, PCI_COMMAND, &status);
 		if ((status & command) != command) {
-			printf ("%s: Couldn't enable IO access or Bus Mastering\n", dev->name);
-			free (dev);
+			printf("%s: Couldn't enable IO access or Bus Mastering\n",
+			       dev->name);
+			free(dev);
 			continue;
 		}
 
-		pci_write_config_byte (devbusfn, PCI_LATENCY_TIMER, 0x40);
+		pci_write_config_byte(devbusfn, PCI_LATENCY_TIMER, 0x40);
 
 		/*
 		 * Probe the PCnet chip.
 		 */
-		if (pcnet_probe (dev, bis, dev_nr) < 0) {
-			free (dev);
+		if (pcnet_probe(dev, bis, dev_nr) < 0) {
+			free(dev);
 			continue;
 		}
 
@@ -207,15 +208,15 @@ int pcnet_initialize (bd_t * bis)
 		dev->send = pcnet_send;
 		dev->recv = pcnet_recv;
 
-		eth_register (dev);
+		eth_register(dev);
 	}
 
-	udelay (10 * 1000);
+	udelay(10 * 1000);
 
 	return dev_nr;
 }
 
-static int pcnet_probe (struct eth_device *dev, bd_t * bis, int dev_nr)
+static int pcnet_probe(struct eth_device *dev, bd_t *bis, int dev_nr)
 {
 	int chip_version;
 	char *chipname;
@@ -225,17 +226,17 @@ static int pcnet_probe (struct eth_device *dev, bd_t * bis, int dev_nr)
 #endif
 
 	/* Reset the PCnet controller */
-	pcnet_reset (dev);
+	pcnet_reset(dev);
 
 	/* Check if register access is working */
-	if (pcnet_read_csr (dev, 0) != 4 || !pcnet_check (dev)) {
-		printf ("%s: CSR register access check failed\n", dev->name);
+	if (pcnet_read_csr(dev, 0) != 4 || !pcnet_check(dev)) {
+		printf("%s: CSR register access check failed\n", dev->name);
 		return -1;
 	}
 
 	/* Identify the chip */
 	chip_version =
-		pcnet_read_csr (dev, 88) | (pcnet_read_csr (dev, 89) << 16);
+		pcnet_read_csr(dev, 88) | (pcnet_read_csr(dev, 89) << 16);
 	if ((chip_version & 0xfff) != 0x003)
 		return -1;
 	chip_version = (chip_version >> 12) & 0xffff;
@@ -254,12 +255,12 @@ static int pcnet_probe (struct eth_device *dev, bd_t * bis, int dev_nr)
 		break;
 #endif
 	default:
-		printf ("%s: PCnet version %#x not supported\n",
-			dev->name, chip_version);
+		printf("%s: PCnet version %#x not supported\n",
+		       dev->name, chip_version);
 		return -1;
 	}
 
-	PCNET_DEBUG1 ("AMD %s\n", chipname);
+	PCNET_DEBUG1("AMD %s\n", chipname);
 
 #ifdef PCNET_HAS_PROM
 	/*
@@ -270,7 +271,7 @@ static int pcnet_probe (struct eth_device *dev, bd_t * bis, int dev_nr)
 	for (i = 0; i < 3; i++) {
 		unsigned int val;
 
-		val = pcnet_read_csr (dev, i + 12) & 0x0ffff;
+		val = pcnet_read_csr(dev, i + 12) & 0x0ffff;
 		/* There may be endianness issues here. */
 		dev->enetaddr[2 * i] = val & 0x0ff;
 		dev->enetaddr[2 * i + 1] = (val >> 8) & 0x0ff;
@@ -280,35 +281,35 @@ static int pcnet_probe (struct eth_device *dev, bd_t * bis, int dev_nr)
 	return 0;
 }
 
-static int pcnet_init (struct eth_device *dev, bd_t * bis)
+static int pcnet_init(struct eth_device *dev, bd_t *bis)
 {
 	int i, val;
 	u32 addr;
 
-	PCNET_DEBUG1 ("%s: pcnet_init...\n", dev->name);
+	PCNET_DEBUG1("%s: pcnet_init...\n", dev->name);
 
 	/* Switch pcnet to 32bit mode */
-	pcnet_write_bcr (dev, 20, 2);
+	pcnet_write_bcr(dev, 20, 2);
 
 #ifdef CONFIG_PN62
 	/* Setup LED registers */
-	val = pcnet_read_bcr (dev, 2) | 0x1000;
-	pcnet_write_bcr (dev, 2, val);	/* enable LEDPE */
-	pcnet_write_bcr (dev, 4, 0x5080);	/* 100MBit */
-	pcnet_write_bcr (dev, 5, 0x40c0);	/* LNKSE */
-	pcnet_write_bcr (dev, 6, 0x4090);	/* TX Activity */
-	pcnet_write_bcr (dev, 7, 0x4084);	/* RX Activity */
+	val = pcnet_read_bcr(dev, 2) | 0x1000;
+	pcnet_write_bcr(dev, 2, val);	/* enable LEDPE */
+	pcnet_write_bcr(dev, 4, 0x5080);	/* 100MBit */
+	pcnet_write_bcr(dev, 5, 0x40c0);	/* LNKSE */
+	pcnet_write_bcr(dev, 6, 0x4090);	/* TX Activity */
+	pcnet_write_bcr(dev, 7, 0x4084);	/* RX Activity */
 #endif
 
 	/* Set/reset autoselect bit */
-	val = pcnet_read_bcr (dev, 2) & ~2;
+	val = pcnet_read_bcr(dev, 2) & ~2;
 	val |= 2;
-	pcnet_write_bcr (dev, 2, val);
+	pcnet_write_bcr(dev, 2, val);
 
 	/* Enable auto negotiate, setup, disable fd */
-	val = pcnet_read_bcr (dev, 32) & ~0x98;
+	val = pcnet_read_bcr(dev, 32) & ~0x98;
 	val |= 0x20;
-	pcnet_write_bcr (dev, 32, val);
+	pcnet_write_bcr(dev, 32, val);
 
 	/*
 	 * We only maintain one structure because the drivers will never
@@ -316,12 +317,12 @@ static int pcnet_init (struct eth_device *dev, bd_t * bis)
 	 * must be aligned on 16-byte boundaries.
 	 */
 	if (lp == NULL) {
-		addr = (u32) malloc (sizeof (pcnet_priv_t) + 0x10);
+		addr = (u32)malloc(sizeof(pcnet_priv_t) + 0x10);
 		addr = (addr + 0xf) & ~0xf;
-		lp = (pcnet_priv_t *) addr;
+		lp = (pcnet_priv_t *)addr;
 	}
 
-	lp->init_block.mode = cpu_to_le16 (0x0000);
+	lp->init_block.mode = cpu_to_le16(0x0000);
 	lp->init_block.filter[0] = 0x00000000;
 	lp->init_block.filter[1] = 0x00000000;
 
@@ -330,9 +331,9 @@ static int pcnet_init (struct eth_device *dev, bd_t * bis)
 	 */
 	lp->cur_rx = 0;
 	for (i = 0; i < RX_RING_SIZE; i++) {
-		lp->rx_ring[i].base = PCI_TO_MEM_LE (dev, lp->rx_buf[i]);
-		lp->rx_ring[i].buf_length = cpu_to_le16 (-PKT_BUF_SZ);
-		lp->rx_ring[i].status = cpu_to_le16 (0x8000);
+		lp->rx_ring[i].base = PCI_TO_MEM_LE(dev, lp->rx_buf[i]);
+		lp->rx_ring[i].buf_length = cpu_to_le16(-PKT_BUF_SZ);
+		lp->rx_ring[i].status = cpu_to_le16(0x8000);
 		PCNET_DEBUG1
 			("Rx%d: base=0x%x buf_length=0x%hx status=0x%hx\n", i,
 			 lp->rx_ring[i].base, lp->rx_ring[i].buf_length,
@@ -352,48 +353,48 @@ static int pcnet_init (struct eth_device *dev, bd_t * bis)
 	/*
 	 * Setup Init Block.
 	 */
-	PCNET_DEBUG1 ("Init block at 0x%p: MAC", &lp->init_block);
+	PCNET_DEBUG1("Init block at 0x%p: MAC", &lp->init_block);
 
 	for (i = 0; i < 6; i++) {
 		lp->init_block.phys_addr[i] = dev->enetaddr[i];
-		PCNET_DEBUG1 (" %02x", lp->init_block.phys_addr[i]);
+		PCNET_DEBUG1(" %02x", lp->init_block.phys_addr[i]);
 	}
 
-	lp->init_block.tlen_rlen = cpu_to_le16 (TX_RING_LEN_BITS |
-						RX_RING_LEN_BITS);
-	lp->init_block.rx_ring = PCI_TO_MEM_LE (dev, lp->rx_ring);
-	lp->init_block.tx_ring = PCI_TO_MEM_LE (dev, lp->tx_ring);
+	lp->init_block.tlen_rlen = cpu_to_le16(TX_RING_LEN_BITS |
+					       RX_RING_LEN_BITS);
+	lp->init_block.rx_ring = PCI_TO_MEM_LE(dev, lp->rx_ring);
+	lp->init_block.tx_ring = PCI_TO_MEM_LE(dev, lp->tx_ring);
 
-	PCNET_DEBUG1 ("\ntlen_rlen=0x%x rx_ring=0x%x tx_ring=0x%x\n",
-		      lp->init_block.tlen_rlen,
-		      lp->init_block.rx_ring, lp->init_block.tx_ring);
+	PCNET_DEBUG1("\ntlen_rlen=0x%x rx_ring=0x%x tx_ring=0x%x\n",
+		     lp->init_block.tlen_rlen,
+		     lp->init_block.rx_ring, lp->init_block.tx_ring);
 
 	/*
 	 * Tell the controller where the Init Block is located.
 	 */
-	addr = PCI_TO_MEM (dev, &lp->init_block);
-	pcnet_write_csr (dev, 1, addr & 0xffff);
-	pcnet_write_csr (dev, 2, (addr >> 16) & 0xffff);
+	addr = PCI_TO_MEM(dev, &lp->init_block);
+	pcnet_write_csr(dev, 1, addr & 0xffff);
+	pcnet_write_csr(dev, 2, (addr >> 16) & 0xffff);
 
-	pcnet_write_csr (dev, 4, 0x0915);
-	pcnet_write_csr (dev, 0, 0x0001);	/* start */
+	pcnet_write_csr(dev, 4, 0x0915);
+	pcnet_write_csr(dev, 0, 0x0001);	/* start */
 
 	/* Wait for Init Done bit */
 	for (i = 10000; i > 0; i--) {
-		if (pcnet_read_csr (dev, 0) & 0x0100)
+		if (pcnet_read_csr(dev, 0) & 0x0100)
 			break;
-		udelay (10);
+		udelay(10);
 	}
 	if (i <= 0) {
-		printf ("%s: TIMEOUT: controller init failed\n", dev->name);
-		pcnet_reset (dev);
+		printf("%s: TIMEOUT: controller init failed\n", dev->name);
+		pcnet_reset(dev);
 		return -1;
 	}
 
 	/*
 	 * Finally start network controller operation.
 	 */
-	pcnet_write_csr (dev, 0, 0x0002);
+	pcnet_write_csr(dev, 0, 0x0002);
 
 	return 0;
 }
@@ -403,20 +404,20 @@ static int pcnet_send(struct eth_device *dev, void *packet, int pkt_len)
 	int i, status;
 	struct pcnet_tx_head *entry = &lp->tx_ring[lp->cur_tx];
 
-	PCNET_DEBUG2 ("Tx%d: %d bytes from 0x%p ", lp->cur_tx, pkt_len,
-		      packet);
+	PCNET_DEBUG2("Tx%d: %d bytes from 0x%p ", lp->cur_tx, pkt_len,
+		     packet);
 
 	/* Wait for completion by testing the OWN bit */
 	for (i = 1000; i > 0; i--) {
-		status = le16_to_cpu (entry->status);
+		status = le16_to_cpu(entry->status);
 		if ((status & 0x8000) == 0)
 			break;
-		udelay (100);
-		PCNET_DEBUG2 (".");
+		udelay(100);
+		PCNET_DEBUG2(".");
 	}
 	if (i <= 0) {
-		printf ("%s: TIMEOUT: Tx%d failed (status = 0x%x)\n",
-			dev->name, lp->cur_tx, status);
+		printf("%s: TIMEOUT: Tx%d failed (status = 0x%x)\n",
+		       dev->name, lp->cur_tx, status);
 		pkt_len = 0;
 		goto failure;
 	}
@@ -426,19 +427,19 @@ static int pcnet_send(struct eth_device *dev, void *packet, int pkt_len)
 	 * set the status with the "ownership" bits last.
 	 */
 	status = 0x8300;
-	entry->length = le16_to_cpu (-pkt_len);
+	entry->length = le16_to_cpu(-pkt_len);
 	entry->misc = 0x00000000;
-	entry->base = PCI_TO_MEM_LE (dev, packet);
-	entry->status = le16_to_cpu (status);
+	entry->base = PCI_TO_MEM_LE(dev, packet);
+	entry->status = le16_to_cpu(status);
 
