* Re: cpu_clock confusion
From: David Miller @ 2008-04-24 9:24 UTC (permalink / raw)
To: johannes; +Cc: linux-kernel, benh, linuxppc-dev, arndb, mingo, tglx
In-Reply-To: <1209028912.3357.52.camel@johannes.berg>
From: Johannes Berg <johannes@sipsolutions.net>
Date: Thu, 24 Apr 2008 11:21:52 +0200
> Hmm. Why is that whole cpu_clock stuff in place anyway? powerpc has
> perfectly synchronised time across processors with dirt cheap access to
> it as well, so why build all this code that only messes it up on top of
> it?
Same on sparc64. These changes add more bugs than they fix.
^ permalink raw reply
* Re: cpu_clock confusion (was: printk time confusion?)
From: Johannes Berg @ 2008-04-24 9:21 UTC (permalink / raw)
To: Ingo Molnar
Cc: linuxppc-dev list, Thomas Gleixner, Arnd Bergmann,
Linux Kernel list, Benjamin Herrenschmidt
In-Reply-To: <20080404144641.GA15844@elte.hu>
[-- Attachment #1: Type: text/plain, Size: 714 bytes --]
> > [ 2.764009 (3/3)]
> > [ 4.272241 (2/2)]
> > [ 4.272322 (2/2)]
> > [ 4.272375 (2/2)]
> > [ 2.948002 (3/3)]
> >
> > As you can see, I added printk_cpu and smp_processor_id() to the
> > printk timestamp output and thus it is obvious that the different
> > times come from different CPUs.
>
> the fixes are queued for v2.6.26. You can pick them up from
> sched-devel/latest as well:
>
> http://people.redhat.com/mingo/sched-devel.git/README
Hmm. Why is that whole cpu_clock stuff in place anyway? powerpc has
perfectly synchronised time across processors with dirt cheap access to
it as well, so why build all this code that only messes it up on top of
it?
johannes
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 828 bytes --]
^ permalink raw reply
* [PATCH 1/9] cell: generalize io-workarounds code
From: Ishizaki Kou @ 2008-04-24 9:21 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev
This patch splits cell io-workaround code into spider-pci dependent
code and a generic part, and also moves io-workarounds initialization
into cell_setup_phb. This patch includes BenH's fix.
Signed-off-by: Kou Ishizaki <kou.ishizaki@toshiba.co.jp>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
arch/powerpc/kernel/of_platform.c | 2
arch/powerpc/platforms/cell/Makefile | 3
arch/powerpc/platforms/cell/io-workarounds.c | 360 +++++++--------------------
arch/powerpc/platforms/cell/io-workarounds.h | 49 +++
arch/powerpc/platforms/cell/setup.c | 43 +++
arch/powerpc/platforms/cell/spider-pci.c | 184 +++++++++++++
include/asm-powerpc/io-defs.h | 107 ++++----
include/asm-powerpc/io.h | 8
8 files changed, 437 insertions(+), 319 deletions(-)
Index: b/arch/powerpc/platforms/cell/Makefile
===================================================================
--- a/arch/powerpc/platforms/cell/Makefile 2008-04-24 12:23:23.000000000 +0900
+++ b/arch/powerpc/platforms/cell/Makefile 2008-04-24 12:23:25.000000000 +0900
@@ -1,6 +1,7 @@
obj-$(CONFIG_PPC_CELL_NATIVE) += interrupt.o iommu.o setup.o \
cbe_regs.o spider-pic.o \
- pervasive.o pmu.o io-workarounds.o
+ pervasive.o pmu.o io-workarounds.o \
+ spider-pci.o
obj-$(CONFIG_CBE_RAS) += ras.o
obj-$(CONFIG_CBE_THERM) += cbe_thermal.o
Index: b/arch/powerpc/platforms/cell/io-workarounds.c
===================================================================
--- a/arch/powerpc/platforms/cell/io-workarounds.c 2008-04-24 12:23:23.000000000 +0900
+++ b/arch/powerpc/platforms/cell/io-workarounds.c 2008-04-24 12:23:25.000000000 +0900
@@ -1,6 +1,9 @@
/*
+ * Support PCI IO workaround
+ *
* Copyright (C) 2006 Benjamin Herrenschmidt <benh@kernel.crashing.org>
* IBM, Corp.
+ * (C) Copyright 2007-2008 TOSHIBA CORPORATION
*
* 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
@@ -9,335 +12,174 @@
#undef DEBUG
#include <linux/kernel.h>
-#include <linux/mm.h>
-#include <linux/pci.h>
+
#include <asm/io.h>
#include <asm/machdep.h>
-#include <asm/pci-bridge.h>
+#include <asm/pgtable.h>
#include <asm/ppc-pci.h>
+#include "io-workarounds.h"
-#define SPIDER_PCI_REG_BASE 0xd000
-#define SPIDER_PCI_VCI_CNTL_STAT 0x0110
-#define SPIDER_PCI_DUMMY_READ 0x0810
-#define SPIDER_PCI_DUMMY_READ_BASE 0x0814
-
-/* Undefine that to re-enable bogus prefetch
- *
- * Without that workaround, the chip will do bogus prefetch past
- * page boundary from system memory. This setting will disable that,
- * though the documentation is unclear as to the consequences of doing
- * so, either purely performances, or possible misbehaviour... It's not
- * clear wether the chip can handle unaligned accesses at all without
- * prefetching enabled.
- *
- * For now, things appear to be behaving properly with that prefetching
- * disabled and IDE, possibly because IDE isn't doing any unaligned
- * access.
- */
-#define SPIDER_DISABLE_PREFETCH
-
-#define MAX_SPIDERS 3
+#define IOWA_MAX_BUS 8
-static struct spider_pci_bus {
- void __iomem *regs;
- unsigned long mmio_start;
- unsigned long mmio_end;
- unsigned long pio_vstart;
- unsigned long pio_vend;
-} spider_pci_busses[MAX_SPIDERS];
-static int spider_pci_count;
+static struct iowa_bus iowa_busses[IOWA_MAX_BUS];
+static unsigned int iowa_bus_count;
-static struct spider_pci_bus *spider_pci_find(unsigned long vaddr,
- unsigned long paddr)
+static struct iowa_bus *iowa_pci_find(unsigned long vaddr, unsigned long paddr)
{
- int i;
-
- for (i = 0; i < spider_pci_count; i++) {
- struct spider_pci_bus *bus = &spider_pci_busses[i];
- if (paddr && paddr >= bus->mmio_start && paddr < bus->mmio_end)
- return bus;
- if (vaddr && vaddr >= bus->pio_vstart && vaddr < bus->pio_vend)
- return bus;
+ int i, j;
+ struct resource *res;
+ unsigned long vstart, vend;
+
+ for (i = 0; i < iowa_bus_count; i++) {
+ struct iowa_bus *bus = &iowa_busses[i];
+ struct pci_controller *phb = bus->phb;
+
+ if (vaddr) {
+ vstart = (unsigned long)phb->io_base_virt;
+ vend = vstart + phb->pci_io_size - 1;
+ if ((vaddr >= vstart) && (vaddr <= vend))
+ return bus;
+ }
+
+ if (paddr)
+ for (j = 0; j < 3; j++) {
+ res = &phb->mem_resources[j];
+ if (paddr >= res->start && paddr <= res->end)
+ return bus;
+ }
}
+
return NULL;
}
-static void spider_io_flush(const volatile void __iomem *addr)
+struct iowa_bus *iowa_mem_find_bus(const PCI_IO_ADDR addr)
{
- struct spider_pci_bus *bus;
+ struct iowa_bus *bus;
int token;
- /* Get platform token (set by ioremap) from address */
token = PCI_GET_ADDR_TOKEN(addr);
- /* Fast path if we have a non-0 token, it indicates which bus we
- * are on.
- *
- * If the token is 0, that means either that the ioremap was done
- * before we initialized this layer, or it's a PIO operation. We
- * fallback to a low path in this case. Hopefully, internal devices
- * which are ioremap'ed early should use in_XX/out_XX functions
- * instead of the PCI ones and thus not suffer from the slowdown.
- *
- * Also note that currently, the workaround will not work for areas
- * that are not mapped with PTEs (bolted in the hash table). This
- * is the case for ioremaps done very early at boot (before
- * mem_init_done) and includes the mapping of the ISA IO space.
- *
- * Fortunately, none of the affected devices is expected to do DMA
- * and thus there should be no problem in practice.
- *
- * In order to improve performances, we only do the PTE search for
- * addresses falling in the PHB IO space area. That means it will
- * not work for hotplug'ed PHBs but those don't exist with Spider.