 	/* Trigger an immediate send poll. */
-	pcnet_write_csr (dev, 0, 0x0008);
+	pcnet_write_csr(dev, 0, 0x0008);
 
       failure:
 	if (++lp->cur_tx >= TX_RING_SIZE)
 		lp->cur_tx = 0;
 
-	PCNET_DEBUG2 ("done\n");
+	PCNET_DEBUG2("done\n");
 	return pkt_len;
 }
 
@@ -453,40 +454,38 @@ static int pcnet_recv (struct eth_device *dev)
 		/*
 		 * If we own the next entry, it's a new packet. Send it up.
 		 */
-		if (((status = le16_to_cpu (entry->status)) & 0x8000) != 0) {
+		status = le16_to_cpu(entry->status);
+		if ((status & 0x8000) != 0)
 			break;
-		}
 		status >>= 8;
 
 		if (status != 0x03) {	/* There was an error. */
-
-			printf ("%s: Rx%d", dev->name, lp->cur_rx);
-			PCNET_DEBUG1 (" (status=0x%x)", status);
+			printf("%s: Rx%d", dev->name, lp->cur_rx);
+			PCNET_DEBUG1(" (status=0x%x)", status);
 			if (status & 0x20)
-				printf (" Frame");
+				printf(" Frame");
 			if (status & 0x10)
-				printf (" Overflow");
+				printf(" Overflow");
 			if (status & 0x08)
-				printf (" CRC");
+				printf(" CRC");
 			if (status & 0x04)
-				printf (" Fifo");
-			printf (" Error\n");
-			entry->status &= le16_to_cpu (0x03ff);
+				printf(" Fifo");
+			printf(" Error\n");
+			entry->status &= le16_to_cpu(0x03ff);
 
 		} else {
-
-			pkt_len =
-				(le32_to_cpu (entry->msg_length) & 0xfff) - 4;
+			pkt_len = (le32_to_cpu(entry->msg_length) & 0xfff) - 4;
 			if (pkt_len < 60) {
-				printf ("%s: Rx%d: invalid packet length %d\n", dev->name, lp->cur_rx, pkt_len);
+				printf("%s: Rx%d: invalid packet length %d\n",
+				       dev->name, lp->cur_rx, pkt_len);
 			} else {
-				NetReceive (lp->rx_buf[lp->cur_rx], pkt_len);
-				PCNET_DEBUG2 ("Rx%d: %d bytes from 0x%p\n",
-					      lp->cur_rx, pkt_len,
-					      lp->rx_buf[lp->cur_rx]);
+				NetReceive(lp->rx_buf[lp->cur_rx], pkt_len);
+				PCNET_DEBUG2("Rx%d: %d bytes from 0x%p\n",
+					     lp->cur_rx, pkt_len,
+					     lp->rx_buf[lp->cur_rx]);
 			}
 		}
-		entry->status |= cpu_to_le16 (0x8000);
+		entry->status |= cpu_to_le16(0x8000);
 
 		if (++lp->cur_rx >= RX_RING_SIZE)
 			lp->cur_rx = 0;
@@ -494,22 +493,21 @@ static int pcnet_recv (struct eth_device *dev)
 	return pkt_len;
 }
 
-static void pcnet_halt (struct eth_device *dev)
+static void pcnet_halt(struct eth_device *dev)
 {
 	int i;
 
-	PCNET_DEBUG1 ("%s: pcnet_halt...\n", dev->name);
+	PCNET_DEBUG1("%s: pcnet_halt...\n", dev->name);
 
 	/* Reset the PCnet controller */
-	pcnet_reset (dev);
+	pcnet_reset(dev);
 
 	/* Wait for Stop bit */
 	for (i = 1000; i > 0; i--) {
-		if (pcnet_read_csr (dev, 0) & 0x4)
+		if (pcnet_read_csr(dev, 0) & 0x4)
 			break;
-		udelay (10);
-	}
-	if (i <= 0) {
-		printf ("%s: TIMEOUT: controller reset failed\n", dev->name);
+		udelay(10);
 	}
+	if (i <= 0)
+		printf("%s: TIMEOUT: controller reset failed\n", dev->name);
 }
Signature · 1 line
1.8.4.1
#3 lore raw

PATCH v2 3/18 pcnet: s/le16_to_cpu/cpu_to_le16/ in pcnet_send

In reply to Paul Burton

This should cause no change to the generated code, but is semantically correct.

---
Changes in v2:
- rebase atop master

---
 drivers/net/pcnet.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/pcnet.c +2 -2
index a30a0bc..843a6fc 100644
@@ -427,10 +427,10 @@ static int pcnet_send(struct eth_device *dev, void *packet, int pkt_len)
 	 * set the status with the "ownership" bits last.
 	 */
 	status = 0x8300;
-	entry->length = le16_to_cpu(-pkt_len);
+	entry->length = cpu_to_le16(-pkt_len);
 	entry->misc = 0x00000000;
 	entry->base = PCI_TO_MEM_LE(dev, packet);
-	entry->status = le16_to_cpu(status);
+	entry->status = cpu_to_le16(status);
 
 	/* Trigger an immediate send poll. */
 	pcnet_write_csr(dev, 0, 0x0008);
Signature · 1 line
1.8.4.1
#4 lore raw

PATCH v2 4/18 pcnet: add cache flushing & invalidation

In reply to Paul Burton

Ensure that the view of memory from the CPU & the ethernet controller is coherent at the various points where they exchange data. This prevents stale data from being transmitted or received, and prevents the driver from getting stuck waiting for the ethernet controller to update descriptors when in reality it has but the old values are being read from cache.

---
Changes in v2:
- rebase atop master

---
 drivers/net/pcnet.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
drivers/net/pcnet.c +16 -0
index 843a6fc..7b87660 100644
@@ -364,6 +364,7 @@ static int pcnet_init(struct eth_device *dev, bd_t *bis)
 					       RX_RING_LEN_BITS);
 	lp->init_block.rx_ring = PCI_TO_MEM_LE(dev, lp->rx_ring);
 	lp->init_block.tx_ring = PCI_TO_MEM_LE(dev, lp->tx_ring);
+	flush_dcache_range((unsigned long)lp, (unsigned long)&lp->rx_buf);
 
 	PCNET_DEBUG1("\ntlen_rlen=0x%x rx_ring=0x%x tx_ring=0x%x\n",
 		     lp->init_block.tlen_rlen,
@@ -407,8 +408,13 @@ static int pcnet_send(struct eth_device *dev, void *packet, int pkt_len)
 	PCNET_DEBUG2("Tx%d: %d bytes from 0x%p ", lp->cur_tx, pkt_len,
 		     packet);
 
+	flush_dcache_range((unsigned long)packet,
+			   (unsigned long)packet + pkt_len);
+
 	/* Wait for completion by testing the OWN bit */
 	for (i = 1000; i > 0; i--) {
+		invalidate_dcache_range((unsigned long)entry,
+					(unsigned long)entry + sizeof(*entry));
 		status = le16_to_cpu(entry->status);
 		if ((status & 0x8000) == 0)
 			break;
@@ -431,6 +437,8 @@ static int pcnet_send(struct eth_device *dev, void *packet, int pkt_len)
 	entry->misc = 0x00000000;
 	entry->base = PCI_TO_MEM_LE(dev, packet);
 	entry->status = cpu_to_le16(status);
+	flush_dcache_range((unsigned long)entry,
+			   (unsigned long)entry + sizeof(*entry));
 
 	/* Trigger an immediate send poll. */
 	pcnet_write_csr(dev, 0, 0x0008);
@@ -451,6 +459,8 @@ static int pcnet_recv (struct eth_device *dev)
 
 	while (1) {
 		entry = &lp->rx_ring[lp->cur_rx];
+		invalidate_dcache_range((unsigned long)entry,
+					(unsigned long)entry + sizeof(*entry));
 		/*
 		 * If we own the next entry, it's a new packet. Send it up.
 		 */
@@ -479,6 +489,10 @@ static int pcnet_recv (struct eth_device *dev)
 				printf("%s: Rx%d: invalid packet length %d\n",
 				       dev->name, lp->cur_rx, pkt_len);
 			} else {
+				invalidate_dcache_range(
+					(unsigned long)lp->rx_buf[lp->cur_rx],
+					(unsigned long)lp->rx_buf[lp->cur_rx] +
+					pkt_len);
 				NetReceive(lp->rx_buf[lp->cur_rx], pkt_len);
 				PCNET_DEBUG2("Rx%d: %d bytes from 0x%p\n",
 					     lp->cur_rx, pkt_len,
@@ -486,6 +500,8 @@ static int pcnet_recv (struct eth_device *dev)
 			}
 		}
 		entry->status |= cpu_to_le16(0x8000);
+		flush_dcache_range((unsigned long)entry,
+				   (unsigned long)entry + sizeof(*entry));
 
 		if (++lp->cur_rx >= RX_RING_SIZE)
 			lp->cur_rx = 0;
Signature · 1 line
1.8.4.1
#5 lore raw

PATCH v2 5/18 pcnet: enable the NOUFLO feature

In reply to Paul Burton

On relatively slow boards (such as the MIPS Malta with an FPGA core card) it can be extremely common for transmits to underflow - to the point where it appears they simply do not work at all. Setting the NOUFLO bit causes the ethernet controller to not begin transmission on the wire until a transmit start point is reached. Setting that transmit start point to the full packet will cause the controller to only transmit the packet once it has buffered it entirely thus preventing any transmit underflows from occuring and allowing the controller to function on slower boards.

---
Changes in v2:
- rebase atop master

---
 drivers/net/pcnet.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)
drivers/net/pcnet.c +15 -0
index 7b87660..5b248be 100644
@@ -312,6 +312,21 @@ static int pcnet_init(struct eth_device *dev, bd_t *bis)
 	pcnet_write_bcr(dev, 32, val);
 
 	/*
+	 * Enable NOUFLO on supported controllers, with the transmit
+	 * start point set to the full packet. This will cause entire
+	 * packets to be buffered by the ethernet controller before
+	 * transmission, eliminating underflows which are common on
+	 * slower devices. Controllers which do not support NOUFLO will
+	 * simply be left with a larger transmit FIFO threshold.
+	 */
+	val = pcnet_read_bcr(dev, 18);
+	val |= 1 << 11;
+	pcnet_write_bcr(dev, 18, val);
+	val = pcnet_read_csr(dev, 80);
+	val |= 0x3 << 10;
+	pcnet_write_csr(dev, 80, val);
+
+	/*
 	 * We only maintain one structure because the drivers will never
 	 * be used concurrently. In 32bit mode the RX and TX ring entries
 	 * must be aligned on 16-byte boundaries.
Signature · 1 line
1.8.4.1
#6 lore raw

PATCH v2 5/18 pcnet: enable the NOUFLO feature

Dear Paul Burton,

Quoted 29 lines

On relatively slow boards (such as the MIPS Malta with an FPGA core card) it can be extremely common for transmits to underflow - to the point where it appears they simply do not work at all. Setting the NOUFLO bit causes the ethernet controller to not begin transmission on the wire until a transmit start point is reached. Setting that transmit start point to the full packet will cause the controller to only transmit the packet once it has buffered it entirely thus preventing any transmit underflows from occuring and allowing the controller to function on slower boards.

---
Changes in v2:
- rebase atop master

---
 drivers/net/pcnet.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)
drivers/net/pcnet.c +3 -0
index 7b87660..5b248be 100644
@@ -312,6 +312,21 @@ static int pcnet_init(struct eth_device *dev, bd_t
 *bis) pcnet_write_bcr(dev, 32, val);
 
 	/*
+	 * Enable NOUFLO on supported controllers, with the transmit
+	 * start point set to the full packet. This will cause entire
+	 * packets to be buffered by the ethernet controller before

drivers/net/pcnet.c static int pcnet_init(struct eth_device *dev, bd_t · line 327

Quoted 10 lines
drivers/net/pcnet.c +10 -0
+	 * transmission, eliminating underflows which are common on
+	 * slower devices. Controllers which do not support NOUFLO will
+	 * simply be left with a larger transmit FIFO threshold.
+	 */
+	val = pcnet_read_bcr(dev, 18);
+	val |= 1 << 11;
+	pcnet_write_bcr(dev, 18, val);
+	val = pcnet_read_csr(dev, 80);
+	val |= 0x3 << 10;
+	pcnet_write_csr(dev, 80, val);

Would be nice to fix this so we don't have so many magic numbers there.

Best regards,
Marek Vasut

#7 lore raw

PATCH v2 5/18 pcnet: enable the NOUFLO feature

Quoted 44 lines · Marek Vasut

Dear Paul Burton,

Quoted 29 lines

On relatively slow boards (such as the MIPS Malta with an FPGA core card) it can be extremely common for transmits to underflow - to the point where it appears they simply do not work at all. Setting the NOUFLO bit causes the ethernet controller to not begin transmission on the wire until a transmit start point is reached. Setting that transmit start point to the full packet will cause the controller to only transmit the packet once it has buffered it entirely thus preventing any transmit underflows from occuring and allowing the controller to function on slower boards.