- */
- if (token && token <= spider_pci_count)
- bus = &spider_pci_busses[token - 1];
+ if (token && token <= iowa_bus_count)
+ bus = &iowa_busses[token - 1];
else {
unsigned long vaddr, paddr;
pte_t *ptep;
- /* Fixup physical address */
vaddr = (unsigned long)PCI_FIX_ADDR(addr);
+ if (vaddr < PHB_IO_BASE || vaddr >= PHB_IO_END)
+ return NULL;
- /* Check if it's in allowed range for PIO */
- if (vaddr < PHB_IO_BASE || vaddr > PHB_IO_END)
- return;
-
- /* Try to find a PTE. If not, clear the paddr, we'll do
- * a vaddr only lookup (PIO only)
- */
ptep = find_linux_pte(init_mm.pgd, vaddr);
if (ptep == NULL)
paddr = 0;
else
paddr = pte_pfn(*ptep) << PAGE_SHIFT;
+ bus = iowa_pci_find(vaddr, paddr);
- bus = spider_pci_find(vaddr, paddr);
if (bus == NULL)
- return;
+ return NULL;
}
- /* Now do the workaround
- */
- (void)in_be32(bus->regs + SPIDER_PCI_DUMMY_READ);
+ return bus;
}
-static u8 spider_readb(const volatile void __iomem *addr)
+struct iowa_bus *iowa_pio_find_bus(unsigned long port)
{
- u8 val = __do_readb(addr);
- spider_io_flush(addr);
- return val;
+ unsigned long vaddr = (unsigned long)pci_io_base + port;
+ return iowa_pci_find(vaddr, 0);
}
-static u16 spider_readw(const volatile void __iomem *addr)
-{
- u16 val = __do_readw(addr);
- spider_io_flush(addr);
- return val;
-}
-static u32 spider_readl(const volatile void __iomem *addr)
-{
- u32 val = __do_readl(addr);
- spider_io_flush(addr);
- return val;
+#define DEF_PCI_AC_RET(name, ret, at, al, space, aa) \
+static ret iowa_##name at \
+{ \
+ struct iowa_bus *bus; \
+ bus = iowa_##space##_find_bus(aa); \
+ if (bus && bus->ops && bus->ops->name) \
+ return bus->ops->name al; \
+ return __do_##name al; \
}
-static u64 spider_readq(const volatile void __iomem *addr)
-{
- u64 val = __do_readq(addr);
- spider_io_flush(addr);
- return val;
+#define DEF_PCI_AC_NORET(name, at, al, space, aa) \
+static void iowa_##name at \
+{ \
+ struct iowa_bus *bus; \
+ bus = iowa_##space##_find_bus(aa); \
+ if (bus && bus->ops && bus->ops->name) { \
+ bus->ops->name al; \
+ return; \
+ } \
+ __do_##name al; \
}
-static u16 spider_readw_be(const volatile void __iomem *addr)
-{
- u16 val = __do_readw_be(addr);
- spider_io_flush(addr);
- return val;
-}
+#include <asm/io-defs.h>
-static u32 spider_readl_be(const volatile void __iomem *addr)
-{
- u32 val = __do_readl_be(addr);
- spider_io_flush(addr);
- return val;
-}
+#undef DEF_PCI_AC_RET
+#undef DEF_PCI_AC_NORET
-static u64 spider_readq_be(const volatile void __iomem *addr)
-{
- u64 val = __do_readq_be(addr);
- spider_io_flush(addr);
- return val;
-}
+static struct ppc_pci_io __initdata iowa_pci_io = {
-static void spider_readsb(const volatile void __iomem *addr, void *buf,
- unsigned long count)
-{
- __do_readsb(addr, buf, count);
- spider_io_flush(addr);
-}
+#define DEF_PCI_AC_RET(name, ret, at, al, space, aa) .name = iowa_##name,
+#define DEF_PCI_AC_NORET(name, at, al, space, aa) .name = iowa_##name,
-static void spider_readsw(const volatile void __iomem *addr, void *buf,
- unsigned long count)
-{
- __do_readsw(addr, buf, count);
- spider_io_flush(addr);
-}
+#include <asm/io-defs.h>
-static void spider_readsl(const volatile void __iomem *addr, void *buf,
- unsigned long count)
-{
- __do_readsl(addr, buf, count);
- spider_io_flush(addr);
-}
-
-static void spider_memcpy_fromio(void *dest, const volatile void __iomem *src,
- unsigned long n)
-{
- __do_memcpy_fromio(dest, src, n);
- spider_io_flush(src);
-}
+#undef DEF_PCI_AC_RET
+#undef DEF_PCI_AC_NORET
+};
-static void __iomem * spider_ioremap(unsigned long addr, unsigned long size,
- unsigned long flags)
+static void __iomem *iowa_ioremap(unsigned long addr, unsigned long size,
+ unsigned long flags)
{
- struct spider_pci_bus *bus;
+ struct iowa_bus *bus;
void __iomem *res = __ioremap(addr, size, flags);
int busno;
- pr_debug("spider_ioremap(0x%lx, 0x%lx, 0x%lx) -> 0x%p\n",
- addr, size, flags, res);
-
- bus = spider_pci_find(0, addr);
+ bus = iowa_pci_find(0, addr);
if (bus != NULL) {
- busno = bus - spider_pci_busses;
- pr_debug(" found bus %d, setting token\n", busno);
+ busno = bus - iowa_busses;
PCI_SET_ADDR_TOKEN(res, busno + 1);
}
- pr_debug(" result=0x%p\n", res);
-
return res;
}
-static void __init spider_pci_setup_chip(struct spider_pci_bus *bus)
-{
-#ifdef SPIDER_DISABLE_PREFETCH
- u32 val = in_be32(bus->regs + SPIDER_PCI_VCI_CNTL_STAT);
- pr_debug(" PVCI_Control_Status was 0x%08x\n", val);
- out_be32(bus->regs + SPIDER_PCI_VCI_CNTL_STAT, val | 0x8);
-#endif
-
- /* Configure the dummy address for the workaround */
- out_be32(bus->regs + SPIDER_PCI_DUMMY_READ_BASE, 0x80000000);
-}
-
-static void __init spider_pci_add_one(struct pci_controller *phb)
+/* Regist new bus to support workaround */
+void __init iowa_register_bus(struct pci_controller *phb,
+ struct ppc_pci_io *ops,
+ int (*initfunc)(struct iowa_bus *, void *), void *data)
{
- struct spider_pci_bus *bus = &spider_pci_busses[spider_pci_count];
+ struct iowa_bus *bus;
struct device_node *np = phb->dn;
- struct resource rsrc;
- void __iomem *regs;
- if (spider_pci_count >= MAX_SPIDERS) {
- printk(KERN_ERR "Too many spider bridges, workarounds"
- " disabled for %s\n", np->full_name);
+ if (iowa_bus_count >= IOWA_MAX_BUS) {
+ pr_err("IOWA:Too many pci bridges, "
+ "workarounds disabled for %s\n", np->full_name);
return;
}
- /* Get the registers for the beast */
- if (of_address_to_resource(np, 0, &rsrc)) {
- printk(KERN_ERR "Failed to get registers for spider %s"
- " workarounds disabled\n", np->full_name);
- return;
- }
+ bus = &iowa_busses[iowa_bus_count];
+ bus->phb = phb;
+ bus->ops = ops;
- /* Mask out some useless bits in there to get to the base of the
- * spider chip
- */
- rsrc.start &= ~0xfffffffful;
-
- /* Map them */
- regs = ioremap(rsrc.start + SPIDER_PCI_REG_BASE, 0x1000);
- if (regs == NULL) {
- printk(KERN_ERR "Failed to map registers for spider %s"
- " workarounds disabled\n", np->full_name);
- return;
- }
-
- spider_pci_count++;
-
- /* We assume spiders only have one MMIO resource */
- bus->mmio_start = phb->mem_resources[0].start;
- bus->mmio_end = phb->mem_resources[0].end + 1;
-
- bus->pio_vstart = (unsigned long)phb->io_base_virt;
- bus->pio_vend = bus->pio_vstart + phb->pci_io_size;
-
- bus->regs = regs;
-
- printk(KERN_INFO "PCI: Spider MMIO workaround for %s\n",np->full_name);
+ if (initfunc)
+ if ((*initfunc)(bus, data))
+ return;
- pr_debug(" mmio (P) = 0x%016lx..0x%016lx\n",
- bus->mmio_start, bus->mmio_end);
- pr_debug(" pio (V) = 0x%016lx..0x%016lx\n",
- bus->pio_vstart, bus->pio_vend);
- pr_debug(" regs (P) = 0x%016lx (V) = 0x%p\n",
- rsrc.start + SPIDER_PCI_REG_BASE, bus->regs);
+ iowa_bus_count++;
- spider_pci_setup_chip(bus);
+ pr_debug("IOWA:[%d]Add bus, %s.\n", iowa_bus_count-1, np->full_name);
}
-static struct ppc_pci_io __initdata spider_pci_io = {
- .readb = spider_readb,
- .readw = spider_readw,
- .readl = spider_readl,
- .readq = spider_readq,
- .readw_be = spider_readw_be,
- .readl_be = spider_readl_be,
- .readq_be = spider_readq_be,
- .readsb = spider_readsb,
- .readsw = spider_readsw,
- .readsl = spider_readsl,
- .memcpy_fromio = spider_memcpy_fromio,
-};
-
-static int __init spider_pci_workaround_init(void)
+/* enable IO workaround */
+void __init io_workaround_init(void)
{
- struct pci_controller *phb;
-
- /* Find spider bridges. We assume they have been all probed
- * in setup_arch(). If that was to change, we would need to
- * update this code to cope with dynamically added busses
- */
- list_for_each_entry(phb, &hose_list, list_node) {
- struct device_node *np = phb->dn;
- const char *model = of_get_property(np, "model", NULL);
-
- /* If no model property or name isn't exactly "pci", skip */
- if (model == NULL || strcmp(np->name, "pci"))
- continue;
- /* If model is not "Spider", skip */
- if (strcmp(model, "Spider"))
- continue;
- spider_pci_add_one(phb);
- }
+ static int io_workaround_inited;
- /* No Spider PCI found, exit */
- if (spider_pci_count == 0)
- return 0;
-
- /* Setup IO callbacks. We only setup MMIO reads. PIO reads will
- * fallback to MMIO reads (though without a token, thus slower)
- */
- ppc_pci_io = spider_pci_io;
-
- /* Setup ioremap callback */
- ppc_md.ioremap = spider_ioremap;
-
- return 0;
+ if (io_workaround_inited)
+ return;
+ ppc_pci_io = iowa_pci_io;
+ ppc_md.ioremap = iowa_ioremap;
+ io_workaround_inited = 1;
}
-machine_arch_initcall(cell, spider_pci_workaround_init);
Index: b/arch/powerpc/platforms/cell/io-workarounds.h
===================================================================
--- /dev/null 1970-01-01 00:00:00.000000000 +0000
+++ b/arch/powerpc/platforms/cell/io-workarounds.h 2008-04-24 12:23:25.000000000 +0900
@@ -0,0 +1,49 @@
+/*
+ * Support PCI IO workaround
+ *
+ * (C) Copyright 2007-2008 TOSHIBA CORPORATION
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#ifndef _IO_WORKAROUNDS_H
+#define _IO_WORKAROUNDS_H
+
+#include <linux/io.h>
+#include <asm/pci-bridge.h>
+
+/* Bus info */
+struct iowa_bus {
+ struct pci_controller *phb;
+ struct ppc_pci_io *ops;
+ void *private;
+};
+
+void __init io_workaround_init(void);
+void __init iowa_register_bus(struct pci_controller *, struct ppc_pci_io *,
+ int (*)(struct iowa_bus *, void *), void *);
+struct iowa_bus *iowa_mem_find_bus(const PCI_IO_ADDR);
+struct iowa_bus *iowa_pio_find_bus(unsigned long);
+
+extern struct ppc_pci_io spiderpci_ops;
+extern int spiderpci_iowa_init(struct iowa_bus *, void *);
+
+#define SPIDER_PCI_REG_BASE 0xd000
+#define SPIDER_PCI_REG_SIZE 0x1000
+#define SPIDER_PCI_VCI_CNTL_STAT 0x0110
+#define SPIDER_PCI_DUMMY_READ 0x0810
+#define SPIDER_PCI_DUMMY_READ_BASE 0x0814
+
+#endif /* _IO_WORKAROUNDS_H */
Index: b/arch/powerpc/platforms/cell/spider-pci.c
===================================================================
--- /dev/null 1970-01-01 00:00:00.000000000 +0000
+++ b/arch/powerpc/platforms/cell/spider-pci.c 2008-04-24 12:23:25.000000000 +0900
@@ -0,0 +1,184 @@
+/*
+ * IO workarounds for PCI on Celleb/Cell platform
+ *
+ * (C) Copyright 2006-2007 TOSHIBA CORPORATION
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#undef DEBUG
+
+#include <linux/kernel.h>
+#include <linux/of_platform.h>
+#include <linux/io.h>
+
+#include <asm/ppc-pci.h>
+#include <asm/pci-bridge.h>
+
+#include "io-workarounds.h"
+
+#define SPIDER_PCI_DISABLE_PREFETCH
+
+struct spiderpci_iowa_private {
+ void __iomem *regs;
+};
+
+static void spiderpci_io_flush(struct iowa_bus *bus)
+{
+ struct spiderpci_iowa_private *priv;
+ u32 val;
+
+ priv = bus->private;
+ val = in_be32(priv->regs + SPIDER_PCI_DUMMY_READ);
+ iosync();
+}
+
+#define SPIDER_PCI_MMIO_READ(name, ret) \
+static ret spiderpci_##name(const PCI_IO_ADDR addr) \
+{ \
+ ret val = __do_##name(addr); \
+ spiderpci_io_flush(iowa_mem_find_bus(addr)); \
+ return val; \
+}
+
+#define SPIDER_PCI_MMIO_READ_STR(name) \
+static void spiderpci_##name(const PCI_IO_ADDR addr, void *buf, \
+ unsigned long count) \
+{ \
+ __do_##name(addr, buf, count); \
+ spiderpci_io_flush(iowa_mem_find_bus(addr)); \
+}
+
+SPIDER_PCI_MMIO_READ(readb, u8)
+SPIDER_PCI_MMIO_READ(readw, u16)
+SPIDER_PCI_MMIO_READ(readl, u32)
+SPIDER_PCI_MMIO_READ(readq, u64)
+SPIDER_PCI_MMIO_READ(readw_be, u16)
+SPIDER_PCI_MMIO_READ(readl_be, u32)
+SPIDER_PCI_MMIO_READ(readq_be, u64)
+SPIDER_PCI_MMIO_READ_STR(readsb)
+SPIDER_PCI_MMIO_READ_STR(readsw)
+SPIDER_PCI_MMIO_READ_STR(readsl)
+
+static void spiderpci_memcpy_fromio(void *dest, const PCI_IO_ADDR src,
+ unsigned long n)
+{
+ __do_memcpy_fromio(dest, src, n);
+ spiderpci_io_flush(iowa_mem_find_bus(src));
+}
+
+static int __init spiderpci_pci_setup_chip(struct pci_controller *phb,
+ void __iomem *regs)
+{
+ void *dummy_page_va;
+ dma_addr_t dummy_page_da;
+
+#ifdef SPIDER_PCI_DISABLE_PREFETCH
+ u32 val = in_be32(regs + SPIDER_PCI_VCI_CNTL_STAT);
+ pr_debug("SPIDER_IOWA:PVCI_Control_Status was 0x%08x\n", val);
+ out_be32(regs + SPIDER_PCI_VCI_CNTL_STAT, val | 0x8);
+#endif /* SPIDER_PCI_DISABLE_PREFETCH */
+
+ /* setup dummy read */
+ /*
+ * On CellBlade, we can't know that which XDR memory is used by
+ * kmalloc() to allocate dummy_page_va.
+ * In order to imporve the performance, the XDR which is used to
+ * allocate dummy_page_va is the nearest the spider-pci.
+ * We have to select the CBE which is the nearest the spider-pci
+ * to allocate memory from the best XDR, but I don't know that
+ * how to do.
+ *
+ * Celleb does not have this problem, because it has only one XDR.
+ */
+ dummy_page_va = kmalloc(PAGE_SIZE, GFP_KERNEL);
+ if (!dummy_page_va) {
+ pr_err("SPIDERPCI-IOWA:Alloc dummy_page_va failed.\n");
+ return -1;
+ }
+
+ dummy_page_da = dma_map_single(phb->parent, dummy_page_va,
+ PAGE_SIZE, DMA_FROM_DEVICE);
+ if (dma_mapping_error(dummy_page_da)) {
+ pr_err("SPIDER-IOWA:Map dummy page filed.\n");
+ kfree(dummy_page_va);
+ return -1;
+ }
+
+ out_be32(regs + SPIDER_PCI_DUMMY_READ_BASE, dummy_page_da);
+
+ return 0;
+}
+
+int __init spiderpci_iowa_init(struct iowa_bus *bus, void *data)
+{
+ void __iomem *regs = NULL;
+ struct spiderpci_iowa_private *priv;
+ struct device_node *np = bus->phb->dn;
+ struct resource r;
+ unsigned long offset = (unsigned long)data;
+
+ pr_debug("SPIDERPCI-IOWA:Bus initialize for spider(%s)\n",
+ np->full_name);
+
+ priv = kzalloc(sizeof(struct spiderpci_iowa_private), GFP_KERNEL);
+ if (!priv) {
+ pr_err("SPIDERPCI-IOWA:"
+ "Can't allocate struct spiderpci_iowa_private");
+ return -1;
+ }
+
+ if (of_address_to_resource(np, 0, &r)) {
+ pr_err("SPIDERPCI-IOWA:Can't get resource.\n");
+ goto error;
+ }
+
+ regs = ioremap(r.start + offset, SPIDER_PCI_REG_SIZE);
+ if (!regs) {
+ pr_err("SPIDERPCI-IOWA:ioremap failed.\n");
+ goto error;
+ }
+ priv->regs = regs;
+ bus->private = priv;
+
+ if (spiderpci_pci_setup_chip(bus->phb, regs))
+ goto error;
+
+ return 0;
+
+error:
+ kfree(priv);
+ bus->private = NULL;
+
+ if (regs)
+ iounmap(regs);
+
+ return -1;
+}
+
+struct ppc_pci_io spiderpci_ops = {
+ .readb = spiderpci_readb,
+ .readw = spiderpci_readw,
+ .readl = spiderpci_readl,
+ .readq = spiderpci_readq,
+ .readw_be = spiderpci_readw_be,
+ .readl_be = spiderpci_readl_be,
+ .readq_be = spiderpci_readq_be,
+ .readsb = spiderpci_readsb,
+ .readsw = spiderpci_readsw,
+ .readsl = spiderpci_readsl,
+ .memcpy_fromio = spiderpci_memcpy_fromio,
+};
+
Index: b/include/asm-powerpc/io-defs.h
===================================================================
--- a/include/asm-powerpc/io-defs.h 2008-04-24 12:23:23.000000000 +0900
+++ b/include/asm-powerpc/io-defs.h 2008-04-24 12:23:25.000000000 +0900
@@ -1,59 +1,60 @@
/* This file is meant to be include multiple times by other headers */
+/* last 2 argments are used by platforms/cell/io-workarounds.[ch] */
-DEF_PCI_AC_RET(readb, u8, (const PCI_IO_ADDR addr), (addr))
-DEF_PCI_AC_RET(readw, u16, (const PCI_IO_ADDR addr), (addr))
-DEF_PCI_AC_RET(readl, u32, (const PCI_IO_ADDR addr), (addr))
-DEF_PCI_AC_RET(readw_be, u16, (const PCI_IO_ADDR addr), (addr))
-DEF_PCI_AC_RET(readl_be, u32, (const PCI_IO_ADDR addr), (addr))
-DEF_PCI_AC_NORET(writeb, (u8 val, PCI_IO_ADDR addr), (val, addr))
-DEF_PCI_AC_NORET(writew, (u16 val, PCI_IO_ADDR addr), (val, addr))
-DEF_PCI_AC_NORET(writel, (u32 val, PCI_IO_ADDR addr), (val, addr))
-DEF_PCI_AC_NORET(writew_be, (u16 val, PCI_IO_ADDR addr), (val, addr))
-DEF_PCI_AC_NORET(writel_be, (u32 val, PCI_IO_ADDR addr), (val, addr))
+DEF_PCI_AC_RET(readb, u8, (const PCI_IO_ADDR addr), (addr), mem, addr)
+DEF_PCI_AC_RET(readw, u16, (const PCI_IO_ADDR addr), (addr), mem, addr)
+DEF_PCI_AC_RET(readl, u32, (const PCI_IO_ADDR addr), (addr), mem, addr)
+DEF_PCI_AC_RET(readw_be, u16, (const PCI_IO_ADDR addr), (addr), mem, addr)
+DEF_PCI_AC_RET(readl_be, u32, (const PCI_IO_ADDR addr), (addr), mem, addr)
+DEF_PCI_AC_NORET(writeb, (u8 val, PCI_IO_ADDR addr), (val, addr), mem, addr)
+DEF_PCI_AC_NORET(writew, (u16 val, PCI_IO_ADDR addr), (val, addr), mem, addr)
+DEF_PCI_AC_NORET(writel, (u32 val, PCI_IO_ADDR addr), (val, addr), mem, addr)
+DEF_PCI_AC_NORET(writew_be, (u16 val, PCI_IO_ADDR addr), (val, addr), mem, addr)
+DEF_PCI_AC_NORET(writel_be, (u32 val, PCI_IO_ADDR addr), (val, addr), mem, addr)
#ifdef __powerpc64__
-DEF_PCI_AC_RET(readq, u64, (const PCI_IO_ADDR addr), (addr))
-DEF_PCI_AC_RET(readq_be, u64, (const PCI_IO_ADDR addr), (addr))
-DEF_PCI_AC_NORET(writeq, (u64 val, PCI_IO_ADDR addr), (val, addr))
-DEF_PCI_AC_NORET(writeq_be, (u64 val, PCI_IO_ADDR addr), (val, addr))
+DEF_PCI_AC_RET(readq, u64, (const PCI_IO_ADDR addr), (addr), mem, addr)
+DEF_PCI_AC_RET(readq_be, u64, (const PCI_IO_ADDR addr), (addr), mem, addr)
+DEF_PCI_AC_NORET(writeq, (u64 val, PCI_IO_ADDR addr), (val, addr), mem, addr)
+DEF_PCI_AC_NORET(writeq_be, (u64 val, PCI_IO_ADDR addr), (val, addr), mem, addr)
#endif /* __powerpc64__ */
-DEF_PCI_AC_RET(inb, u8, (unsigned long port), (port))
-DEF_PCI_AC_RET(inw, u16, (unsigned long port), (port))
-DEF_PCI_AC_RET(inl, u32, (unsigned long port), (port))
-DEF_PCI_AC_NORET(outb, (u8 val, unsigned long port), (val, port))
-DEF_PCI_AC_NORET(outw, (u16 val, unsigned long port), (val, port))
-DEF_PCI_AC_NORET(outl, (u32 val, unsigned long port), (val, port))
-
-DEF_PCI_AC_NORET(readsb, (const PCI_IO_ADDR a, void *b, unsigned long c), \
- (a, b, c))
-DEF_PCI_AC_NORET(readsw, (const PCI_IO_ADDR a, void *b, unsigned long c), \
- (a, b, c))
-DEF_PCI_AC_NORET(readsl, (const PCI_IO_ADDR a, void *b, unsigned long c), \
- (a, b, c))
-DEF_PCI_AC_NORET(writesb, (PCI_IO_ADDR a, const void *b, unsigned long c), \
- (a, b, c))
-DEF_PCI_AC_NORET(writesw, (PCI_IO_ADDR a, const void *b, unsigned long c), \
- (a, b, c))
-DEF_PCI_AC_NORET(writesl, (PCI_IO_ADDR a, const void *b, unsigned long c), \
- (a, b, c))
-
-DEF_PCI_AC_NORET(insb, (unsigned long p, void *b, unsigned long c), \
- (p, b, c))
-DEF_PCI_AC_NORET(insw, (unsigned long p, void *b, unsigned long c), \
- (p, b, c))
-DEF_PCI_AC_NORET(insl, (unsigned long p, void *b, unsigned long c), \
- (p, b, c))
-DEF_PCI_AC_NORET(outsb, (unsigned long p, const void *b, unsigned long c), \
- (p, b, c))
-DEF_PCI_AC_NORET(outsw, (unsigned long p, const void *b, unsigned long c), \
- (p, b, c))
-DEF_PCI_AC_NORET(outsl, (unsigned long p, const void *b, unsigned long c), \
- (p, b, c))
-
-DEF_PCI_AC_NORET(memset_io, (PCI_IO_ADDR a, int c, unsigned long n), \
- (a, c, n))
-DEF_PCI_AC_NORET(memcpy_fromio,(void *d,const PCI_IO_ADDR s,unsigned long n), \
- (d, s, n))
-DEF_PCI_AC_NORET(memcpy_toio,(PCI_IO_ADDR d,const void *s,unsigned long n), \
- (d, s, n))
+DEF_PCI_AC_RET(inb, u8, (unsigned long port), (port), pio, port)
+DEF_PCI_AC_RET(inw, u16, (unsigned long port), (port), pio, port)
+DEF_PCI_AC_RET(inl, u32, (unsigned long port), (port), pio, port)
+DEF_PCI_AC_NORET(outb, (u8 val, unsigned long port), (val, port), pio, port)
+DEF_PCI_AC_NORET(outw, (u16 val, unsigned long port), (val, port), pio, port)
+DEF_PCI_AC_NORET(outl, (u32 val, unsigned long port), (val, port), pio, port)
+
+DEF_PCI_AC_NORET(readsb, (const PCI_IO_ADDR a, void *b, unsigned long c),
+ (a, b, c), mem, a)
+DEF_PCI_AC_NORET(readsw, (const PCI_IO_ADDR a, void *b, unsigned long c),
+ (a, b, c), mem, a)
+DEF_PCI_AC_NORET(readsl, (const PCI_IO_ADDR a, void *b, unsigned long c),
+ (a, b, c), mem, a)
+DEF_PCI_AC_NORET(writesb, (PCI_IO_ADDR a, const void *b, unsigned long c),
+ (a, b, c), mem, a)
+DEF_PCI_AC_NORET(writesw, (PCI_IO_ADDR a, const void *b, unsigned long c),
+ (a, b, c), mem, a)
+DEF_PCI_AC_NORET(writesl, (PCI_IO_ADDR a, const void *b, unsigned long c),
+ (a, b, c), mem, a)
+
+DEF_PCI_AC_NORET(insb, (unsigned long p, void *b, unsigned long c),
+ (p, b, c), pio, p)
+DEF_PCI_AC_NORET(insw, (unsigned long p, void *b, unsigned long c),
+ (p, b, c), pio, p)
+DEF_PCI_AC_NORET(insl, (unsigned long p, void *b, unsigned long c),
+ (p, b, c), pio, p)
+DEF_PCI_AC_NORET(outsb, (unsigned long p, const void *b, unsigned long c),
+ (p, b, c), pio, p)
+DEF_PCI_AC_NORET(outsw, (unsigned long p, const void *b, unsigned long c),
+ (p, b, c), pio, p)
+DEF_PCI_AC_NORET(outsl, (unsigned long p, const void *b, unsigned long c),
+ (p, b, c), pio, p)
+
+DEF_PCI_AC_NORET(memset_io, (PCI_IO_ADDR a, int c, unsigned long n),
+ (a, c, n), mem, a)
+DEF_PCI_AC_NORET(memcpy_fromio, (void *d, const PCI_IO_ADDR s, unsigned long n),
+ (d, s, n), mem, s)
+DEF_PCI_AC_NORET(memcpy_toio, (PCI_IO_ADDR d, const void *s, unsigned long n),
+ (d, s, n), mem, d)
Index: b/include/asm-powerpc/io.h
===================================================================