---
Changes in v2:
- rebase atop master

---
  drivers/net/pcnet.c | 15 +++++++++++++++
  1 file changed, 15 insertions(+)
drivers/net/pcnet.c +3 -0
index 7b87660..5b248be 100644
@@ -312,6 +312,21 @@ static int pcnet_init(struct eth_device *dev, bd_t
 *bis) pcnet_write_bcr(dev, 32, val);
 
  	/*
+	 * Enable NOUFLO on supported controllers, with the transmit
+	 * start point set to the full packet. This will cause entire
+	 * packets to be buffered by the ethernet controller before

drivers/net/pcnet.c static int pcnet_init(struct eth_device *dev, bd_t · line 327

Quoted 10 lines
drivers/net/pcnet.c +10 -0
+	 * transmission, eliminating underflows which are common on
+	 * slower devices. Controllers which do not support NOUFLO will
+	 * simply be left with a larger transmit FIFO threshold.
+	 */
+	val = pcnet_read_bcr(dev, 18);
+	val |= 1 << 11;
+	pcnet_write_bcr(dev, 18, val);
+	val = pcnet_read_csr(dev, 80);
+	val |= 0x3 << 10;
+	pcnet_write_csr(dev, 80, val);

Would be nice to fix this so we don't have so many magic numbers there.

Quoted 3 lines

Best regards,
Marek Vasut

I agree, the whole file is full of them. I'll submit a cleanup patch later if nobody else beats me to it.

Thanks,
Paul

#8 lore raw

PATCH v2 5/18 pcnet: enable the NOUFLO feature

Dear Paul Burton,

Quoted 49 lines · Marek Vasut
Quoted 48 lines

Dear Paul Burton,

Quoted 33 lines

On relatively slow boards (such as the MIPS Malta with an FPGA core card) it can be extremely common for transmits to underflow - to the point where it appears they simply do not work at all. Setting the NOUFLO bit causes the ethernet controller to not begin transmission on the wire until a transmit start point is reached. Setting that transmit start point to the full packet will cause the controller to only transmit the packet once it has buffered it entirely thus preventing any transmit underflows from occuring and allowing the controller to function on slower boards.

---

Changes in v2:
- rebase atop master

---

  drivers/net/pcnet.c | 15 +++++++++++++++
  1 file changed, 15 insertions(+)
drivers/net/pcnet.c +3 -0
index 7b87660..5b248be 100644
@@ -312,6 +312,21 @@ static int pcnet_init(struct eth_device *dev, bd_t
 *bis) pcnet_write_bcr(dev, 32, val);
 
  	/*
 
+	 * Enable NOUFLO on supported controllers, with the transmit
+	 * start point set to the full packet. This will cause entire
+	 * packets to be buffered by the ethernet controller before

drivers/net/pcnet.c static int pcnet_init(struct eth_device *dev, bd_t · line 328

Quoted 10 lines
drivers/net/pcnet.c +10 -0
+	 * transmission, eliminating underflows which are common on
+	 * slower devices. Controllers which do not support NOUFLO will
+	 * simply be left with a larger transmit FIFO threshold.
+	 */
+	val = pcnet_read_bcr(dev, 18);
+	val |= 1 << 11;
+	pcnet_write_bcr(dev, 18, val);
+	val = pcnet_read_csr(dev, 80);
+	val |= 0x3 << 10;
+	pcnet_write_csr(dev, 80, val);

Would be nice to fix this so we don't have so many magic numbers there.

Quoted 5 lines
Quoted 2 lines

Best regards,
Marek Vasut

I agree, the whole file is full of them. I'll submit a cleanup patch later if nobody else beats me to it.

Thank you, that'd be really nice of you to do.

Best regards,
Marek Vasut

#9 lore raw

PATCH v2 6/18 pci.h: allow inclusion in assembly source

In reply to Paul Burton

This patch simply #ifdef's out the C-specific parts of pci.h when it is included by an assembly file. This will allow the macros it contains to be used from assembly source as will be done in a followup commit adding support for more modern MIPS Malta boards.

---
Changes in v2:
- rebase atop master

---
 include/pci.h | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
include/pci.h +5 -1
index d462479..461f17c 100644
@@ -417,6 +417,8 @@
 
 #include <pci_ids.h>
 
+#ifndef __ASSEMBLY__
+
 #ifdef CONFIG_SYS_PCI_64BIT
 typedef u64 pci_addr_t;
 typedef u64 pci_size_t;
@@ -667,4 +669,6 @@ extern void pci_mpc824x_init (struct pci_controller *hose);
 #ifdef CONFIG_MPC85xx
 extern void pci_mpc85xx_init (struct pci_controller *hose);
 #endif
-#endif	/* _PCI_H */
+
+#endif /* __ASSEMBLY__ */
+#endif /* _PCI_H */
Signature · 1 line
1.8.4.1
#10 lore raw

PATCH v4 7/18 qemu-malta: rename to just "malta"

In reply to Paul Burton

This is in preparation for adapting this board to function correctly on a physical MIPS Malta board. The board is moved into an "imgtec" vendor directory at the same time in order to ready us for any other boards supported by Imagination in the future.

---
Changes in v4:
- rebase atop master

Changes in v3:
- modify a missed CONFIG_QEMU_MALTA in arch/mips/cpu/mips32/start.S which would affect old QEMU builds

Changes in v2:
- move to an imgtec vendor directory to reduce churn should other Imagination boards be added later
- enable rename detection for the patch

---
 arch/mips/cpu/mips32/start.S                            |  2 +-
 arch/mips/lib/bootm.c                                   | 12 ++++++------
 board/{qemu-malta => imgtec/malta}/Makefile             |  2 +-
 board/{qemu-malta => imgtec/malta}/lowlevel_init.S      |  0
 board/{qemu-malta/qemu-malta.c => imgtec/malta/malta.c} |  0
 boards.cfg                                              |  4 ++--
 include/configs/{qemu-malta.h => malta.h}               |  2 +-
 7 files changed, 11 insertions(+), 11 deletions(-)
 rename board/{qemu-malta => imgtec/malta}/Makefile (87%)
 rename board/{qemu-malta => imgtec/malta}/lowlevel_init.S (100%)
 rename board/{qemu-malta/qemu-malta.c => imgtec/malta/malta.c} (100%)
 rename include/configs/{qemu-malta.h => malta.h} (98%)
arch/mips/cpu/mips32/start.S +1 -1
index 70ad198..68e59b5 100644
@@ -51,7 +51,7 @@ _start:
 	 */
 	.word CONFIG_SYS_XWAY_EBU_BOOTCFG
 	.word 0x0
-#elif defined(CONFIG_QEMU_MALTA)
+#elif defined(CONFIG_MALTA)
 	/*
 	 * Linux expects the Board ID here.
 	 */
arch/mips/lib/bootm.c +6 -6
index 66340ea..1febf29 100644
@@ -17,10 +17,10 @@ DECLARE_GLOBAL_DATA_PTR;
 #define	LINUX_MAX_ENVS		256
 #define	LINUX_MAX_ARGS		256
 
-#if defined(CONFIG_QEMU_MALTA)
-#define mips_boot_qemu_malta	1
+#if defined(CONFIG_MALTA)
+#define mips_boot_malta		1
 #else
-#define mips_boot_qemu_malta	0
+#define mips_boot_malta		0
 #endif
 
 static int linux_argc;
@@ -139,7 +139,7 @@ static void linux_env_set(const char *env_name, const char *env_val)
 		strcpy(linux_env_p, env_name);
 		linux_env_p += strlen(env_name);
 
-		if (mips_boot_qemu_malta) {
+		if (mips_boot_malta) {
 			linux_env_p++;
 			linux_env[++linux_env_idx] = linux_env_p;
 		} else {
@@ -196,7 +196,7 @@ static void boot_prep_linux(bootm_headers_t *images)
 	if (cp)
 		linux_env_set("eth1addr", cp);
 
-	if (mips_boot_qemu_malta)
+	if (mips_boot_malta)
 		linux_env_set("modetty0", "38400n8r");
 }
 
@@ -210,7 +210,7 @@ static void boot_jump_linux(bootm_headers_t *images)
 
 	bootstage_mark(BOOTSTAGE_ID_RUN_OS);
 
-	if (mips_boot_qemu_malta)
+	if (mips_boot_malta)
 		linux_extra = gd->ram_size;
 
 	/* we assume that the kernel is in place */
board/imgtec/malta/Makefile renamed from board/qemu-malta/Makefile +1 -1
similarity index 87%
rename from board/qemu-malta/Makefile
rename to board/imgtec/malta/Makefile
index 5d727f6..091830d 100644
@@ -5,5 +5,5 @@
 # SPDX-License-Identifier:	GPL-2.0+
 #
 
-obj-y	= qemu-malta.o
+obj-y	= malta.o
 obj-y	+= lowlevel_init.o
board/imgtec/malta/lowlevel_init.S renamed from board/qemu-malta/lowlevel_init.S +0 -0
similarity index 100%
rename from board/qemu-malta/lowlevel_init.S
rename to board/imgtec/malta/lowlevel_init.S
board/imgtec/malta/malta.c renamed from board/qemu-malta/qemu-malta.c +0 -0
similarity index 100%
rename from board/qemu-malta/qemu-malta.c
rename to board/imgtec/malta/malta.c
boards.cfg +2 -2
index 5e10125..54c6bf7 100644
@@ -481,10 +481,10 @@ Active  m68k        mcf547x_8x     -           freescale       m548xevb
 Active  m68k        mcf547x_8x     -           freescale       m548xevb            M5485GFE                             M5485EVB:SYS_BUSCLK=100000000,SYS_BOOTSZ=4,SYS_DRAMSZ=64                                                                          TsiChung Liew <[email protected]>
 Active  m68k        mcf547x_8x     -           freescale       m548xevb            M5485HFE                             M5485EVB:SYS_BUSCLK=100000000,SYS_BOOTSZ=2,SYS_DRAMSZ=64,SYS_NOR1SZ=16,SYS_VIDEO                                                  TsiChung Liew <[email protected]>
 Active  microblaze  microblaze     -           xilinx          microblaze-generic  microblaze-generic                   -                                                                                                                                 Michal Simek <[email protected]>
-Active  mips        mips32         -           -               qemu-malta          qemu_malta                           qemu-malta:MIPS32,SYS_BIG_ENDIAN                                                                                                  -
-Active  mips        mips32         -           -               qemu-malta          qemu_maltael                         qemu-malta:MIPS32,SYS_LITTLE_ENDIAN                                                                                               -
 Active  mips        mips32         -           -               qemu-mips           qemu_mips                            qemu-mips:SYS_BIG_ENDIAN                                                                                                          Vlad Lungu <[email protected]>
 Active  mips        mips32         -           -               qemu-mips           qemu_mipsel                          qemu-mips:SYS_LITTLE_ENDIAN                                                                                                       -
+Active  mips        mips32         -           imgtec          malta               malta                                malta:MIPS32,SYS_BIG_ENDIAN                                                                                                       -
+Active  mips        mips32         -           imgtec          malta               maltael                              malta:MIPS32,SYS_LITTLE_ENDIAN                                                                                                    -
 Active  mips        mips32         -           micronas        vct                 vct_platinum                         vct:VCT_PLATINUM                                                                                                                  -
 Active  mips        mips32         -           micronas        vct                 vct_platinum_onenand                 vct:VCT_PLATINUM,VCT_ONENAND                                                                                                      -
 Active  mips        mips32         -           micronas        vct                 vct_platinum_onenand_small           vct:VCT_PLATINUM,VCT_ONENAND,VCT_SMALL_IMAGE                                                                                      -
include/configs/malta.h renamed from include/configs/qemu-malta.h +1 -1
similarity index 98%
rename from include/configs/qemu-malta.h
rename to include/configs/malta.h
index 03514d1..4ca8bad 100644
@@ -13,7 +13,7 @@
 /*
  * System configuration
  */
-#define CONFIG_QEMU_MALTA
+#define CONFIG_MALTA
 
 #define CONFIG_PCI
 #define CONFIG_PCI_GT64120
Signature · 1 line
1.8.4.1
#11 lore raw

PATCH v4 7/18 qemu-malta: rename to just "malta"

2013/11/8 Paul Burton <[email protected]>:

Quoted 19 lines

This is in preparation for adapting this board to function correctly on a physical MIPS Malta board. The board is moved into an "imgtec" vendor directory at the same time in order to ready us for any other boards supported by Imagination in the future.

---
Changes in v4:
- rebase atop master

Changes in v3:
- modify a missed CONFIG_QEMU_MALTA in arch/mips/cpu/mips32/start.S which would affect old QEMU builds

Changes in v2:
- move to an imgtec vendor directory to reduce churn should other Imagination boards be added later
- enable rename detection for the patch
---

...

Quoted 17 lines
include/configs/malta.h renamed from include/configs/qemu-malta.h +1 -2
similarity index 98%
rename from include/configs/qemu-malta.h
rename to include/configs/malta.h
index 03514d1..4ca8bad 100644
@@ -13,7 +13,7 @@
 /*
  * System configuration
  */
-#define CONFIG_QEMU_MALTA
+#define CONFIG_MALTA
 
 #define CONFIG_PCI
 #define CONFIG_PCI_GT64120
--
Quoted 3 lines

1.8.4.1

I've missed that last time, you should also change the prompt:

#if defined(CONFIG_SYS_LITTLE_ENDIAN)
#define CONFIG_SYS_PROMPT "qemu-maltael # "
#else
#define CONFIG_SYS_PROMPT "qemu-malta # "
#endif

Signature · 1 line
- Daniel
#12 lore raw

PATCH v5 7/18 qemu-malta: rename to just "malta"

This is in preparation for adapting this board to function correctly on a physical MIPS Malta board. The board is moved into an "imgtec" vendor directory at the same time in order to ready us for any other boards supported by Imagination in the future.