--- a/include/asm-powerpc/io.h 2008-04-24 12:23:23.000000000 +0900
+++ b/include/asm-powerpc/io.h 2008-04-24 12:23:25.000000000 +0900
@@ -458,8 +458,8 @@
/* Structure containing all the hooks */
extern struct ppc_pci_io {
-#define DEF_PCI_AC_RET(name, ret, at, al) ret (*name) at;
-#define DEF_PCI_AC_NORET(name, at, al) void (*name) at;
+#define DEF_PCI_AC_RET(name, ret, at, al, space, aa) ret (*name) at;
+#define DEF_PCI_AC_NORET(name, at, al, space, aa) void (*name) at;
#include <asm/io-defs.h>
@@ -469,7 +469,7 @@
} ppc_pci_io;
/* The inline wrappers */
-#define DEF_PCI_AC_RET(name, ret, at, al) \
+#define DEF_PCI_AC_RET(name, ret, at, al, space, aa) \
static inline ret name at \
{ \
if (DEF_PCI_HOOK(ppc_pci_io.name) != NULL) \
@@ -477,7 +477,7 @@
return __do_##name al; \
}
-#define DEF_PCI_AC_NORET(name, at, al) \
+#define DEF_PCI_AC_NORET(name, at, al, space, aa) \
static inline void name at \
{ \
if (DEF_PCI_HOOK(ppc_pci_io.name) != NULL) \
Index: b/arch/powerpc/kernel/of_platform.c
===================================================================
--- a/arch/powerpc/kernel/of_platform.c 2008-04-24 12:23:23.000000000 +0900
+++ b/arch/powerpc/kernel/of_platform.c 2008-04-24 12:23:25.000000000 +0900
@@ -275,6 +275,8 @@
/* Scan the bus */
scan_phb(phb);
+ if (phb->bus == NULL)
+ return -ENXIO;
/* Claim resources. This might need some rework as well depending
* wether we are doing probe-only or not, like assigning unassigned
Index: b/arch/powerpc/platforms/cell/setup.c
===================================================================
--- a/arch/powerpc/platforms/cell/setup.c 2008-04-24 12:23:23.000000000 +0900
+++ b/arch/powerpc/platforms/cell/setup.c 2008-04-24 12:23:25.000000000 +0900
@@ -57,6 +57,7 @@
#include "interrupt.h"
#include "pervasive.h"
#include "ras.h"
+#include "io-workarounds.h"
#ifdef DEBUG
#define DBG(fmt...) udbg_printf(fmt)
@@ -117,13 +118,50 @@
}
DECLARE_PCI_FIXUP_HEADER(PCI_ANY_ID, PCI_ANY_ID, cell_fixup_pcie_rootcomplex);
+static int __devinit cell_setup_phb(struct pci_controller *phb)
+{
+ const char *model;
+ struct device_node *np;
+
+ int rc = rtas_setup_phb(phb);
+ if (rc)
+ return rc;
+
+ np = phb->dn;
+ model = of_get_property(np, "model", NULL);
+ if (model == NULL || strcmp(np->name, "pci"))
+ return 0;
+
+ /* Setup workarounds for spider */
+ if (strcmp(model, "Spider"))
+ return 0;
+
+ iowa_register_bus(phb, &spiderpci_ops, &spiderpci_iowa_init,
+ (void *)SPIDER_PCI_REG_BASE);
+ io_workaround_init();
+
+ return 0;
+}
+
static int __init cell_publish_devices(void)
{
+ struct device_node *root = of_find_node_by_path("/");
+ struct device_node *np;
int node;
/* Publish OF platform devices for southbridge IOs */
of_platform_bus_probe(NULL, NULL, NULL);
+ /* On spider based blades, we need to manually create the OF
+ * platform devices for the PCI host bridges
+ */
+ for_each_child_of_node(root, np) {
+ if (np->type == NULL || (strcmp(np->type, "pci") != 0 &&
+ strcmp(np->type, "pciex") != 0))
+ continue;
+ of_platform_device_create(np, NULL, NULL);
+ }
+
/* There is no device for the MIC memory controller, thus we create
* a platform device for it to attach the EDAC driver to.
*/
@@ -132,6 +170,7 @@
continue;
platform_device_register_simple("cbe-mic", node, NULL, 0);
}
+
return 0;
}
machine_subsys_initcall(cell, cell_publish_devices);
@@ -213,7 +252,7 @@
/* Find and initialize PCI host bridges */
init_pci_config_tokens();
- find_and_init_phbs();
+
cbe_pervasive_init();
#ifdef CONFIG_DUMMY_CONSOLE
conswitchp = &dummy_con;
@@ -249,7 +288,7 @@
.calibrate_decr = generic_calibrate_decr,
.progress = cell_progress,
.init_IRQ = cell_init_irq,
- .pci_setup_phb = rtas_setup_phb,
+ .pci_setup_phb = cell_setup_phb,
#ifdef CONFIG_KEXEC
.machine_kexec = default_machine_kexec,
.machine_kexec_prepare = default_machine_kexec_prepare,
^ permalink raw reply
* ML405 failed reboot
From: Guillaume Dargaud @ 2008-04-24 8:34 UTC (permalink / raw)
To: linuxppc-dev
Now that my Xilinx card is operational with Kernel 2.6.24 and a
buildroot/busybox/uClibc config, I'm getting on the application side.
First hitch: I can't do a soft reboot of the card:
# reboot
The system is going down NOW!
Sending SIGTERM to all processes
Requesting system reboot
[ 434.713999] Restarting system.
[ 434.7[ 434.718123] Signal: 4
[ 434.718133] Code: 30001
[ 434.718144] Addr: c74cbdc8
And then it just hangs.
I also tried the additional -n (no sync) and -f (force) without difference.
Design problem ?
Any other way to force a reboot ?
--
Guillaume Dargaud
http://www.gdargaud.net/
^ permalink raw reply
* Re: [RFC][WIP][PATCH] Add IRQSTACKS to ppc32
From: Benjamin Herrenschmidt @ 2008-04-24 7:18 UTC (permalink / raw)
To: Kumar Gala; +Cc: linuxppc-dev
In-Reply-To: <Pine.LNX.4.64.0804240034160.23185@blarg.am.freescale.net>
On Thu, 2008-04-24 at 00:37 -0500, Kumar Gala wrote:
> + /*
> + * interrupt stacks must be under 256MB, we cannot afford to
> take
> + * SLB misses on them.
> + */
> + for_each_possible_cpu(i) {
> + softirq_ctx[i] = (struct thread_info *)
> + __va(lmb_alloc_base(THREAD_SIZE,
> + THREAD_SIZE, 0x10000000));
> + hardirq_ctx[i] = (struct thread_info *)
> + __va(lmb_alloc_base(THREAD_SIZE,
> + THREAD_SIZE, 0x10000000));
> + }
> +
The comment is a bit bogus :-) (about SLB misses). lowmem is your limit
I think. Also, why not share the code with ppc64 ?
Ben.
^ permalink raw reply
* Re: LMB regression...
From: David Miller @ 2008-04-24 7:01 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev
In-Reply-To: <18448.12058.743037.267216@cargo.ozlabs.ibm.com>
From: Paul Mackerras <paulus@samba.org>
Date: Thu, 24 Apr 2008 16:56:26 +1000
> David Miller writes:
>
> > Specifically, you removed the aligning of the size argument given to
> > lmb_add_region() in the lmb allocators, and that is critical when
> > allocating many small chunks, we run out of LMB slots otherwise
> > and allocations start failing.
>
> Sorry. It's still there in lmb_alloc_nid_unreserved but I lost the
> one in __lmb_alloc_base. Are you going to send a fix to Linus, or
> will I include it in the pull request I'm going to send tomorrow?
I've got it, see my other email.
^ permalink raw reply
* Re: LMB regression...
From: Paul Mackerras @ 2008-04-24 6:56 UTC (permalink / raw)
To: David Miller; +Cc: linuxppc-dev
In-Reply-To: <20080423.232419.193712136.davem@davemloft.net>
David Miller writes:
> Specifically, you removed the aligning of the size argument given to
> lmb_add_region() in the lmb allocators, and that is critical when
> allocating many small chunks, we run out of LMB slots otherwise
> and allocations start failing.
Sorry. It's still there in lmb_alloc_nid_unreserved but I lost the
one in __lmb_alloc_base. Are you going to send a fix to Linus, or
will I include it in the pull request I'm going to send tomorrow?
> I find it quite ironic that you spent so much time and effort fixing
> an absolutely totally theoretical bug that nobody was triggering, and
> in the process reintroduced a real one that triggers immediately on
> real systems, and that had even been explicitly fixed previously.
In fact the bug should be relatively easily triggerable in
lmb_alloc_nid as far as I can see, since it scans upwards
from zero. It depends a bit on how memory is distributed across
nodes, but it you are allocating on the node that has the memory
starting at physical address 0, and all of the first contiguous chunk
on that node is reserved, then I think you'll hit the bug. The only
saving grace is that lmb_alloc_nid is presently unused.
Paul.
^ permalink raw reply
* Re: LMB regression...
From: David Miller @ 2008-04-24 6:40 UTC (permalink / raw)
To: michael; +Cc: linuxppc-dev, paulus
In-Reply-To: <1209018933.22532.0.camel@concordia.ozlabs.ibm.com>
From: Michael Ellerman <michael@ellerman.id.au>
Date: Thu, 24 Apr 2008 16:35:33 +1000
> Sounds like we need a test suite :)
Maybe :-) Anyways, here is the bug fix I plan to push to
Linus with my sparc64 NUMA changes, unless someone has an
objection:
[LMB]: Fix lmb allocation regression.
Changeset d9024df02ffe74d723d97d552f86de3b34beb8cc ("[LMB] Restructure
allocation loops to avoid unsigned underflow") removed the alignment
of the 'size' argument to call lmb_add_region() done by __lmb_alloc_base().
In doing so it reintroduced the bug fixed by changeset
eea89e13a9c61d3928223d2f9bf2295e22e0efb6 ("[LMB]: Fix bug in
__lmb_alloc_base().").
This puts it back.
Signed-off-by: David S. Miller <davem@davemloft.net>
---
lib/lmb.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/lib/lmb.c b/lib/lmb.c
index 896e283..207147a 100644
--- a/lib/lmb.c
+++ b/lib/lmb.c
@@ -346,7 +346,7 @@ u64 __init __lmb_alloc_base(u64 size, u64 align, u64 max_addr)
if (j < 0) {
/* this area isn't reserved, take it */
if (lmb_add_region(&lmb.reserved, base,
- size) < 0)
+ lmb_align_up(size, align)) < 0)
return 0;
return base;
}
--
1.5.5.1.57.g5909c
^ permalink raw reply related
* Re: LMB regression...
From: Michael Ellerman @ 2008-04-24 6:35 UTC (permalink / raw)
To: David Miller; +Cc: linuxppc-dev, paulus
In-Reply-To: <20080423.232419.193712136.davem@davemloft.net>
[-- Attachment #1: Type: text/plain, Size: 1900 bytes --]
On Wed, 2008-04-23 at 23:24 -0700, David Miller wrote:
> Paul, I finally got around to testing your changeset on sparc64, it
> breaks things:
>
> commit d9024df02ffe74d723d97d552f86de3b34beb8cc
> Author: Paul Mackerras <paulus@samba.org>
> Date: Sat Apr 12 15:20:59 2008 +1000
>
> [LMB] Restructure allocation loops to avoid unsigned underflow
> ...
>
> Specifically, you removed the aligning of the size argument given to
> lmb_add_region() in the lmb allocators, and that is critical when
> allocating many small chunks, we run out of LMB slots otherwise
> and allocations start failing.
>
> I added the alignment there as a bug fix earlier:
>
> commit eea89e13a9c61d3928223d2f9bf2295e22e0efb6
> Author: David S. Miller <davem@davemloft.net>
> Date: Wed Feb 13 16:57:09 2008 -0800
>
> [LMB]: Fix bug in __lmb_alloc_base().
>
> We need to check lmb_add_region() for errors, it can run out
> of regions etc.
>
> Also, the size needs to be padded to the given alignment
> or else the lmb.reserved regions don't get expanded and
> instead we get tons of holes and eventually run out of
> regions prematurely.
>
> Signed-off-by: David S. Miller <davem@davemloft.net>
>
> Please be more careful in the future :-(
>
> I find it quite ironic that you spent so much time and effort fixing
> an absolutely totally theoretical bug that nobody was triggering, and
> in the process reintroduced a real one that triggers immediately on
> real systems, and that had even been explicitly fixed previously.
Sounds like we need a test suite :)
cheers
--
Michael Ellerman
OzLabs, IBM Australia Development Lab
wwweb: http://michael.ellerman.id.au
phone: +61 2 6212 1183 (tie line 70 21183)
We do not inherit the earth from our ancestors,
we borrow it from our children. - S.M.A.R.T Person
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]
^ permalink raw reply
* Re: [PATCH] powerpc TLF_RESTORE_SIGMASK
From: Paul Mackerras @ 2008-04-24 6:34 UTC (permalink / raw)
To: Roland McGrath; +Cc: linuxppc-dev, akpm, linux-kernel
In-Reply-To: <20080423063753.C2088270381@magilla.localdomain>
Roland McGrath writes:
> Indeed so. Here's a replacement patch.
Oh, and could you please also fix the occurrence of
TIF_RESTORE_SIGMASK in arch/ppc/kernel/entry.S when this stuff goes
in?
Thanks,
Paul.
^ permalink raw reply
* Re: [PATCH] powerpc TLF_RESTORE_SIGMASK
From: Paul Mackerras @ 2008-04-24 6:31 UTC (permalink / raw)
To: Roland McGrath; +Cc: linuxppc-dev, akpm, linux-kernel
In-Reply-To: <20080423063753.C2088270381@magilla.localdomain>
Roland McGrath writes:
> > This should be _TLF_RESTORE_SIGMASK (leading '_'), I think.
>
> Indeed so. Here's a replacement patch.
Thanks. That can't go in until some generic changes have gone in
first, right? So are you going to push the lot in one go via Andrew,
or should I wait until the generic part is upstream and then push the
powerpc bit?
Paul.
^ permalink raw reply
* LMB regression...
From: David Miller @ 2008-04-24 6:24 UTC (permalink / raw)
To: linuxppc-dev; +Cc: paulus
Paul, I finally got around to testing your changeset on sparc64, it
breaks things:
commit d9024df02ffe74d723d97d552f86de3b34beb8cc
Author: Paul Mackerras <paulus@samba.org>
Date: Sat Apr 12 15:20:59 2008 +1000
[LMB] Restructure allocation loops to avoid unsigned underflow
...
Specifically, you removed the aligning of the size argument given to
lmb_add_region() in the lmb allocators, and that is critical when
allocating many small chunks, we run out of LMB slots otherwise
and allocations start failing.
I added the alignment there as a bug fix earlier:
commit eea89e13a9c61d3928223d2f9bf2295e22e0efb6
Author: David S. Miller <davem@davemloft.net>
Date: Wed Feb 13 16:57:09 2008 -0800
[LMB]: Fix bug in __lmb_alloc_base().
We need to check lmb_add_region() for errors, it can run out
of regions etc.
Also, the size needs to be padded to the given alignment
or else the lmb.reserved regions don't get expanded and
instead we get tons of holes and eventually run out of
regions prematurely.
Signed-off-by: David S. Miller <davem@davemloft.net>
Please be more careful in the future :-(
I find it quite ironic that you spent so much time and effort fixing
an absolutely totally theoretical bug that nobody was triggering, and
in the process reintroduced a real one that triggers immediately on
real systems, and that had even been explicitly fixed previously.
^ permalink raw reply
* Re: [BUG] 2.6.25-rc2-git4 - Regression Kernel oops while running kernbench and tbench on powerpc
From: Kamalesh Babulal @ 2008-04-24 6:05 UTC (permalink / raw)
To: Paul Mackerras
Cc: kernel list, linuxppc-dev, linux-next, nacc, Andrew Morton,
Balbir Singh
In-Reply-To: <18446.61538.620549.715043@cargo.ozlabs.ibm.com>
Paul Mackerras wrote:
> Kamalesh Babulal writes:
>
>> After applying the patch above and the patch posted on
>> http://lkml.org/lkml/2008/4/8/42
>> the bug had the following information,
>
> Thanks. The patch below, against Linus' current git tree, fixes one
> bug that might be the cause of the problem, and also attempts to
> detect the erroneous situation earlier and fix it up, and also print
> some debug information. Please try to reproduce the problem with this
> patch applied, and if there are any console log messages starting with
> SLB: or FWNMI:, please send me the console log.
>
> Paul.
>
> diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
> index c0db5b7..f7f0962 100644
> --- a/arch/powerpc/kernel/entry_64.S
> +++ b/arch/powerpc/kernel/entry_64.S
> @@ -439,6 +439,19 @@ END_FTR_SECTION_IFSET(CPU_FTR_1T_SEGMENT)
> mr r1,r8 /* start using new stack pointer */
> std r7,PACAKSAVE(r13)
>
> + /* check that SLB entry 2 contains the right thing */
> + clrrdi r6,r1,28
> + clrldi. r0,r6,2
> + beq 3f
> + li r0,2
> + slbmfee r7,r0
> + oris r6,r6,SLB_ESID_V@h
> + cmpd r6,r7
> + beq 3f
> + bl bad_slb_switch
> + ld r3,PACACURRENT(r13)
> + addi r3,r3,THREAD
> +3:
> ld r6,_CCR(r1)
> mtcrf 0xFF,r6
>
> @@ -540,6 +553,19 @@ END_FW_FTR_SECTION_IFSET(FW_FEATURE_ISERIES)
> ld r4,_XER(r1)
> mtspr SPRN_XER,r4
>
> + /* check that SLB entry 2 contains the right thing */
> + clrrdi r6,r1,28 /* stack ESID */
> + clrldi. r0,r6,2
> + beq 57f
> + li r0,2
> + slbmfee r7,r0
> + oris r6,r6,SLB_ESID_V@h
> + cmpd r6,r7
> + beq 57f
> + addi r3,r1,STACK_FRAME_OVERHEAD
> + bl bad_slb_exc
> + ld r3,_MSR(r1)
> +57:
> REST_8GPRS(5, r1)
>
> andi. r0,r3,MSR_RI
> diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
> index be35ffa..c938134 100644
> --- a/arch/powerpc/kernel/smp.c
> +++ b/arch/powerpc/kernel/smp.c
> @@ -45,6 +45,7 @@
> #include <asm/system.h>
> #include <asm/mpic.h>
> #include <asm/vdso_datapage.h>
> +#include <asm/mmu.h>
> #ifdef CONFIG_PPC64
> #include <asm/paca.h>
> #endif
> @@ -580,6 +581,10 @@ int __devinit start_secondary(void *unused)
> atomic_inc(&init_mm.mm_count);
> current->active_mm = &init_mm;
>
> + /* Bolt in the entry for the kernel stack now */
> + if (cpu_has_feature(CPU_FTR_SLB))
> + slb_flush_and_rebolt();
> +
> smp_store_cpu_info(cpu);
> set_dec(tb_ticks_per_jiffy);
> preempt_disable();
> diff --git a/arch/powerpc/mm/slb.c b/arch/powerpc/mm/slb.c
> index 906daed..bb7765b 100644
> --- a/arch/powerpc/mm/slb.c
> +++ b/arch/powerpc/mm/slb.c
> @@ -309,3 +309,34 @@ void slb_initialize(void)
> * one. */
> asm volatile("isync":::"memory");
> }
> +
> +static void dump_slb(void)
> +{
> + long entry;
> + unsigned long esid, vsid;
> +
> + printk(KERN_EMERG "SLB contents now:\n");
> + for (entry = 0; entry < 64; ++entry) {
> + asm volatile("slbmfee %0,%1" : "=r" (esid) : "r" (entry));
> + if (esid == 0)
> + /* valid bit is clear along with everything else */
> + continue;
> + asm volatile("slbmfev %0,%1" : "=r" (vsid) : "r" (entry));
> + printk(KERN_EMERG "%d: %.16lx %.16lx\n", entry, esid, vsid);
> + }
> +}
> +
> +void bad_slb_exc(struct pt_regs *regs)
> +{
> + printk(KERN_EMERG "SLB: stack not bolted on exception return\n");
> + dump_slb();
> + slb_flush_and_rebolt();
> + show_regs(regs);
> +}
> +
> +void bad_slb_switch(void)
> +{
> + printk(KERN_EMERG "SLB: stack not bolted on context switch\n");
> + dump_slb();
> + slb_flush_and_rebolt();
> +}
> diff --git a/arch/powerpc/platforms/pseries/ras.c b/arch/powerpc/platforms/pseries/ras.c
> index a1ab25c..ed68083 100644
> --- a/arch/powerpc/platforms/pseries/ras.c
> +++ b/arch/powerpc/platforms/pseries/ras.c
> @@ -325,6 +325,8 @@ static int recover_mce(struct pt_regs *regs, struct rtas_error_log * err)
>
> if (err->disposition == RTAS_DISP_FULLY_RECOVERED) {
> /* Platform corrected itself */
> + printk(KERN_ALERT "FWNMI: platform corrected error %.16lx\n",
> + *(unsigned long *)err);
> nonfatal = 1;
> } else if ((regs->msr & MSR_RI) &&
> user_mode(regs) &&
Hi Paul,
Thanks, after applying the patch the oops is not reproducible on the machine. The console
log had no message starting with SLB: or FWNMI:. I have updated the bugzilla also.
Tested-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
--
Thanks & Regards,
Kamalesh Babulal,
Linux Technology Center,
IBM, ISTL.
^ permalink raw reply
* [RFC][WIP][PATCH] Add IRQSTACKS to ppc32
From: Kumar Gala @ 2008-04-24 5:37 UTC (permalink / raw)
To: linuxppc-dev
Posting this to get any review and suggestions on the functionality of the
patch.