---
Changes in v5:
- change CONFIG_SYS_PROMPT
- change include/configs/malta.h include guard

Changes in v4:
- rebase atop master

Changes in v3:
- modify a missed CONFIG_QEMU_MALTA in arch/mips/cpu/mips32/start.S which would affect old QEMU builds

Changes in v2:
- move to an imgtec vendor directory to reduce churn should other Imagination boards be added later
- enable rename detection for the patch

---
 arch/mips/cpu/mips32/start.S                            |  2 +-
 arch/mips/lib/bootm.c                                   | 12 ++++++------
 board/{qemu-malta => imgtec/malta}/Makefile             |  2 +-
 board/{qemu-malta => imgtec/malta}/lowlevel_init.S      |  0
 board/{qemu-malta/qemu-malta.c => imgtec/malta/malta.c} |  0
 boards.cfg                                              |  4 ++--
 include/configs/{qemu-malta.h => malta.h}               | 12 ++++++------
 7 files changed, 16 insertions(+), 16 deletions(-)
 rename board/{qemu-malta => imgtec/malta}/Makefile (87%)
 rename board/{qemu-malta => imgtec/malta}/lowlevel_init.S (100%)
 rename board/{qemu-malta/qemu-malta.c => imgtec/malta/malta.c} (100%)
 rename include/configs/{qemu-malta.h => malta.h} (91%)
arch/mips/cpu/mips32/start.S +1 -1
index 70ad198..68e59b5 100644
@@ -51,7 +51,7 @@ _start:
 	 */
 	.word CONFIG_SYS_XWAY_EBU_BOOTCFG
 	.word 0x0
-#elif defined(CONFIG_QEMU_MALTA)
+#elif defined(CONFIG_MALTA)
 	/*
 	 * Linux expects the Board ID here.
 	 */
arch/mips/lib/bootm.c +6 -6
index 66340ea..1febf29 100644
@@ -17,10 +17,10 @@ DECLARE_GLOBAL_DATA_PTR;
 #define	LINUX_MAX_ENVS		256
 #define	LINUX_MAX_ARGS		256
 
-#if defined(CONFIG_QEMU_MALTA)
-#define mips_boot_qemu_malta	1
+#if defined(CONFIG_MALTA)
+#define mips_boot_malta		1
 #else
-#define mips_boot_qemu_malta	0
+#define mips_boot_malta		0
 #endif
 
 static int linux_argc;
@@ -139,7 +139,7 @@ static void linux_env_set(const char *env_name, const char *env_val)
 		strcpy(linux_env_p, env_name);
 		linux_env_p += strlen(env_name);
 
-		if (mips_boot_qemu_malta) {
+		if (mips_boot_malta) {
 			linux_env_p++;
 			linux_env[++linux_env_idx] = linux_env_p;
 		} else {
@@ -196,7 +196,7 @@ static void boot_prep_linux(bootm_headers_t *images)
 	if (cp)
 		linux_env_set("eth1addr", cp);
 
-	if (mips_boot_qemu_malta)
+	if (mips_boot_malta)
 		linux_env_set("modetty0", "38400n8r");
 }
 
@@ -210,7 +210,7 @@ static void boot_jump_linux(bootm_headers_t *images)
 
 	bootstage_mark(BOOTSTAGE_ID_RUN_OS);
 
-	if (mips_boot_qemu_malta)
+	if (mips_boot_malta)
 		linux_extra = gd->ram_size;
 
 	/* we assume that the kernel is in place */
board/imgtec/malta/Makefile renamed from board/qemu-malta/Makefile +1 -1
similarity index 87%
rename from board/qemu-malta/Makefile
rename to board/imgtec/malta/Makefile
index 5d727f6..091830d 100644
@@ -5,5 +5,5 @@
 # SPDX-License-Identifier:	GPL-2.0+
 #
 
-obj-y	= qemu-malta.o
+obj-y	= malta.o
 obj-y	+= lowlevel_init.o
board/imgtec/malta/lowlevel_init.S renamed from board/qemu-malta/lowlevel_init.S +0 -0
similarity index 100%
rename from board/qemu-malta/lowlevel_init.S
rename to board/imgtec/malta/lowlevel_init.S
board/imgtec/malta/malta.c renamed from board/qemu-malta/qemu-malta.c +0 -0
similarity index 100%
rename from board/qemu-malta/qemu-malta.c
rename to board/imgtec/malta/malta.c
boards.cfg +2 -2
index 5e10125..54c6bf7 100644
@@ -481,10 +481,10 @@ Active  m68k        mcf547x_8x     -           freescale       m548xevb
 Active  m68k        mcf547x_8x     -           freescale       m548xevb            M5485GFE                             M5485EVB:SYS_BUSCLK=100000000,SYS_BOOTSZ=4,SYS_DRAMSZ=64                                                                          TsiChung Liew <[email protected]>
 Active  m68k        mcf547x_8x     -           freescale       m548xevb            M5485HFE                             M5485EVB:SYS_BUSCLK=100000000,SYS_BOOTSZ=2,SYS_DRAMSZ=64,SYS_NOR1SZ=16,SYS_VIDEO                                                  TsiChung Liew <[email protected]>
 Active  microblaze  microblaze     -           xilinx          microblaze-generic  microblaze-generic                   -                                                                                                                                 Michal Simek <[email protected]>
-Active  mips        mips32         -           -               qemu-malta          qemu_malta                           qemu-malta:MIPS32,SYS_BIG_ENDIAN                                                                                                  -
-Active  mips        mips32         -           -               qemu-malta          qemu_maltael                         qemu-malta:MIPS32,SYS_LITTLE_ENDIAN                                                                                               -
 Active  mips        mips32         -           -               qemu-mips           qemu_mips                            qemu-mips:SYS_BIG_ENDIAN                                                                                                          Vlad Lungu <[email protected]>
 Active  mips        mips32         -           -               qemu-mips           qemu_mipsel                          qemu-mips:SYS_LITTLE_ENDIAN                                                                                                       -
+Active  mips        mips32         -           imgtec          malta               malta                                malta:MIPS32,SYS_BIG_ENDIAN                                                                                                       -
+Active  mips        mips32         -           imgtec          malta               maltael                              malta:MIPS32,SYS_LITTLE_ENDIAN                                                                                                    -
 Active  mips        mips32         -           micronas        vct                 vct_platinum                         vct:VCT_PLATINUM                                                                                                                  -
 Active  mips        mips32         -           micronas        vct                 vct_platinum_onenand                 vct:VCT_PLATINUM,VCT_ONENAND                                                                                                      -
 Active  mips        mips32         -           micronas        vct                 vct_platinum_onenand_small           vct:VCT_PLATINUM,VCT_ONENAND,VCT_SMALL_IMAGE                                                                                      -
include/configs/malta.h renamed from include/configs/qemu-malta.h +6 -6
similarity index 91%
rename from include/configs/qemu-malta.h
rename to include/configs/malta.h
index 03514d1..d067d98 100644
@@ -4,8 +4,8 @@
  * SPDX-License-Identifier:	GPL-2.0
  */
 
-#ifndef _QEMU_MALTA_CONFIG_H
-#define _QEMU_MALTA_CONFIG_H
+#ifndef _MALTA_CONFIG_H
+#define _MALTA_CONFIG_H
 
 #include <asm/addrspace.h>
 #include <asm/malta.h>
@@ -13,7 +13,7 @@
 /*
  * System configuration
  */
-#define CONFIG_QEMU_MALTA
+#define CONFIG_MALTA
 
 #define CONFIG_PCI
 #define CONFIG_PCI_GT64120
@@ -54,9 +54,9 @@
  * Console configuration
  */
 #if defined(CONFIG_SYS_LITTLE_ENDIAN)
-#define CONFIG_SYS_PROMPT		"qemu-maltael # "
+#define CONFIG_SYS_PROMPT		"maltael # "
 #else
-#define CONFIG_SYS_PROMPT		"qemu-malta # "
+#define CONFIG_SYS_PROMPT		"malta # "
 #endif
 
 #define CONFIG_SYS_CBSIZE		256
@@ -110,4 +110,4 @@
 
 #define CONFIG_SYS_LONGHELP		/* verbose help, undef to save memory */
 
-#endif /* _QEMU_MALTA_CONFIG_H */
+#endif /* _MALTA_CONFIG_H */
Signature · 1 line
1.8.4.1
#13 lore raw

PATCH v5 7/18 qemu-malta: rename to just "malta"

Dear Paul Burton,

Quoted 6 lines

This is in preparation for adapting this board to function correctly on a physical MIPS Malta board. The board is moved into an "imgtec" vendor directory at the same time in order to ready us for any other boards supported by Imagination in the future.

Will the board still work in QEMU?

Best regards,
Marek Vasut

#14 lore raw

PATCH v5 7/18 qemu-malta: rename to just "malta"

Quoted 7 lines · Marek Vasut

Dear Paul Burton,

Quoted 5 lines

This is in preparation for adapting this board to function correctly on a physical MIPS Malta board. The board is moved into an "imgtec" vendor directory at the same time in order to ready us for any other boards supported by Imagination in the future.

Quoted 7 lines
Quoted 1 line

Will the board still work in QEMU?

Best regards,
Marek Vasut

Yes. It doesn't work in current QEMU master either with or without this series, but it works in older QEMU releases (v1.5.3 for example).

Thanks,
Paul

#15 lore raw

PATCH v5 7/18 qemu-malta: rename to just "malta"

Dear Paul Burton,

Quoted 10 lines · Marek Vasut
Quoted 9 lines

Dear Paul Burton,

Quoted 6 lines

This is in preparation for adapting this board to function correctly on a physical MIPS Malta board. The board is moved into an "imgtec" vendor directory at the same time in order to ready us for any other boards supported by Imagination in the future.

Quoted 7 lines
Quoted 4 lines

Will the board still work in QEMU?

Best regards,
Marek Vasut

Yes. It doesn't work in current QEMU master either with or without this series, but it works in older QEMU releases (v1.5.3 for example).

OK, nice, thanks!

Best regards,
Marek Vasut

#16 lore raw

PATCH v3 8/18 malta: setup super I/O UARTs

In reply to Paul Burton

On a real Malta the Super I/O needs to be configured before we are able to access the UARTs. This patch performs that configuration, setting up the UARTs in the same way that YAMON would.

---
Changes in v3:
- rebase atop master

Changes in v2:
- rebased after malta moved to board/imgtec/malta

---
 board/imgtec/malta/Makefile  |  1 +
 board/imgtec/malta/malta.c   | 10 +++++++
 board/imgtec/malta/superio.c | 63 ++++++++++++++++++++++++++++++++++++++++++++
 board/imgtec/malta/superio.h | 15 +++++++++++
 4 files changed, 89 insertions(+)
 create mode 100644 board/imgtec/malta/superio.c
 create mode 100644 board/imgtec/malta/superio.h
board/imgtec/malta/Makefile +1 -0
index 091830d..19dd3a3 100644
@@ -7,3 +7,4 @@
 
 obj-y	= malta.o
 obj-y	+= lowlevel_init.o
+obj-y	+= superio.o
board/imgtec/malta/malta.c +10 -0
index 7eddf1c..09da9ea 100644
@@ -12,6 +12,8 @@
 #include <asm/malta.h>
 #include <pci_gt64120.h>
 
+#include "superio.h"
+
 phys_size_t initdram(int board_type)
 {
 	return CONFIG_SYS_MEM_SIZE;
@@ -36,6 +38,14 @@ void _machine_restart(void)
 	__raw_writel(GORESET, reset_base);
 }
 
+int board_early_init_f(void)
+{
+	/* setup FDC37M817 super I/O controller */
+	malta_superio_init((void *)CKSEG1ADDR(MALTA_IO_PORT_BASE));
+
+	return 0;
+}
+
 void pci_init_board(void)
 {
 	set_io_port_base(CKSEG1ADDR(MALTA_IO_PORT_BASE));
board/imgtec/malta/superio.c new file +63 -0

63 lines not on this page · raw message

board/imgtec/malta/superio.h new file +15 -0
new file mode 100644
index 0000000..1450da5
@@ -0,0 +1,15 @@
+/*
+ * Copyright (C) 2013 Imagination Technologies
+ * Author: Paul Burton <[email protected]>
+ *
+ * Setup code for the FDC37M817 super I/O controller
+ *
+ * SPDX-License-Identifier:	GPL-2.0+
+ */
+
+#ifndef __BOARD_MALTA_SUPERIO_H__
+#define __BOARD_MALTA_SUPERIO_H__
+
+extern void malta_superio_init(void *io_base);
+
+#endif /* __BOARD_MALTA_SUPERIO_H__ */
Signature · 1 line
1.8.4.1
#17 lore raw

PATCH v3 8/18 malta: setup super I/O UARTs

Dear Paul Burton,

Quoted 3 lines

On a real Malta the Super I/O needs to be configured before we are able to access the UARTs. This patch performs that configuration, setting up the UARTs in the same way that YAMON would.