Questions/issues:
* what to do about the stack_ovf check in entry_32.S
* do we really have any constraints on ppc32 (beyond being in lowmem) on
the locations of the stacks [see irqstack_early_init()]
- k
diff --git a/arch/powerpc/Kconfig.debug b/arch/powerpc/Kconfig.debug
index a86d8d8..2cf72d2 100644
--- a/arch/powerpc/Kconfig.debug
+++ b/arch/powerpc/Kconfig.debug
@@ -118,7 +118,6 @@ config XMON_DISASSEMBLY
config IRQSTACKS
bool "Use separate kernel stacks when processing interrupts"
- depends on PPC64
help
If you say Y here the kernel will use separate kernel stacks
for handling hard and soft interrupts. This can help avoid
diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index 425616f..ad40eb4 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c
@@ -352,7 +352,7 @@ void __init init_IRQ(void)
{
if (ppc_md.init_IRQ)
ppc_md.init_IRQ();
-#ifdef CONFIG_PPC64
+#ifdef CONFIG_IRQSTACKS
irq_ctx_init();
#endif
}
diff --git a/arch/powerpc/kernel/misc_32.S b/arch/powerpc/kernel/misc_32.S
index 92ccc6f..89aaaa6 100644
--- a/arch/powerpc/kernel/misc_32.S
+++ b/arch/powerpc/kernel/misc_32.S
@@ -32,6 +32,31 @@
.text
+#ifdef CONFIG_IRQSTACKS
+_GLOBAL(call_do_softirq)
+ mflr r0
+ stw r0,4(r1)
+ stwu r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r3)
+ mr r1,r3
+ bl __do_softirq
+ lwz r1,0(r1)
+ lwz r0,4(r1)
+ mtlr r0
+ blr
+
+_GLOBAL(call_handle_irq)
+ mflr r0
+ stw r0,4(r1)
+ mtctr r6
+ stwu r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r5)
+ mr r1,r5
+ bctrl
+ lwz r1,0(r1)
+ lwz r0,4(r1)
+ mtlr r0
+ blr
+#endif /* CONFIG_IRQSTACKS */
+
/*
* This returns the high 64 bits of the product of two 64-bit numbers.
*/
diff --git a/arch/powerpc/kernel/setup_32.c b/arch/powerpc/kernel/setup_32.c
index 36f6779..ebd1b1d 100644
--- a/arch/powerpc/kernel/setup_32.c
+++ b/arch/powerpc/kernel/setup_32.c
@@ -16,6 +16,7 @@
#include <linux/root_dev.h>
#include <linux/cpu.h>
#include <linux/console.h>
+#include <linux/lmb.h>
#include <asm/io.h>
#include <asm/prom.h>
@@ -229,6 +230,28 @@ int __init ppc_init(void)
arch_initcall(ppc_init);
+#ifdef CONFIG_IRQSTACKS
+static void __init irqstack_early_init(void)
+{
+ unsigned int i;
+
+ /*
+ * interrupt stacks must be under 256MB, we cannot afford to take
+ * SLB misses on them.
+ */
+ for_each_possible_cpu(i) {
+ softirq_ctx[i] = (struct thread_info *)
+ __va(lmb_alloc_base(THREAD_SIZE,
+ THREAD_SIZE, 0x10000000));
+ hardirq_ctx[i] = (struct thread_info *)
+ __va(lmb_alloc_base(THREAD_SIZE,
+ THREAD_SIZE, 0x10000000));
+ }
+}
+#else
+#define irqstack_early_init()
+#endif
+
/* Warning, IO base is not yet inited */
void __init setup_arch(char **cmdline_p)
{
@@ -286,6 +309,8 @@ void __init setup_arch(char **cmdline_p)
init_mm.end_data = (unsigned long) _edata;
init_mm.brk = klimit;
+ irqstack_early_init();
+
/* set up the bootmem stuff with available memory */
do_init_bootmem();
if ( ppc_md.progress ) ppc_md.progress("setup_arch: bootmem", 0x3eab);
^ permalink raw reply related
* [PATCH 5/5] Add CONFIG_PPC_PSERIES_DEBUG to enable debugging for platforms/pseries
From: Michael Ellerman @ 2008-04-24 5:13 UTC (permalink / raw)
To: linuxppc-dev
In-Reply-To: <c01702547f6296b2c18e5ff7504ad52cb0402664.1209013992.git.michael@ellerman.id.au>
Add a DEBUG config setting which turns on all (most) of the debugging
under platforms/pseries.
To have this take effect we need to remove all the #undef DEBUG's, in
various files. We leave the #undef DEBUG in platforms/pseries/lpar.c,
as this enables debugging printks from the low-level hash table routines,
and tends to make your system unusable. If you want those enabled you
still have to turn them on by hand.
Also some of the RAS code has a DEBUG block which causes a functional
change, so I've keyed this off a different (non-existant) debug #define.
This is only enabled if you have PPC_EARLY_DEBUG enabled also.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
---
arch/powerpc/platforms/pseries/Kconfig | 5 +++++
arch/powerpc/platforms/pseries/Makefile | 4 ++++
arch/powerpc/platforms/pseries/eeh.c | 1 -
arch/powerpc/platforms/pseries/eeh_cache.c | 1 -
arch/powerpc/platforms/pseries/firmware.c | 1 -
arch/powerpc/platforms/pseries/ras.c | 4 +---
arch/powerpc/platforms/pseries/setup.c | 2 --
arch/powerpc/platforms/pseries/smp.c | 1 -
arch/powerpc/platforms/pseries/xics.c | 1 -
9 files changed, 10 insertions(+), 10 deletions(-)
diff --git a/arch/powerpc/platforms/pseries/Kconfig b/arch/powerpc/platforms/pseries/Kconfig
index 306a9d0..07fe5b6 100644
--- a/arch/powerpc/platforms/pseries/Kconfig
+++ b/arch/powerpc/platforms/pseries/Kconfig
@@ -34,3 +34,8 @@ config LPARCFG
help
Provide system capacity information via human readable
<key word>=<value> pairs through a /proc/ppc64/lparcfg interface.
+
+config PPC_PSERIES_DEBUG
+ depends on PPC_PSERIES && PPC_EARLY_DEBUG
+ bool "Enable extra debug logging in platforms/pseries"
+ default y
diff --git a/arch/powerpc/platforms/pseries/Makefile b/arch/powerpc/platforms/pseries/Makefile
index bdae04b..bd2593e 100644
--- a/arch/powerpc/platforms/pseries/Makefile
+++ b/arch/powerpc/platforms/pseries/Makefile
@@ -2,6 +2,10 @@ ifeq ($(CONFIG_PPC64),y)
EXTRA_CFLAGS += -mno-minimal-toc
endif
+ifeq ($(CONFIG_PPC_PSERIES_DEBUG),y)
+EXTRA_CFLAGS += -DDEBUG
+endif
+
obj-y := lpar.o hvCall.o nvram.o reconfig.o \
setup.o iommu.o ras.o rtasd.o \
firmware.o power.o
diff --git a/arch/powerpc/platforms/pseries/eeh.c b/arch/powerpc/platforms/pseries/eeh.c
index 550b2f7..a3fd56b 100644
--- a/arch/powerpc/platforms/pseries/eeh.c
+++ b/arch/powerpc/platforms/pseries/eeh.c
@@ -39,7 +39,6 @@
#include <asm/ppc-pci.h>
#include <asm/rtas.h>
-#undef DEBUG
/** Overview:
* EEH, or "Extended Error Handling" is a PCI bridge technology for
diff --git a/arch/powerpc/platforms/pseries/eeh_cache.c b/arch/powerpc/platforms/pseries/eeh_cache.c
index 1e83fcd..ce37040 100644
--- a/arch/powerpc/platforms/pseries/eeh_cache.c
+++ b/arch/powerpc/platforms/pseries/eeh_cache.c
@@ -28,7 +28,6 @@
#include <asm/pci-bridge.h>
#include <asm/ppc-pci.h>
-#undef DEBUG
/**
* The pci address cache subsystem. This subsystem places
diff --git a/arch/powerpc/platforms/pseries/firmware.c b/arch/powerpc/platforms/pseries/firmware.c
index 743d494..9d3a40f 100644
--- a/arch/powerpc/platforms/pseries/firmware.c
+++ b/arch/powerpc/platforms/pseries/firmware.c
@@ -21,7 +21,6 @@
* 2 of the License, or (at your option) any later version.
*/
-#undef DEBUG
#include <asm/firmware.h>
#include <asm/prom.h>
diff --git a/arch/powerpc/platforms/pseries/ras.c b/arch/powerpc/platforms/pseries/ras.c
index a1ab25c..2b548af 100644
--- a/arch/powerpc/platforms/pseries/ras.c
+++ b/arch/powerpc/platforms/pseries/ras.c
@@ -67,8 +67,6 @@ static int ras_check_exception_token;
static irqreturn_t ras_epow_interrupt(int irq, void *dev_id);
static irqreturn_t ras_error_interrupt(int irq, void *dev_id);
-/* #define DEBUG */
-
static void request_ras_irqs(struct device_node *np,
irq_handler_t handler,
@@ -237,7 +235,7 @@ static irqreturn_t ras_error_interrupt(int irq, void *dev_id)
printk(KERN_EMERG "Error: Fatal hardware error <0x%lx 0x%x>\n",
*((unsigned long *)&ras_log_buf), status);
-#ifndef DEBUG
+#ifndef DEBUG_RTAS_POWER_OFF
/* Don't actually power off when debugging so we can test
* without actually failing while injecting errors.
* Error data will not be logged to syslog.
diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index 65e8795..f5d29f5 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -16,8 +16,6 @@
* bootup setup stuff..
*/
-#undef DEBUG
-
#include <linux/cpu.h>
#include <linux/errno.h>
#include <linux/sched.h>
diff --git a/arch/powerpc/platforms/pseries/smp.c b/arch/powerpc/platforms/pseries/smp.c
index e9bc2a5..9d8f8c8 100644
--- a/arch/powerpc/platforms/pseries/smp.c
+++ b/arch/powerpc/platforms/pseries/smp.c
@@ -12,7 +12,6 @@
* 2 of the License, or (at your option) any later version.
*/
-#undef DEBUG
#include <linux/kernel.h>
#include <linux/module.h>
diff --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c
index 43df53c..ebebc28 100644
--- a/arch/powerpc/platforms/pseries/xics.c
+++ b/arch/powerpc/platforms/pseries/xics.c
@@ -9,7 +9,6 @@
* 2 of the License, or (at your option) any later version.
*/
-#undef DEBUG
#include <linux/types.h>
#include <linux/threads.h>
--
1.5.5
^ permalink raw reply related
* [PATCH 4/5] Convert from DBG() to pr_debug() in platforms/pseries/
From: Michael Ellerman @ 2008-04-24 5:13 UTC (permalink / raw)
To: linuxppc-dev
In-Reply-To: <c01702547f6296b2c18e5ff7504ad52cb0402664.1209013992.git.michael@ellerman.id.au>
In pseries/lpar.c, fix some printf specifier mismatches, and add
a newline to one printk.
In pseries/rtasd.c add "rtasd" to some messages to make it clear
where they're coming from.