[...]

board/imgtec/malta/superio.c

Quoted 2 lines
board/imgtec/malta/superio.c new file +0 -0

Do we not already have some kind of a superIO controller driver in U-Boot ? Just wondering, I seem to remember seeing a few. Is this driver re-usable by other boards too maybe?

Best regards,
Marek Vasut

#18 lore raw

PATCH v3 8/18 malta: setup super I/O UARTs

Grepping all I found was board/mpl/common/isa.c which seems to do something similar. However calling this a driver is quite generous - all it does it write some board-specific values to some registers. So there's really not much to be shared.

Thanks,
Paul

Quoted 18 lines · Marek Vasut

Dear Paul Burton,

Quoted 3 lines

On a real Malta the Super I/O needs to be configured before we are able to access the UARTs. This patch performs that configuration, setting up the UARTs in the same way that YAMON would.

[...]

board/imgtec/malta/superio.c

Quoted 2 lines
board/imgtec/malta/superio.c new file +0 -0

Do we not already have some kind of a superIO controller driver in U-Boot ? Just wondering, I seem to remember seeing a few. Is this driver re-usable by other boards too maybe?

Best regards,
Marek Vasut

#19 lore raw

PATCH v3 8/18 malta: setup super I/O UARTs

Dear Paul Burton,

Quoted 4 lines

Grepping all I found was board/mpl/common/isa.c which seems to do something similar. However calling this a driver is quite generous - all it does it write some board-specific values to some registers. So there's really not much to be shared.

Thanks for clearing this. So all super-IOs do share among themselves is the name ;-)

Best regards,
Marek Vasut

#20 lore raw

PATCH v3 9/18 malta: support for coreFPGA6 boards

In reply to Paul Burton

This patch adds support for running on Malta boards using coreFPGA6 core cards, including support for the msc01 system controller used with them. The system controller is detected at runtime allowing one U-boot binary to run on a Malta with either.

Due to the PCI I/O base differing between Maltas using gt64120 & msc01 system controllers, the UART setup is modified slightly. A second UART is added so that there is one pointing at the correct address for each system controller. The Malta board then defines its own default_serial_console function to select the correct one at runtime. The incorrect UART will simply not function.

Tested on:
- A coreFPGA6 Malta running interAptiv and proAptiv bitstreams, both with and without an L2 cache.
- QEMU.

---
Changes in v3:
- rebase atop master
- remove commit message note about not working in QEMU, this does work in older QEMU versions as noted by Gabor Juhos

Changes in v2:
- rebased after malta moved to board/imgtec/malta
- removed unused status variable from msc01_config_access

---
 arch/mips/include/asm/malta.h      |  35 ++++++--
 board/imgtec/malta/lowlevel_init.S | 164 ++++++++++++++++++++++++++++++++++++-
 board/imgtec/malta/malta.c         | 126 ++++++++++++++++++++++++++--
 drivers/pci/Makefile               |   1 +
 drivers/pci/pci_msc01.c            | 125 ++++++++++++++++++++++++++++
 include/configs/malta.h            |   4 +-
 include/msc01.h                    | 135 ++++++++++++++++++++++++++++++
 include/pci_ids.h                  |   3 +
 include/pci_msc01.h                |  17 ++++
 9 files changed, 594 insertions(+), 16 deletions(-)
 create mode 100644 drivers/pci/pci_msc01.c
 create mode 100644 include/msc01.h
 create mode 100644 include/pci_msc01.h
arch/mips/include/asm/malta.h +29 -6
index d4d44a2..0b50a66 100644
@@ -9,15 +9,38 @@
 #ifndef _MIPS_ASM_MALTA_H
 #define _MIPS_ASM_MALTA_H
 
-#define MALTA_IO_PORT_BASE	0x18000000
+#define MALTA_GT_BASE			0x1be00000
+#define MALTA_GT_PCIIO_BASE		0x18000000
+#define MALTA_GT_UART0_BASE		(MALTA_GT_PCIIO_BASE + 0x3f8)
 
-#define MALTA_UART_BASE		(MALTA_IO_PORT_BASE + 0x3f8)
+#define MALTA_MSC01_BIU_BASE		0x1bc80000
+#define MALTA_MSC01_PCI_BASE		0x1bd00000
+#define MALTA_MSC01_PBC_BASE		0x1bd40000
+#define MALTA_MSC01_IP1_BASE		0x1bc00000
+#define MALTA_MSC01_IP1_SIZE		0x00400000
+#define MALTA_MSC01_IP2_BASE1		0x10000000
+#define MALTA_MSC01_IP2_SIZE1		0x08000000
+#define MALTA_MSC01_IP2_BASE2		0x18000000
+#define MALTA_MSC01_IP2_SIZE2		0x04000000
+#define MALTA_MSC01_IP3_BASE		0x1c000000
+#define MALTA_MSC01_IP3_SIZE		0x04000000
+#define MALTA_MSC01_PCIMEM_BASE		0x10000000
+#define MALTA_MSC01_PCIMEM_SIZE		0x10000000
+#define MALTA_MSC01_PCIMEM_MAP		0x10000000
+#define MALTA_MSC01_PCIIO_BASE		0x1b000000
+#define MALTA_MSC01_PCIIO_SIZE		0x00800000
+#define MALTA_MSC01_PCIIO_MAP		0x00000000
+#define MALTA_MSC01_UART0_BASE		(MALTA_MSC01_PCIIO_BASE + 0x3f8)
 
-#define MALTA_GT_BASE		0x1be00000
+#define MALTA_RESET_BASE		0x1f000500
+#define GORESET				0x42
 
-#define MALTA_RESET_BASE	0x1f000500
-#define GORESET			0x42
+#define MALTA_FLASH_BASE		0x1fc00000
 
-#define MALTA_FLASH_BASE	0x1fc00000
+#define MALTA_REVISION			0x1fc00010
+#define MALTA_REVISION_CORID_SHF	10
+#define MALTA_REVISION_CORID_MSK	(0x3f << MALTA_REVISION_CORID_SHF)
+#define MALTA_REVISION_CORID_CORE_LV		1
+#define MALTA_REVISION_CORID_CORE_FPGA6		14
 
 #endif /* _MIPS_ASM_MALTA_H */
board/imgtec/malta/lowlevel_init.S +163 -1

185 lines not on this page · raw message

board/imgtec/malta/malta.c +118 -8

162 lines not on this page · raw message

drivers/pci/Makefile +1 -0
index 99d51a6..6182a59 100644
@@ -9,6 +9,7 @@ obj-$(CONFIG_FSL_PCI_INIT) += fsl_pci_init.o
 obj-$(CONFIG_PCI) += pci.o pci_auto.o
 obj-$(CONFIG_PCI_INDIRECT_BRIDGE) += pci_indirect.o
 obj-$(CONFIG_PCI_GT64120) += pci_gt64120.o
+obj-$(CONFIG_PCI_MSC01) += pci_msc01.o
 obj-$(CONFIG_FTPCI100) += pci_ftpci100.o
 obj-$(CONFIG_IXP_PCI) += pci_ixp.o
 obj-$(CONFIG_SH4_PCI) += pci_sh4.o
drivers/pci/pci_msc01.c new file +125 -0

125 lines not on this page · raw message

include/configs/malta.h +3 -1
index 4ca8bad..737dd39 100644
@@ -17,6 +17,7 @@
 
 #define CONFIG_PCI
 #define CONFIG_PCI_GT64120
+#define CONFIG_PCI_MSC01
 #define CONFIG_PCI_PNP
 #define CONFIG_PCNET
 
@@ -76,7 +77,8 @@
 #define CONFIG_SYS_NS16550_SERIAL
 #define CONFIG_SYS_NS16550_REG_SIZE	1
 #define CONFIG_SYS_NS16550_CLK		115200
-#define CONFIG_SYS_NS16550_COM1		CKSEG1ADDR(MALTA_UART_BASE)
+#define CONFIG_SYS_NS16550_COM1		CKSEG1ADDR(MALTA_GT_UART0_BASE)
+#define CONFIG_SYS_NS16550_COM2		CKSEG1ADDR(MALTA_MSC01_UART0_BASE)
 #define CONFIG_CONS_INDEX		1
 
 /*
include/msc01.h new file +135 -0

135 lines not on this page · raw message

include/pci_ids.h +3 -0
index 2c6dfd4..6bab677 100644
@@ -2170,6 +2170,9 @@
 #define PCI_DEVICE_ID_ENE_720		0x1421
 #define PCI_DEVICE_ID_ENE_722		0x1422
 
+#define PCI_VENDOR_ID_MIPS		0x153f
+#define PCI_DEVICE_ID_MIPS_MSC01	0x0001
+
 #define PCI_SUBVENDOR_ID_PERLE          0x155f
 #define PCI_SUBDEVICE_ID_PCI_RAS4       0xf001
 #define PCI_SUBDEVICE_ID_PCI_RAS8       0xf010
include/pci_msc01.h new file +17 -0
new file mode 100644
index 0000000..54945a7
@@ -0,0 +1,17 @@
+/*
+ * Copyright (C) 2013 Imagination Technologies
+ * Author: Paul Burton <[email protected]>
+ *
+ * SPDX-License-Identifier:	GPL-2.0+
+ */
+
+#ifndef __PCI_MSC01_H__
+#define __PCI_MSC01_H__
+
+extern void msc01_pci_init(void *base, unsigned long sys_bus,
+			   unsigned long sys_phys, unsigned long sys_size,
+			   unsigned long mem_bus, unsigned long mem_phys,
+			   unsigned long mem_size, unsigned long io_bus,
+			   unsigned long io_phys, unsigned long io_size);
+
+#endif /* __PCI_MSC01_H__ */
Signature · 1 line
1.8.4.1
#21 lore raw

PATCH v3 9/18 malta: support for coreFPGA6 boards

Dear Paul Burton,

arch/mips/include/asm/malta.h

Quoted 47 lines

This patch adds support for running on Malta boards using coreFPGA6 core cards, including support for the msc01 system controller used with them. The system controller is detected at runtime allowing one U-boot binary to run on a Malta with either.

Due to the PCI I/O base differing between Maltas using gt64120 & msc01 system controllers, the UART setup is modified slightly. A second UART is added so that there is one pointing at the correct address for each system controller. The Malta board then defines its own default_serial_console function to select the correct one at runtime. The incorrect UART will simply not function.

Tested on:
- A coreFPGA6 Malta running interAptiv and proAptiv bitstreams, both with and without an L2 cache.
- QEMU.

---
Changes in v3:
- rebase atop master
- remove commit message note about not working in QEMU, this does work in older QEMU versions as noted by Gabor Juhos

Changes in v2:
- rebased after malta moved to board/imgtec/malta
- removed unused status variable from msc01_config_access

---
 arch/mips/include/asm/malta.h      |  35 ++++++--
 board/imgtec/malta/lowlevel_init.S | 164

++++++++++++++++++++++++++++++++++++- board/imgtec/malta/malta.c |
126 ++++++++++++++++++++++++++-- drivers/pci/Makefile | 1

(quoted diff) +1 -0
+
 drivers/pci/pci_msc01.c            | 125 ++++++++++++++++++++++++++++
 include/configs/malta.h            |   4 +-
 include/msc01.h                    | 135 ++++++++++++++++++++++++++++++
 include/pci_ids.h                  |   3 +
 include/pci_msc01.h                |  17 ++++
 9 files changed, 594 insertions(+), 16 deletions(-)
 create mode 100644 drivers/pci/pci_msc01.c
 create mode 100644 include/msc01.h
 create mode 100644 include/pci_msc01.h
arch/mips/include/asm/malta.h +0 -0
index d4d44a2..0b50a66 100644

You are missing license in some files.

Best regards,
Marek Vasut

#22 lore raw

PATCH v3 9/18 malta: support for coreFPGA6 boards

Quoted 52 lines · Marek Vasut

Dear Paul Burton,

arch/mips/include/asm/malta.h

Quoted 47 lines

This patch adds support for running on Malta boards using coreFPGA6 core cards, including support for the msc01 system controller used with them. The system controller is detected at runtime allowing one U-boot binary to run on a Malta with either.

Due to the PCI I/O base differing between Maltas using gt64120 & msc01 system controllers, the UART setup is modified slightly. A second UART is added so that there is one pointing at the correct address for each system controller. The Malta board then defines its own default_serial_console function to select the correct one at runtime. The incorrect UART will simply not function.

Tested on:
- A coreFPGA6 Malta running interAptiv and proAptiv bitstreams, both with and without an L2 cache.
- QEMU.