In pseries/scanlog.c remove the hand-rolled runtime debugging support
in there. This file has been largely unchanged for eons, if we need to
debug it in future we can recompile.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
---
arch/powerpc/platforms/pseries/firmware.c | 9 +-----
arch/powerpc/platforms/pseries/iommu.c | 39 ++++++++++++++---------------
arch/powerpc/platforms/pseries/lpar.c | 34 +++++++++++--------------
arch/powerpc/platforms/pseries/rtasd.c | 14 +++------
arch/powerpc/platforms/pseries/scanlog.c | 23 ++++++-----------
arch/powerpc/platforms/pseries/setup.c | 15 +++-------
arch/powerpc/platforms/pseries/smp.c | 10 +------
7 files changed, 56 insertions(+), 88 deletions(-)
diff --git a/arch/powerpc/platforms/pseries/firmware.c b/arch/powerpc/platforms/pseries/firmware.c
index b765b7c..743d494 100644
--- a/arch/powerpc/platforms/pseries/firmware.c
+++ b/arch/powerpc/platforms/pseries/firmware.c
@@ -27,11 +27,6 @@
#include <asm/prom.h>
#include <asm/udbg.h>
-#ifdef DEBUG
-#define DBG(fmt...) udbg_printf(fmt)
-#else
-#define DBG(fmt...)
-#endif
typedef struct {
unsigned long val;
@@ -72,7 +67,7 @@ void __init fw_feature_init(const char *hypertas, unsigned long len)
const char *s;
int i;
- DBG(" -> fw_feature_init()\n");
+ pr_debug(" -> fw_feature_init()\n");
for (s = hypertas; s < hypertas + len; s += strlen(s) + 1) {
for (i = 0; i < FIRMWARE_MAX_FEATURES; i++) {
@@ -88,5 +83,5 @@ void __init fw_feature_init(const char *hypertas, unsigned long len)
}
}
- DBG(" <- fw_feature_init()\n");
+ pr_debug(" <- fw_feature_init()\n");
}
diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c
index a65c763..176f1f3 100644
--- a/arch/powerpc/platforms/pseries/iommu.c
+++ b/arch/powerpc/platforms/pseries/iommu.c
@@ -47,7 +47,6 @@
#include "plpar_wrappers.h"
-#define DBG(fmt...)
static void tce_build_pSeries(struct iommu_table *tbl, long index,
long npages, unsigned long uaddr,
@@ -322,7 +321,7 @@ static void pci_dma_bus_setup_pSeries(struct pci_bus *bus)
dn = pci_bus_to_OF_node(bus);
- DBG("pci_dma_bus_setup_pSeries: setting up bus %s\n", dn->full_name);
+ pr_debug("pci_dma_bus_setup_pSeries: setting up bus %s\n", dn->full_name);
if (bus->self) {
/* This is not a root bus, any setup will be done for the
@@ -347,7 +346,7 @@ static void pci_dma_bus_setup_pSeries(struct pci_bus *bus)
for (children = 0, tmp = dn->child; tmp; tmp = tmp->sibling)
children++;
- DBG("Children: %d\n", children);
+ pr_debug("Children: %d\n", children);
/* Calculate amount of DMA window per slot. Each window must be
* a power of two (due to pci_alloc_consistent requirements).
@@ -361,8 +360,8 @@ static void pci_dma_bus_setup_pSeries(struct pci_bus *bus)
while (pci->phb->dma_window_size * children > 0x80000000ul)
pci->phb->dma_window_size >>= 1;
- DBG("No ISA/IDE, window size is 0x%lx\n",
- pci->phb->dma_window_size);
+ pr_debug("No ISA/IDE, window size is 0x%lx\n",
+ pci->phb->dma_window_size);
pci->phb->dma_window_base_cur = 0;
return;
@@ -387,8 +386,7 @@ static void pci_dma_bus_setup_pSeries(struct pci_bus *bus)
while (pci->phb->dma_window_size * children > 0x70000000ul)
pci->phb->dma_window_size >>= 1;
- DBG("ISA/IDE, window size is 0x%lx\n", pci->phb->dma_window_size);
-
+ pr_debug("ISA/IDE, window size is 0x%lx\n", pci->phb->dma_window_size);
}
@@ -401,7 +399,8 @@ static void pci_dma_bus_setup_pSeriesLP(struct pci_bus *bus)
dn = pci_bus_to_OF_node(bus);
- DBG("pci_dma_bus_setup_pSeriesLP: setting up bus %s\n", dn->full_name);
+ pr_debug("pci_dma_bus_setup_pSeriesLP: setting up bus %s\n",
+ dn->full_name);
/* Find nearest ibm,dma-window, walking up the device tree */
for (pdn = dn; pdn != NULL; pdn = pdn->parent) {
@@ -411,14 +410,14 @@ static void pci_dma_bus_setup_pSeriesLP(struct pci_bus *bus)
}
if (dma_window == NULL) {
- DBG(" no ibm,dma-window property !\n");
+ pr_debug(" no ibm,dma-window property !\n");
return;
}
ppci = PCI_DN(pdn);
- DBG(" parent is %s, iommu_table: 0x%p\n",
- pdn->full_name, ppci->iommu_table);
+ pr_debug(" parent is %s, iommu_table: 0x%p\n",
+ pdn->full_name, ppci->iommu_table);
if (!ppci->iommu_table) {
tbl = kmalloc_node(sizeof(struct iommu_table), GFP_KERNEL,
@@ -426,7 +425,7 @@ static void pci_dma_bus_setup_pSeriesLP(struct pci_bus *bus)
iommu_table_setparms_lpar(ppci->phb, pdn, tbl, dma_window,
bus->number);
ppci->iommu_table = iommu_init_table(tbl, ppci->phb->node);
- DBG(" created table: %p\n", ppci->iommu_table);
+ pr_debug(" created table: %p\n", ppci->iommu_table);
}
if (pdn != dn)
@@ -439,7 +438,7 @@ static void pci_dma_dev_setup_pSeries(struct pci_dev *dev)
struct device_node *dn;
struct iommu_table *tbl;
- DBG("pci_dma_dev_setup_pSeries: %s\n", pci_name(dev));
+ pr_debug("pci_dma_dev_setup_pSeries: %s\n", pci_name(dev));
dn = dev->dev.archdata.of_node;
@@ -450,7 +449,7 @@ static void pci_dma_dev_setup_pSeries(struct pci_dev *dev)
if (!dev->bus->self) {
struct pci_controller *phb = PCI_DN(dn)->phb;
- DBG(" --> first child, no bridge. Allocating iommu table.\n");
+ pr_debug(" --> first child, no bridge. Allocating iommu table.\n");
tbl = kmalloc_node(sizeof(struct iommu_table), GFP_KERNEL,
phb->node);
iommu_table_setparms(phb, dn, tbl);
@@ -480,7 +479,7 @@ static void pci_dma_dev_setup_pSeriesLP(struct pci_dev *dev)
const void *dma_window = NULL;
struct pci_dn *pci;
- DBG("pci_dma_dev_setup_pSeriesLP: %s\n", pci_name(dev));
+ pr_debug("pci_dma_dev_setup_pSeriesLP: %s\n", pci_name(dev));
/* dev setup for LPAR is a little tricky, since the device tree might
* contain the dma-window properties per-device and not neccesarily
@@ -489,7 +488,7 @@ static void pci_dma_dev_setup_pSeriesLP(struct pci_dev *dev)
* already allocated.
*/
dn = pci_device_to_OF_node(dev);
- DBG(" node is %s\n", dn->full_name);
+ pr_debug(" node is %s\n", dn->full_name);
for (pdn = dn; pdn && PCI_DN(pdn) && !PCI_DN(pdn)->iommu_table;
pdn = pdn->parent) {
@@ -504,13 +503,13 @@ static void pci_dma_dev_setup_pSeriesLP(struct pci_dev *dev)
pci_name(dev), dn? dn->full_name : "<null>");
return;
}
- DBG(" parent is %s\n", pdn->full_name);
+ pr_debug(" parent is %s\n", pdn->full_name);
/* Check for parent == NULL so we don't try to setup the empty EADS
* slots on POWER4 machines.
*/
if (dma_window == NULL || pdn->parent == NULL) {
- DBG(" no dma window for device, linking to parent\n");
+ pr_debug(" no dma window for device, linking to parent\n");
dev->dev.archdata.dma_data = PCI_DN(pdn)->iommu_table;
return;
}
@@ -522,9 +521,9 @@ static void pci_dma_dev_setup_pSeriesLP(struct pci_dev *dev)
iommu_table_setparms_lpar(pci->phb, pdn, tbl, dma_window,
pci->phb->bus->number);
pci->iommu_table = iommu_init_table(tbl, pci->phb->node);
- DBG(" created table: %p\n", pci->iommu_table);
+ pr_debug(" created table: %p\n", pci->iommu_table);
} else {
- DBG(" found DMA window, table: %p\n", pci->iommu_table);
+ pr_debug(" found DMA window, table: %p\n", pci->iommu_table);
}
dev->dev.archdata.dma_data = pci->iommu_table;
diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c
index c34789f..2cbaedb 100644
--- a/arch/powerpc/platforms/pseries/lpar.c
+++ b/arch/powerpc/platforms/pseries/lpar.c
@@ -19,7 +19,8 @@
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
-#undef DEBUG_LOW
+/* Enables debugging of low-level hash table routines - careful! */
+#undef DEBUG
#include <linux/kernel.h>
#include <linux/dma-mapping.h>
@@ -42,11 +43,6 @@
#include "plpar_wrappers.h"
#include "pseries.h"
-#ifdef DEBUG_LOW
-#define DBG_LOW(fmt...) do { udbg_printf(fmt); } while(0)
-#else
-#define DBG_LOW(fmt...) do { } while(0)
-#endif
/* in hvCall.S */
EXPORT_SYMBOL(plpar_hcall);
@@ -290,15 +286,15 @@ static long pSeries_lpar_hpte_insert(unsigned long hpte_group,
unsigned long hpte_v, hpte_r;
if (!(vflags & HPTE_V_BOLTED))
- DBG_LOW("hpte_insert(group=%lx, va=%016lx, pa=%016lx, "
- "rflags=%lx, vflags=%lx, psize=%d)\n",
- hpte_group, va, pa, rflags, vflags, psize);
+ pr_debug("hpte_insert(group=%lx, va=%016lx, pa=%016lx, "
+ "rflags=%lx, vflags=%lx, psize=%d)\n",
+ hpte_group, va, pa, rflags, vflags, psize);
hpte_v = hpte_encode_v(va, psize, ssize) | vflags | HPTE_V_VALID;
hpte_r = hpte_encode_r(pa, psize) | rflags;
if (!(vflags & HPTE_V_BOLTED))
- DBG_LOW(" hpte_v=%016lx, hpte_r=%016lx\n", hpte_v, hpte_r);
+ pr_debug(" hpte_v=%016lx, hpte_r=%016lx\n", hpte_v, hpte_r);
/* Now fill in the actual HPTE */
/* Set CEC cookie to 0 */
@@ -315,7 +311,7 @@ static long pSeries_lpar_hpte_insert(unsigned long hpte_group,
lpar_rc = plpar_pte_enter(flags, hpte_group, hpte_v, hpte_r, &slot);
if (unlikely(lpar_rc == H_PTEG_FULL)) {
if (!(vflags & HPTE_V_BOLTED))
- DBG_LOW(" full\n");
+ pr_debug(" full\n");
return -1;
}
@@ -326,11 +322,11 @@ static long pSeries_lpar_hpte_insert(unsigned long hpte_group,
*/
if (unlikely(lpar_rc != H_SUCCESS)) {
if (!(vflags & HPTE_V_BOLTED))
- DBG_LOW(" lpar err %d\n", lpar_rc);
+ pr_debug(" lpar err %lu\n", lpar_rc);
return -2;
}
if (!(vflags & HPTE_V_BOLTED))
- DBG_LOW(" -> slot: %d\n", slot & 7);
+ pr_debug(" -> slot: %lu\n", slot & 7);
/* Because of iSeries, we have to pass down the secondary
* bucket bit here as well
@@ -422,17 +418,17 @@ static long pSeries_lpar_hpte_updatepp(unsigned long slot,
want_v = hpte_encode_avpn(va, psize, ssize);
- DBG_LOW(" update: avpnv=%016lx, hash=%016lx, f=%x, psize: %d ... ",
- want_v, slot, flags, psize);
+ pr_debug(" update: avpnv=%016lx, hash=%016lx, f=%lx, psize: %d ...",
+ want_v, slot, flags, psize);
lpar_rc = plpar_pte_protect(flags, slot, want_v);
if (lpar_rc == H_NOT_FOUND) {
- DBG_LOW("not found !\n");
+ pr_debug("not found !\n");
return -1;
}
- DBG_LOW("ok\n");
+ pr_debug("ok\n");
BUG_ON(lpar_rc != H_SUCCESS);
@@ -507,8 +503,8 @@ static void pSeries_lpar_hpte_invalidate(unsigned long slot, unsigned long va,
unsigned long lpar_rc;
unsigned long dummy1, dummy2;
- DBG_LOW(" inval : slot=%lx, va=%016lx, psize: %d, local: %d",
- slot, va, psize, local);
+ pr_debug(" inval : slot=%lx, va=%016lx, psize: %d, local: %d\n",
+ slot, va, psize, local);
want_v = hpte_encode_avpn(va, psize, ssize);
lpar_rc = plpar_pte_remove(H_AVPN, slot, want_v, &dummy1, &dummy2);
diff --git a/arch/powerpc/platforms/pseries/rtasd.c b/arch/powerpc/platforms/pseries/rtasd.c
index e3078ce..befadd4 100644
--- a/arch/powerpc/platforms/pseries/rtasd.c
+++ b/arch/powerpc/platforms/pseries/rtasd.c
@@ -29,11 +29,6 @@
#include <asm/atomic.h>
#include <asm/machdep.h>
-#if 0
-#define DEBUG(A...) printk(KERN_ERR A)
-#else
-#define DEBUG(A...)