---
Changes in v3:
- rebase atop master
- remove commit message note about not working in QEMU, this does work in older QEMU versions as noted by Gabor Juhos

Changes in v2:
- rebased after malta moved to board/imgtec/malta
- removed unused status variable from msc01_config_access

---
  arch/mips/include/asm/malta.h      |  35 ++++++--
  board/imgtec/malta/lowlevel_init.S | 164

++++++++++++++++++++++++++++++++++++- board/imgtec/malta/malta.c |
126 ++++++++++++++++++++++++++-- drivers/pci/Makefile | 1

(quoted diff) +1 -0
+
  drivers/pci/pci_msc01.c            | 125 ++++++++++++++++++++++++++++
  include/configs/malta.h            |   4 +-
  include/msc01.h                    | 135 ++++++++++++++++++++++++++++++
  include/pci_ids.h                  |   3 +
  include/pci_msc01.h                |  17 ++++
  9 files changed, 594 insertions(+), 16 deletions(-)
  create mode 100644 drivers/pci/pci_msc01.c
  create mode 100644 include/msc01.h
  create mode 100644 include/pci_msc01.h
arch/mips/include/asm/malta.h +0 -0
index d4d44a2..0b50a66 100644

You are missing license in some files.

Quoted 3 lines

Best regards,
Marek Vasut

Which files? The only files this patch adds are:

drivers/pci/pci_msc01.c
include/msc01.h
include/pci_msc01.h

...and all 3 of them contain GPL-2.0+ SPDX tags.

Thanks,
Paul

#23 lore raw

PATCH v3 9/18 malta: support for coreFPGA6 boards

2013/11/11 Paul Burton <[email protected]>:

Quoted 65 lines · Marek Vasut
Quoted 56 lines

Dear Paul Burton,

arch/mips/include/asm/malta.h

Quoted 49 lines

This patch adds support for running on Malta boards using coreFPGA6 core cards, including support for the msc01 system controller used with them. The system controller is detected at runtime allowing one U-boot binary to run on a Malta with either.

Due to the PCI I/O base differing between Maltas using gt64120 & msc01 system controllers, the UART setup is modified slightly. A second UART is added so that there is one pointing at the correct address for each system controller. The Malta board then defines its own default_serial_console function to select the correct one at runtime. The incorrect UART will simply not function.

Tested on:
- A coreFPGA6 Malta running interAptiv and proAptiv bitstreams, both with and without an L2 cache.
- QEMU.

---
Changes in v3:
- rebase atop master
- remove commit message note about not working in QEMU, this does work in older QEMU versions as noted by Gabor Juhos

Changes in v2:
- rebased after malta moved to board/imgtec/malta
- removed unused status variable from msc01_config_access

---
  arch/mips/include/asm/malta.h      |  35 ++++++--
  board/imgtec/malta/lowlevel_init.S | 164

++++++++++++++++++++++++++++++++++++- board/imgtec/malta/malta.c |
126 ++++++++++++++++++++++++++-- drivers/pci/Makefile | 1

(quoted diff) +1 -0
+
  drivers/pci/pci_msc01.c            | 125 ++++++++++++++++++++++++++++
  include/configs/malta.h            |   4 +-
  include/msc01.h                    | 135 ++++++++++++++++++++++++++++++
  include/pci_ids.h                  |   3 +
  include/pci_msc01.h                |  17 ++++
  9 files changed, 594 insertions(+), 16 deletions(-)
  create mode 100644 drivers/pci/pci_msc01.c
  create mode 100644 include/msc01.h
  create mode 100644 include/pci_msc01.h
arch/mips/include/asm/malta.h +0 -0
index d4d44a2..0b50a66 100644

You are missing license in some files.

Quoted 3 lines

Best regards,
Marek Vasut

Which files? The only files this patch adds are:

Quoted 9 lines

drivers/pci/pci_msc01.c
include/msc01.h
include/pci_msc01.h

...and all 3 of them contain GPL-2.0+ SPDX tags.

Thanks,
Paul

only arch/mips/include/asm/malta.h is missing the SPDX tag.

Paul, could you send a patch? Maybe you want to add your copyright, because you modified parts of that file.

Signature · 1 line
- Daniel
#24 lore raw

PATCH 1/1 malta: arch/mips/include/asm/malta.h SPDX license tag

This patch replaces the GPL-2.0 text with a GPL-2.0 SPDX-License-Identifier tag, and adds Imagination Technologies copyright following my recent changes.

---
 arch/mips/include/asm/malta.h | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
arch/mips/include/asm/malta.h +2 -3
index e141eb0..9b1100b 100644
@@ -1,9 +1,8 @@
 /*
  * Copyright (C) 2013 Gabor Juhos <[email protected]>
+ * Copyright (C) 2013 Imagination Technologies
  *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License version 2 as published
- * by the Free Software Foundation.
+ * SPDX-License-Identifier:	GPL-2.0
  */
 
 #ifndef _MIPS_ASM_MALTA_H
Signature · 1 line
1.8.4.1
#25 lore raw

PATCH 1/1 malta: arch/mips/include/asm/malta.h SPDX license tag

2013/11/11 Paul Burton <[email protected]>:

Quoted 9 lines

This patch replaces the GPL-2.0 text with a GPL-2.0 SPDX-License-Identifier tag, and adds Imagination Technologies copyright following my recent changes.

---
 arch/mips/include/asm/malta.h | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

applied to u-boot-mips/master, thanks

Signature · 1 line
- Daniel
#26 lore raw

PATCH v3 10/18 malta: display "U-boot" on the LCD screen

In reply to Paul Burton

Displaying a message on the LCD screen is a simple yet effective way to show the user that the board has booted successfully.

---
Changes in v3:
- rebase atop master

Changes in v2:
- rebased after malta moved to board/imgtec/malta

---
 arch/mips/include/asm/malta.h | 10 ++++++++++
 board/imgtec/malta/malta.c    | 19 +++++++++++++++++++
 2 files changed, 29 insertions(+)
arch/mips/include/asm/malta.h +10 -0
index 0b50a66..d8ec57c 100644
@@ -32,6 +32,16 @@
 #define MALTA_MSC01_PCIIO_MAP		0x00000000
 #define MALTA_MSC01_UART0_BASE		(MALTA_MSC01_PCIIO_BASE + 0x3f8)
 
+#define MALTA_ASCIIWORD			0x1f000410
+#define MALTA_ASCIIPOS0			0x1f000418
+#define MALTA_ASCIIPOS1			0x1f000420
+#define MALTA_ASCIIPOS2			0x1f000428
+#define MALTA_ASCIIPOS3			0x1f000430
+#define MALTA_ASCIIPOS4			0x1f000438
+#define MALTA_ASCIIPOS5			0x1f000440
+#define MALTA_ASCIIPOS6			0x1f000448
+#define MALTA_ASCIIPOS7			0x1f000450
+
 #define MALTA_RESET_BASE		0x1f000500
 #define GORESET				0x42
 
board/imgtec/malta/malta.c +19 -0
index 2af0067..119546a 100644
@@ -29,6 +29,24 @@ enum sys_con {
 	SYSCON_MSC01,
 };
 
+static void malta_lcd_puts(const char *str)
+{
+	int i;
+	void *reg = (void *)CKSEG1ADDR(MALTA_ASCIIPOS0);
+
+	/* print up to 8 characters of the string */
+	for (i = 0; i < min(strlen(str), 8); i++) {
+		__raw_writel(str[i], reg);
+		reg += MALTA_ASCIIPOS1 - MALTA_ASCIIPOS0;
+	}
+
+	/* fill the rest of the display with spaces */
+	for (; i < 8; i++) {
+		__raw_writel(' ', reg);
+		reg += MALTA_ASCIIPOS1 - MALTA_ASCIIPOS0;
+	}
+}
+
 static enum core_card malta_core_card(void)
 {
 	u32 corid, rev;
@@ -71,6 +89,7 @@ int checkboard(void)
 {
 	enum core_card core;
 
+	malta_lcd_puts("U-boot");
 	puts("Board: MIPS Malta");
 
 	core = malta_core_card();
Signature · 1 line
1.8.4.1
#27 lore raw

PATCH v3 10/18 malta: display "U-boot" on the LCD screen

Dear Paul Burton,

Quoted 66 lines

Displaying a message on the LCD screen is a simple yet effective way to show the user that the board has booted successfully.

---
Changes in v3:
- rebase atop master

Changes in v2:
- rebased after malta moved to board/imgtec/malta

---
 arch/mips/include/asm/malta.h | 10 ++++++++++
 board/imgtec/malta/malta.c    | 19 +++++++++++++++++++
 2 files changed, 29 insertions(+)
arch/mips/include/asm/malta.h +10 -0
index 0b50a66..d8ec57c 100644
@@ -32,6 +32,16 @@
 #define MALTA_MSC01_PCIIO_MAP		0x00000000
 #define MALTA_MSC01_UART0_BASE		(MALTA_MSC01_PCIIO_BASE + 0x3f8)
 
+#define MALTA_ASCIIWORD			0x1f000410
+#define MALTA_ASCIIPOS0			0x1f000418
+#define MALTA_ASCIIPOS1			0x1f000420
+#define MALTA_ASCIIPOS2			0x1f000428
+#define MALTA_ASCIIPOS3			0x1f000430
+#define MALTA_ASCIIPOS4			0x1f000438
+#define MALTA_ASCIIPOS5			0x1f000440
+#define MALTA_ASCIIPOS6			0x1f000448
+#define MALTA_ASCIIPOS7			0x1f000450
+
 #define MALTA_RESET_BASE		0x1f000500
 #define GORESET				0x42
 
board/imgtec/malta/malta.c +18 -0
index 2af0067..119546a 100644
@@ -29,6 +29,24 @@ enum sys_con {
 	SYSCON_MSC01,
 };
 
+static void malta_lcd_puts(const char *str)
+{
+	int i;
+	void *reg = (void *)CKSEG1ADDR(MALTA_ASCIIPOS0);
+
+	/* print up to 8 characters of the string */
+	for (i = 0; i < min(strlen(str), 8); i++) {
+		__raw_writel(str[i], reg);
+		reg += MALTA_ASCIIPOS1 - MALTA_ASCIIPOS0;
+	}
+
+	/* fill the rest of the display with spaces */
+	for (; i < 8; i++) {
+		__raw_writel(' ', reg);
+		reg += MALTA_ASCIIPOS1 - MALTA_ASCIIPOS0;
+	}
+}
+
 static enum core_card malta_core_card(void)
 {
 	u32 corid, rev;

board/imgtec/malta/malta.c int checkboard(void) · line 95

Quoted 7 lines
board/imgtec/malta/malta.c +1 -0
@@ -71,6 +89,7 @@ int checkboard(void)
 {
 	enum core_card core;
 
+	malta_lcd_puts("U-boot");
 	puts("Board: MIPS Malta");
 
 	core = malta_core_card();

Can this not be implemented as a regular framebuffer console maybe?

Best regards,
Marek Vasut

#28 lore raw

PATCH v3 10/18 malta: display "U-boot" on the LCD screen

Quoted 74 lines · Marek Vasut

Dear Paul Burton,

Quoted 72 lines

Displaying a message on the LCD screen is a simple yet effective way to show the user that the board has booted successfully.

---
Changes in v3:
- rebase atop master

Changes in v2:
- rebased after malta moved to board/imgtec/malta

---
  arch/mips/include/asm/malta.h | 10 ++++++++++
  board/imgtec/malta/malta.c    | 19 +++++++++++++++++++
  2 files changed, 29 insertions(+)
arch/mips/include/asm/malta.h +10 -0
index 0b50a66..d8ec57c 100644
@@ -32,6 +32,16 @@
  #define MALTA_MSC01_PCIIO_MAP		0x00000000
  #define MALTA_MSC01_UART0_BASE		(MALTA_MSC01_PCIIO_BASE + 0x3f8)
 
+#define MALTA_ASCIIWORD			0x1f000410
+#define MALTA_ASCIIPOS0			0x1f000418
+#define MALTA_ASCIIPOS1			0x1f000420
+#define MALTA_ASCIIPOS2			0x1f000428
+#define MALTA_ASCIIPOS3			0x1f000430
+#define MALTA_ASCIIPOS4			0x1f000438
+#define MALTA_ASCIIPOS5			0x1f000440
+#define MALTA_ASCIIPOS6			0x1f000448
+#define MALTA_ASCIIPOS7			0x1f000450
+
  #define MALTA_RESET_BASE		0x1f000500
  #define GORESET				0x42
 
board/imgtec/malta/malta.c +19 -0
index 2af0067..119546a 100644
@@ -29,6 +29,24 @@ enum sys_con {
  	SYSCON_MSC01,
  };
 
+static void malta_lcd_puts(const char *str)
+{
+	int i;
+	void *reg = (void *)CKSEG1ADDR(MALTA_ASCIIPOS0);
+
+	/* print up to 8 characters of the string */
+	for (i = 0; i < min(strlen(str), 8); i++) {
+		__raw_writel(str[i], reg);
+		reg += MALTA_ASCIIPOS1 - MALTA_ASCIIPOS0;
+	}
+
+	/* fill the rest of the display with spaces */
+	for (; i < 8; i++) {
+		__raw_writel(' ', reg);
+		reg += MALTA_ASCIIPOS1 - MALTA_ASCIIPOS0;
+	}
+}
+
  static enum core_card malta_core_card(void)
  {
  	u32 corid, rev;
@@ -71,6 +89,7 @@ int checkboard(void)
  {
  	enum core_card core;
 
+	malta_lcd_puts("U-boot");
  	puts("Board: MIPS Malta");
 
Quoted 7 lines

board/imgtec/malta/malta.c int checkboard(void) · line 95

Quoted 1 line
board/imgtec/malta/malta.c +0 -0
  	core = malta_core_card();

Can this not be implemented as a regular framebuffer console maybe?

Best regards,
Marek Vasut

We're talking about an 8 character segmented LCD display here - it's not something you can draw (non-ASCII) graphics on, nor something you want to be interacting with even if you hooked up some input.

Thanks,
Paul

#29 lore raw

PATCH v3 10/18 malta: display "U-boot" on the LCD screen

Dear Paul Burton,

Quoted 21 lines · Marek Vasut
Quoted 20 lines

Dear Paul Burton,

Quoted 18 lines

Displaying a message on the LCD screen is a simple yet effective way to show the user that the board has booted successfully.

---

Changes in v3:
- rebase atop master

Changes in v2:
- rebased after malta moved to board/imgtec/malta

---

  arch/mips/include/asm/malta.h | 10 ++++++++++
  board/imgtec/malta/malta.c    | 19 +++++++++++++++++++
  2 files changed, 29 insertions(+)
Quoted 8 lines
Quoted 8 lines

arch/mips/include/asm/malta.h line 34

Quoted 8 lines
(quoted diff) +0 -0

b/arch/mips/include/asm/malta.h index 0b50a66..d8ec57c 100644

arch/mips/include/asm/malta.h +0 -0
@@ -32,6 +32,16 @@
 
  #define MALTA_MSC01_PCIIO_MAP		0x00000000
  #define MALTA_MSC01_UART0_BASE		(MALTA_MSC01_PCIIO_BASE + 

0x3f8)

Quoted 63 lines
Quoted 63 lines
Quoted 57 lines
arch/mips/include/asm/malta.h +10 -0
+#define MALTA_ASCIIWORD			0x1f000410
+#define MALTA_ASCIIPOS0			0x1f000418
+#define MALTA_ASCIIPOS1			0x1f000420
+#define MALTA_ASCIIPOS2			0x1f000428
+#define MALTA_ASCIIPOS3			0x1f000430
+#define MALTA_ASCIIPOS4			0x1f000438
+#define MALTA_ASCIIPOS5			0x1f000440
+#define MALTA_ASCIIPOS6			0x1f000448
+#define MALTA_ASCIIPOS7			0x1f000450
+
 
  #define MALTA_RESET_BASE		0x1f000500
  #define GORESET				0x42
board/imgtec/malta/malta.c +18 -0
index 2af0067..119546a 100644
@@ -29,6 +29,24 @@ enum sys_con {
 
  	SYSCON_MSC01,
 
  };
 
+static void malta_lcd_puts(const char *str)
+{
+	int i;
+	void *reg = (void *)CKSEG1ADDR(MALTA_ASCIIPOS0);
+
+	/* print up to 8 characters of the string */
+	for (i = 0; i < min(strlen(str), 8); i++) {
+		__raw_writel(str[i], reg);
+		reg += MALTA_ASCIIPOS1 - MALTA_ASCIIPOS0;
+	}
+
+	/* fill the rest of the display with spaces */
+	for (; i < 8; i++) {
+		__raw_writel(' ', reg);
+		reg += MALTA_ASCIIPOS1 - MALTA_ASCIIPOS0;
+	}
+}
+
 
  static enum core_card malta_core_card(void)
  {

u32 corid, rev;

board/imgtec/malta/malta.c +1 -0
@@ -71,6 +89,7 @@ int checkboard(void)
 
  {
 
  	enum core_card core;
 
+	malta_lcd_puts("U-boot");
 
Quoted 3 lines
  	puts("Board: MIPS Malta");
  	
  	core = malta_core_card();

Can this not be implemented as a regular framebuffer console maybe?

Quoted 6 lines
Quoted 2 lines

Best regards,
Marek Vasut

We're talking about an 8 character segmented LCD display here - it's not something you can draw (non-ASCII) graphics on, nor something you want to be interacting with even if you hooked up some input.

Oh OK, I see.

#30 lore raw

PATCH v2 11/18 malta: enable CONFIG_PCNET_79C973, PCNET_HAS_PROM, CONFIG_CMD_DHCP

In reply to Paul Burton

This model of the pcnet is used in current Malta boards, at least in the Malta-R rev 3. Enable support for it.

The Malta also has the ethernet controller PROM containing its MAC address, so enable support for that in order to read that MAC address.

DHCP is a very useful feature to have available for many networks, enable support for it also.

---
Changes in v2:
- rebase atop master

---
 include/configs/malta.h | 3 +++
 1 file changed, 3 insertions(+)
include/configs/malta.h +3 -0
index 737dd39..a0c5b79 100644
@@ -20,6 +20,8 @@
 #define CONFIG_PCI_MSC01
 #define CONFIG_PCI_PNP
 #define CONFIG_PCNET
+#define CONFIG_PCNET_79C973
+#define PCNET_HAS_PROM
 
 /*
  * CPU Configuration
@@ -107,6 +109,7 @@
 #undef CONFIG_CMD_LOADS
 #undef CONFIG_CMD_NFS
 
+#define CONFIG_CMD_DHCP
 #define CONFIG_CMD_PCI
 #define CONFIG_CMD_PING
 
Signature · 1 line
1.8.4.1
#31 lore raw

PATCH v2 12/18 malta: remove cache size definitions

In reply to Paul Burton

These will now be detected at runtime, allowing a single U-boot configuration to function correctly with different bitstreams. Without this you may need to re-configure, re-build and re-flash U-boot to your Malta if you flash a new bitstream with a different cache configuration to your old bitstream.

---
Changes in v2:
- rebase atop master

---
 include/configs/malta.h | 4 ----
 1 file changed, 4 deletions(-)
include/configs/malta.h +0 -4
index a0c5b79..5368308 100644
@@ -29,10 +29,6 @@
 #define CONFIG_SYS_MHZ			250	/* arbitrary value */
 #define CONFIG_SYS_MIPS_TIMER_FREQ	(CONFIG_SYS_MHZ * 1000000)
 
-#define CONFIG_SYS_DCACHE_SIZE		16384	/* arbitrary value */
-#define CONFIG_SYS_ICACHE_SIZE		16384	/* arbitrary value */
-#define CONFIG_SYS_CACHELINE_SIZE	32	/* arbitrary value */
-
 #define CONFIG_SWAP_IO_SPACE
 
 /*
Signature · 1 line
1.8.4.1
#32 lore raw

PATCH v3 13/18 malta: disable L2 caches

In reply to Paul Burton

Malta boards may be used with cores which support L2 caches, however U-boot does not yet support L2 cache for MIPS. Thus for the moment we'll disable L2 caches by setting the L2B bit in Config2. This is specific to MTI/Imagination MIPS cores which is why this is done for the Malta board rather than generically.

---
Changes in v3:
- rebase atop master

Changes in v2:
- rebased after malta moved to board/imgtec/malta

---
 board/imgtec/malta/lowlevel_init.S | 7 +++++++
 1 file changed, 7 insertions(+)
board/imgtec/malta/lowlevel_init.S +7 -0
index 1af34f1..ae09c27 100644
@@ -12,6 +12,7 @@
 #include <asm/addrspace.h>
 #include <asm/regdef.h>
 #include <asm/malta.h>
+#include <asm/mipsregs.h>
 
 #ifdef CONFIG_SYS_BIG_ENDIAN
 #define CPU_TO_GT32(_x)		((_x))
@@ -27,6 +28,12 @@
 
 	.globl	lowlevel_init
 lowlevel_init:
+	/* disable any L2 cache for now */
+	sync
+	mfc0	t0, CP0_CONFIG, 2
+	ori	t0, t0, 0x1 << 12
+	mtc0	t0, CP0_CONFIG, 2
+
 	/* detect the core card */
 	li	t0, KSEG1ADDR(MALTA_REVISION)
 	lw	t0, 0(t0)
Signature · 1 line
1.8.4.1
#33 lore raw

PATCH v2 14/18 malta: enable RTC support

In reply to Paul Burton

This is actually required in order for a Linux kernel to boot successfully on a physical Malta board. Without enabling the RTC, a Malta Linux kernel will get stuck in its estimate_frequencies function on boot.

---
Changes in v2:
- rebase atop master
- included within "MIPS Malta board support series"

---
 board/imgtec/malta/malta.c | 8 ++++++++
 drivers/rtc/mc146818.c     | 2 +-
 include/configs/malta.h    | 5 +++++
 3 files changed, 14 insertions(+), 1 deletion(-)
board/imgtec/malta/malta.c +8 -0
index 119546a..2f92259 100644
@@ -9,6 +9,7 @@
 #include <netdev.h>
 #include <pci_gt64120.h>
 #include <pci_msc01.h>
+#include <rtc.h>
 #include <serial.h>
 
 #include <asm/addrspace.h>
@@ -147,6 +148,13 @@ int board_early_init_f(void)
 	return 0;
 }
 
+int misc_init_r(void)
+{
+	rtc_reset();
+
+	return 0;
+}
+
 struct serial_device *default_serial_console(void)
 {
 	switch (malta_sys_con()) {
drivers/rtc/mc146818.c +1 -1
index 5f9d359..f7cf106 100644
@@ -15,7 +15,7 @@
 #include <command.h>
 #include <rtc.h>
 
-#ifdef __I386__
+#if defined(__I386__) || defined(CONFIG_MALTA)
 #include <asm/io.h>
 #define in8(p) inb(p)
 #define out8(p, v) outb(v, p)
include/configs/malta.h +5 -0
index 5368308..230aa56 100644
@@ -23,6 +23,10 @@
 #define CONFIG_PCNET_79C973
 #define PCNET_HAS_PROM
 
+#define CONFIG_MISC_INIT_R
+#define CONFIG_RTC_MC146818
+#define CONFIG_SYS_ISA_IO_BASE_ADDRESS	0
+
 /*
  * CPU Configuration
  */
@@ -105,6 +109,7 @@
 #undef CONFIG_CMD_LOADS
 #undef CONFIG_CMD_NFS
 
+#define CONFIG_CMD_DATE
 #define CONFIG_CMD_DHCP
 #define CONFIG_CMD_PCI
 #define CONFIG_CMD_PING
Signature · 1 line
1.8.4.1
#34 lore raw

PATCH 15/18 malta: store environment in flash

In reply to Paul Burton

Allow the environment to be stored in the monitor flash of a Malta board. The environment is stored in the final 128KB of the flash, which both leaves the majority of the flash available for U-boot code and also matches the location which YAMON uses.

---
 include/configs/malta.h | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)
include/configs/malta.h +9 -6
index 230aa56..58428c6 100644
@@ -84,12 +84,6 @@
 #define CONFIG_CONS_INDEX		1
 
 /*
- * Environment
- */
-#define CONFIG_ENV_IS_NOWHERE
-#define CONFIG_ENV_SIZE			0x10000
-
-/*
  * Flash configuration
  */
 #define CONFIG_SYS_FLASH_BASE		(KSEG1 | MALTA_FLASH_BASE)
@@ -100,6 +94,15 @@
 #define CONFIG_SYS_FLASH_USE_BUFFER_WRITE
 
 /*
+ * Environment
+ */
+#define CONFIG_ENV_IS_IN_FLASH
+#define CONFIG_ENV_SECT_SIZE		0x20000
+#define CONFIG_ENV_SIZE			CONFIG_ENV_SECT_SIZE
+#define CONFIG_ENV_ADDR \
+	(CONFIG_SYS_FLASH_BASE + (4 << 20) - CONFIG_ENV_SIZE)
+
+/*
  * Commands
  */
 #include <config_cmd_default.h>
Signature · 1 line
1.8.4.1
#35 lore raw

PATCH 16/18 malta: setup PIIX4 interrupt route

In reply to Paul Burton

Without setting up the PIRQ[A:D] interrupt routes, PCI interrupts will be left disabled. Linux does not set up this routing but relies upon it having been set up by the bootloader, reading back the IRQ lines which the PIRQ[A:D] signals have been routed to.

This patch routes PIRQA & PIRQB to IRQ 10, and PIRQC & PIRQD to IRQ 11. This matches the setup used by YAMON.

---
 arch/mips/include/asm/malta.h |  5 +++++
 board/imgtec/malta/malta.c    | 14 ++++++++++++++
 2 files changed, 19 insertions(+)
arch/mips/include/asm/malta.h +5 -0
index d8ec57c..e141eb0 100644
@@ -53,4 +53,9 @@
 #define MALTA_REVISION_CORID_CORE_LV		1
 #define MALTA_REVISION_CORID_CORE_FPGA6		14
 
+#define PCI_CFG_PIIX4_PIRQRCA		0x60
+#define PCI_CFG_PIIX4_PIRQRCB		0x61
+#define PCI_CFG_PIIX4_PIRQRCC		0x62
+#define PCI_CFG_PIIX4_PIRQRCD		0x63
+
 #endif /* _MIPS_ASM_MALTA_H */
board/imgtec/malta/malta.c +14 -0
index 2f92259..a1a4c01 100644
@@ -7,6 +7,7 @@
 
 #include <common.h>
 #include <netdev.h>
+#include <pci.h>
 #include <pci_gt64120.h>
 #include <pci_msc01.h>
 #include <rtc.h>
@@ -169,6 +170,8 @@ struct serial_device *default_serial_console(void)
 
 void pci_init_board(void)
 {
+	pci_dev_t bdf;
+
 	switch (malta_sys_con()) {
 	case SYSCON_GT64120:
 		set_io_port_base(CKSEG1ADDR(MALTA_GT_PCIIO_BASE));
@@ -191,4 +194,15 @@ void pci_init_board(void)
 			       0x00000000, MALTA_MSC01_PCIIO_SIZE);
 		break;
 	}
+
+	bdf = pci_find_device(PCI_VENDOR_ID_INTEL,
+			      PCI_DEVICE_ID_INTEL_82371AB_0, 0);
+	if (bdf == -1)
+		panic("Failed to find PIIX4 PCI bridge\n");
+
+	/* setup PCI interrupt routing */
+	pci_write_config_byte(bdf, PCI_CFG_PIIX4_PIRQRCA, 10);
+	pci_write_config_byte(bdf, PCI_CFG_PIIX4_PIRQRCB, 10);
+	pci_write_config_byte(bdf, PCI_CFG_PIIX4_PIRQRCC, 11);
+	pci_write_config_byte(bdf, PCI_CFG_PIIX4_PIRQRCD, 11);
 }
Signature · 1 line
1.8.4.1
#36 lore raw

PATCH 16/18 malta: setup PIIX4 interrupt route

Dear Paul Burton,

Quoted 4 lines

Without setting up the PIRQ[A:D] interrupt routes, PCI interrupts will be left disabled. Linux does not set up this routing but relies upon it having been set up by the bootloader, reading back the IRQ lines which the PIRQ[A:D] signals have been routed to.

Did you also submit a fix to Linux guys?

Best regards,
Marek Vasut

#37 lore raw

PATCH 16/18 malta: setup PIIX4 interrupt route

Quoted 7 lines · Marek Vasut

Dear Paul Burton,

Quoted 4 lines

Without setting up the PIRQ[A:D] interrupt routes, PCI interrupts will be left disabled. Linux does not set up this routing but relies upon it having been set up by the bootloader, reading back the IRQ lines which the PIRQ[A:D] signals have been routed to.

Quoted 5 lines

Did you also submit a fix to Linux guys?

Best regards,
Marek Vasut

No, I haven't. Although Linux is reliant upon these values having been programmed it's quite clear from reading its code that that depence is intentional. It explicitely reads the route setup by the bootloader rather than clobbering it, and to me that does make sense in this case.

However the pcnet32 ethernet driver seems to be reliant upon the interrupt and doesn't handle the case of a PCI device with no assigned interrupt, so I'll send a patch to fail its probe in that case shortly.

Thanks,
Paul

#38 lore raw

PATCH 16/18 malta: setup PIIX4 interrupt route

Dear Paul Burton,

Quoted 10 lines · Marek Vasut
Quoted 9 lines

Dear Paul Burton,

Quoted 4 lines

Without setting up the PIRQ[A:D] interrupt routes, PCI interrupts will be left disabled. Linux does not set up this routing but relies upon it having been set up by the bootloader, reading back the IRQ lines which the PIRQ[A:D] signals have been routed to.

Did you also submit a fix to Linux guys?

Quoted 7 lines
Quoted 2 lines

Best regards,
Marek Vasut

No, I haven't. Although Linux is reliant upon these values having been programmed it's quite clear from reading its code that that depence is intentional. It explicitely reads the route setup by the bootloader rather than clobbering it, and to me that does make sense in this case.

Can this setup not be passed via a well-designed DT binding then?

Quoted 3 lines

However the pcnet32 ethernet driver seems to be reliant upon the interrupt and doesn't handle the case of a PCI device with no assigned interrupt, so I'll send a patch to fail its probe in that case shortly.

Thanks!

Best regards,
Marek Vasut

#39 lore raw

PATCH 16/18 malta: setup PIIX4 interrupt route

Quoted 15 lines · Marek Vasut

Dear Paul Burton,

Quoted 13 lines · Marek Vasut
Quoted 11 lines

Dear Paul Burton,

Quoted 4 lines

Without setting up the PIRQ[A:D] interrupt routes, PCI interrupts will be left disabled. Linux does not set up this routing but relies upon it having been set up by the bootloader, reading back the IRQ lines which the PIRQ[A:D] signals have been routed to.

Did you also submit a fix to Linux guys?

Best regards,
Marek Vasut

Quoted 6 lines
Quoted 4 lines

No, I haven't. Although Linux is reliant upon these values having been programmed it's quite clear from reading its code that that depence is intentional. It explicitely reads the route setup by the bootloader rather than clobbering it, and to me that does make sense in this case.

Can this setup not be passed via a well-designed DT binding then?

I'm sure the information could be passed via DT but the Malta board currently doesn't use DT at all either in Linux or U-boot, and I don't have the time right now to convert them both to DT.

Quoted 10 lines
Quoted 4 lines

However the pcnet32 ethernet driver seems to be reliant upon the interrupt and doesn't handle the case of a PCI device with no assigned interrupt, so I'll send a patch to fail its probe in that case shortly.

Thanks!

Best regards,
Marek Vasut

Thanks,
Paul

#40 lore raw

PATCH 16/18 malta: setup PIIX4 interrupt route

Dear Paul Burton,

Quoted 21 lines · Marek Vasut
Quoted 20 lines

Dear Paul Burton,

Quoted 17 lines · Marek Vasut
Quoted 11 lines

Dear Paul Burton,

Quoted 4 lines

Without setting up the PIRQ[A:D] interrupt routes, PCI interrupts will be left disabled. Linux does not set up this routing but relies upon it having been set up by the bootloader, reading back the IRQ lines which the PIRQ[A:D] signals have been routed to.

Did you also submit a fix to Linux guys?

Best regards,
Marek Vasut

No, I haven't. Although Linux is reliant upon these values having been programmed it's quite clear from reading its code that that depence is intentional. It explicitely reads the route setup by the bootloader rather than clobbering it, and to me that does make sense in this case.

Quoted 5 lines
Quoted 1 line

Can this setup not be passed via a well-designed DT binding then?

I'm sure the information could be passed via DT but the Malta board currently doesn't use DT at all either in Linux or U-boot, and I don't have the time right now to convert them both to DT.

That's unfortunate, but OK.

Best regards,
Marek Vasut

#41 lore raw

PATCH v3 17/18 malta: add script & instructions to flash U-boot

In reply to Paul Burton

This patch adds a script which may be used with MIPS Navigator Console and a MIPS Nagivator Probe in order to flash U-boot to a MIPS Malta development board.

Please see the newly added doc/README.malta for usage instructions.

---
Changes in v3:
- rebase atop master

Changes in v2:
- rebased after malta moved to board/imgtec/malta
- moved board/malta/README to doc/README.malta

---
 board/imgtec/malta/flash-malta-boot.tcl | 40 +++++++++++++++++++++++++++++++++
 doc/README.malta                        | 16 +++++++++++++
 2 files changed, 56 insertions(+)
 create mode 100644 board/imgtec/malta/flash-malta-boot.tcl
 create mode 100644 doc/README.malta
board/imgtec/malta/flash-malta-boot.tcl new file +40 -0
new file mode 100644
index 0000000..0eedf07
@@ -0,0 +1,40 @@
+#
+# Copyright (C) 2013 Imagination Technologies
+#
+# Programs a MIPS Malta boot flash with a flat binary image.
+#
+# SPDX-License-Identifier:	GPL-2.0+
+#
+
+proc flash-boot { binfile } {
+  puts "flash monitor binary $binfile"
+  config Coherent on
+  config CoherencyDuringLoad on
+
+  if {[endian]=="big"} {
+    puts "CPU in BE mode"
+    flash device sharp_16x32_be;
+  } else {
+    puts "CPU in LE mode"
+    flash device sharp_16x32;
+  }
+
+  flash clear all;
+  flash set 0xBE000000..0xBE0FFFFF
+  flash erase sector 0xbe000000;
+  flash erase sector 0xbe020000;
+  flash erase sector 0xbe040000;
+  flash erase sector 0xbe060000;
+  flash erase sector 0xbe080000;
+  flash erase sector 0xbe0a0000;
+  flash erase sector 0xbe0c0000;
+  flash erase sector 0xbe0e0000;
+  puts "finished erasing boot flash";
+
+  puts "programming flash, please be patient"
+  load bin 0xbe000000 $binfile size4
+
+  flash clear all
+  config CoherencyDuringLoad off
+  puts "finished programming boot flash";
+}
doc/README.malta new file +16 -0
new file mode 100644
index 0000000..a495d02
@@ -0,0 +1,16 @@
+MIPS Malta board
+
+How to flash using a MIPS Navigator Probe:
+
+  - Ensure that your Malta has jumper JP1 fitted. Without this jumper you will
+    be unable to flash your Malta using a Navigator Probe.
+
+  - Connect Navigator Console to your probe and Malta as usual.
+
+  - Within Navigator Console run the following commands:
+
+      source /path/to/u-boot/board/malta/flash-malta-boot.tcl
+      reset
+      flash-boot /path/to/u-boot/u-boot.bin
+
+  - You should now be able to reboot your Malta to a U-boot shell.
Signature · 1 line
1.8.4.1
#42 lore raw

PATCH v3 18/18 malta: add myself to maintainers

In reply to Paul Burton

This patch adds me as a maintainer of the malta(el) board(s). I have access to physical Malta boards and the desire for U-boot to run well on them.

---
Changes in v3:
- rebase atop master

Changes in v2:
- rebased after malta moved to board/imgtec/malta

---
 boards.cfg | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
boards.cfg +2 -2
index 54c6bf7..e4d99ca 100644
@@ -483,8 +483,8 @@ Active  m68k        mcf547x_8x     -           freescale       m548xevb
 Active  microblaze  microblaze     -           xilinx          microblaze-generic  microblaze-generic                   -                                                                                                                                 Michal Simek <[email protected]>
 Active  mips        mips32         -           -               qemu-mips           qemu_mips                            qemu-mips:SYS_BIG_ENDIAN                                                                                                          Vlad Lungu <[email protected]>
 Active  mips        mips32         -           -               qemu-mips           qemu_mipsel                          qemu-mips:SYS_LITTLE_ENDIAN                                                                                                       -
-Active  mips        mips32         -           imgtec          malta               malta                                malta:MIPS32,SYS_BIG_ENDIAN                                                                                                       -
-Active  mips        mips32         -           imgtec          malta               maltael                              malta:MIPS32,SYS_LITTLE_ENDIAN                                                                                                    -
+Active  mips        mips32         -           imgtec          malta               malta                                malta:MIPS32,SYS_BIG_ENDIAN                                                                                                       Paul Burton <[email protected]>
+Active  mips        mips32         -           imgtec          malta               maltael                              malta:MIPS32,SYS_LITTLE_ENDIAN                                                                                                    Paul Burton <[email protected]>
 Active  mips        mips32         -           micronas        vct                 vct_platinum                         vct:VCT_PLATINUM                                                                                                                  -
 Active  mips        mips32         -           micronas        vct                 vct_platinum_onenand                 vct:VCT_PLATINUM,VCT_ONENAND                                                                                                      -
 Active  mips        mips32         -           micronas        vct                 vct_platinum_onenand_small           vct:VCT_PLATINUM,VCT_ONENAND,VCT_SMALL_IMAGE                                                                                      -
Signature · 1 line
1.8.4.1
#43 lore raw

cover MIPS Malta board support

In reply to Paul Burton

2013/11/8 Paul Burton <[email protected]>:

Quoted 29 lines

This series adds support for booting on a physical MIPS Malta board using a coreFPGA6 core card.

The first 6 patches lay some groundwork, then the next 8 genericise the existing qemu-malta board to also function on a physical Malta.

In the final patch I stake my claim upon, errm I mean step up to support :), the malta(el) board(s).

Since the original posting this series has been rebased atop current master & adapted to the KBuild changes. I also included the RTC support patch I previously posted separately, and a couple of extra patches which fix up issues with booted kernels & add environment support.

After this series Gabor Juhos' "malta: define CONFIG_MEMSIZE_IN_BYTES" patch is also required in order for a booted Linux kernel to function correctly.

Paul Burton (18):
mips32: detect L1 cache sizes if they're not defined
pcnet: code style cleanup
pcnet: s/le16_to_cpu/cpu_to_le16/ in pcnet_send
pcnet: add cache flushing & invalidation
pcnet: enable the NOUFLO feature
pci.h: allow inclusion in assembly source
qemu-malta: rename to just "malta"
malta: setup super I/O UARTs
malta: support for coreFPGA6 boards
malta: display "U-boot" on the LCD screen

Quoted 9 lines

malta: enable CONFIG_PCNET_79C973, PCNET_HAS_PROM, CONFIG_CMD_DHCP
malta: remove cache size definitions
malta: disable L2 caches
malta: enable RTC support
malta: store environment in flash
malta: setup PIIX4 interrupt route
malta: add script & instructions to flash U-boot
malta: add myself to maintainers

all patches applied to u-boot-mips/master, thanks

Signature · 1 line
- Daniel