-#endif
static DEFINE_SPINLOCK(rtasd_log_lock);
@@ -198,7 +193,7 @@ void pSeries_log_error(char *buf, unsigned int err_type, int fatal)
unsigned long s;
int len = 0;
- DEBUG("logging event\n");
+ pr_debug("rtasd: logging event\n");
if (buf == NULL)
return;
@@ -409,7 +404,8 @@ static int rtasd(void *unused)
daemonize("rtasd");
printk(KERN_DEBUG "RTAS daemon started\n");
- DEBUG("will sleep for %d milliseconds\n", (30000/rtas_event_scan_rate));
+ pr_debug("rtasd: will sleep for %d milliseconds\n",
+ (30000 / rtas_event_scan_rate));
/* See if we have any error stored in NVRAM */
memset(logdata, 0, rtas_error_log_max);
@@ -428,9 +424,9 @@ static int rtasd(void *unused)
do_event_scan_all_cpus(1000);
if (surveillance_timeout != -1) {
- DEBUG("enabling surveillance\n");
+ pr_debug("rtasd: enabling surveillance\n");
enable_surveillance(surveillance_timeout);
- DEBUG("surveillance enabled\n");
+ pr_debug("rtasd: surveillance enabled\n");
}
/* Delay should be at least one second since some
diff --git a/arch/powerpc/platforms/pseries/scanlog.c b/arch/powerpc/platforms/pseries/scanlog.c
index e5b0ea8..bec3803 100644
--- a/arch/powerpc/platforms/pseries/scanlog.c
+++ b/arch/powerpc/platforms/pseries/scanlog.c
@@ -38,9 +38,7 @@
#define SCANLOG_HWERROR -1
#define SCANLOG_CONTINUE 1
-#define DEBUG(A...) do { if (scanlog_debug) printk(KERN_ERR "scanlog: " A); } while (0)
-static int scanlog_debug;
static unsigned int ibm_scan_log_dump; /* RTAS token */
static struct proc_dir_entry *proc_ppc64_scan_log_dump; /* The proc file */
@@ -86,14 +84,14 @@ static ssize_t scanlog_read(struct file *file, char __user *buf,
memcpy(data, rtas_data_buf, RTAS_DATA_BUF_SIZE);
spin_unlock(&rtas_data_buf_lock);
- DEBUG("status=%d, data[0]=%x, data[1]=%x, data[2]=%x\n",
- status, data[0], data[1], data[2]);
+ pr_debug("scanlog: status=%d, data[0]=%x, data[1]=%x, " \
+ "data[2]=%x\n", status, data[0], data[1], data[2]);
switch (status) {
case SCANLOG_COMPLETE:
- DEBUG("hit eof\n");
+ pr_debug("scanlog: hit eof\n");
return 0;
case SCANLOG_HWERROR:
- DEBUG("hardware error reading scan log data\n");
+ pr_debug("scanlog: hardware error reading data\n");
return -EIO;
case SCANLOG_CONTINUE:
/* We may or may not have data yet */
@@ -110,7 +108,8 @@ static ssize_t scanlog_read(struct file *file, char __user *buf,
/* Assume extended busy */
wait_time = rtas_busy_delay_time(status);
if (!wait_time) {
- printk(KERN_ERR "scanlog: unknown error from rtas: %d\n", status);
+ printk(KERN_ERR "scanlog: unknown error " \
+ "from rtas: %d\n", status);
return -EIO;
}
}
@@ -134,15 +133,9 @@ static ssize_t scanlog_write(struct file * file, const char __user * buf,
if (buf) {
if (strncmp(stkbuf, "reset", 5) == 0) {
- DEBUG("reset scanlog\n");
+ pr_debug("scanlog: reset scanlog\n");
status = rtas_call(ibm_scan_log_dump, 2, 1, NULL, 0, 0);
- DEBUG("rtas returns %d\n", status);
- } else if (strncmp(stkbuf, "debugon", 7) == 0) {
- printk(KERN_ERR "scanlog: debug on\n");
- scanlog_debug = 1;
- } else if (strncmp(stkbuf, "debugoff", 8) == 0) {
- printk(KERN_ERR "scanlog: debug off\n");
- scanlog_debug = 0;
+ pr_debug("scanlog: rtas returns %d\n", status);
}
}
return count;
diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index f66aa9c..65e8795 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -70,11 +70,6 @@
#include "plpar_wrappers.h"
#include "pseries.h"
-#ifdef DEBUG
-#define DBG(fmt...) udbg_printf(fmt)
-#else
-#define DBG(fmt...)
-#endif
int fwnmi_active; /* TRUE if an FWNMI handler is present */
@@ -326,7 +321,7 @@ static int pseries_set_xdabr(unsigned long dabr)
*/
static void __init pSeries_init_early(void)
{
- DBG(" -> pSeries_init_early()\n");
+ pr_debug(" -> pSeries_init_early()\n");
if (firmware_has_feature(FW_FEATURE_LPAR))
find_udbg_vterm();
@@ -338,7 +333,7 @@ static void __init pSeries_init_early(void)
iommu_init_early_pSeries();
- DBG(" <- pSeries_init_early()\n");
+ pr_debug(" <- pSeries_init_early()\n");
}
/*
@@ -383,7 +378,7 @@ static int __init pSeries_probe(void)
of_flat_dt_is_compatible(root, "IBM,CBEA"))
return 0;
- DBG("pSeries detected, looking for LPAR capability...\n");
+ pr_debug("pSeries detected, looking for LPAR capability...\n");
/* Now try to figure out if we are running on LPAR */
of_scan_flat_dt(pSeries_probe_hypertas, NULL);
@@ -393,8 +388,8 @@ static int __init pSeries_probe(void)
else
hpte_init_native();
- DBG("Machine is%s LPAR !\n",
- (powerpc_firmware_features & FW_FEATURE_LPAR) ? "" : " not");
+ pr_debug("Machine is%s LPAR !\n",
+ (powerpc_firmware_features & FW_FEATURE_LPAR) ? "" : " not");
return 1;
}
diff --git a/arch/powerpc/platforms/pseries/smp.c b/arch/powerpc/platforms/pseries/smp.c
index ea4c659..e9bc2a5 100644
--- a/arch/powerpc/platforms/pseries/smp.c
+++ b/arch/powerpc/platforms/pseries/smp.c
@@ -51,12 +51,6 @@
#include "plpar_wrappers.h"
#include "pseries.h"
-#ifdef DEBUG
-#include <asm/udbg.h>
-#define DBG(fmt...) udbg_printf(fmt)
-#else
-#define DBG(fmt...)
-#endif
/*
* The primary thread of each non-boot processor is recorded here before
@@ -231,7 +225,7 @@ static void __init smp_init_pseries(void)
{
int i;
- DBG(" -> smp_init_pSeries()\n");
+ pr_debug(" -> smp_init_pSeries()\n");
/* Mark threads which are still spinning in hold loops. */
if (cpu_has_feature(CPU_FTR_SMT)) {
@@ -255,7 +249,7 @@ static void __init smp_init_pseries(void)
smp_ops->take_timebase = pSeries_take_timebase;
}
- DBG(" <- smp_init_pSeries()\n");
+ pr_debug(" <- smp_init_pSeries()\n");
}
#ifdef CONFIG_MPIC
--
1.5.5
^ permalink raw reply related
* [PATCH 3/5] Register udbg console early on pseries LPAR
From: Michael Ellerman @ 2008-04-24 5:13 UTC (permalink / raw)
To: linuxppc-dev
In-Reply-To: <c01702547f6296b2c18e5ff7504ad52cb0402664.1209013992.git.michael@ellerman.id.au>
On pseries LPAR we can call the udbg routines, and the udbg console very
early. So mark the udbg console as safe to call early in boot, and register
the udbg console as soon as the udbg routines are hooked up.
This allows platforms/pseries code to use printk() and pr_debug() rather
than needing to call udbg_printf() directly for early debugging. This is
nice because a) it's standard, b) it goes via the printk buffer, and c)
you can get printk time stamps.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
---
arch/powerpc/platforms/pseries/lpar.c | 2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c
index 9235c46..c34789f 100644
--- a/arch/powerpc/platforms/pseries/lpar.c
+++ b/arch/powerpc/platforms/pseries/lpar.c
@@ -196,6 +196,8 @@ void __init udbg_init_debug_lpar(void)
udbg_putc = udbg_putcLP;
udbg_getc = udbg_getcLP;
udbg_getc_poll = udbg_getc_pollLP;
+
+ register_early_udbg_console();
}
/* returns 0 if couldn't find or use /chosen/stdout as console */
--
1.5.5
^ permalink raw reply related
* [PATCH 2/5] Mark udbg console as CON_ANYTIME, ie. callable early in boot
From: Michael Ellerman @ 2008-04-24 5:13 UTC (permalink / raw)
To: linuxppc-dev
In-Reply-To: <c01702547f6296b2c18e5ff7504ad52cb0402664.1209013992.git.michael@ellerman.id.au>
The udbg console should be safe to call basically at any time after boot.
It does not need any per-cpu resources or for the cpu to be online, as
long as there is a udbg_putc routine hooked up it should work. So mark it
as CON_ANYTIME.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
---
arch/powerpc/kernel/udbg.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/arch/powerpc/kernel/udbg.c b/arch/powerpc/kernel/udbg.c
index 9ac5f3a..7d6c9bb 100644
--- a/arch/powerpc/kernel/udbg.c
+++ b/arch/powerpc/kernel/udbg.c
@@ -154,7 +154,7 @@ static void udbg_console_write(struct console *con, const char *s,
static struct console udbg_console = {
.name = "udbg",
.write = udbg_console_write,
- .flags = CON_PRINTBUFFER | CON_ENABLED | CON_BOOT,
+ .flags = CON_PRINTBUFFER | CON_ENABLED | CON_BOOT | CON_ANYTIME,
.index = 0,
};
--
1.5.5
^ permalink raw reply related
* [PATCH 1/5] Set udbg_console index to 0
From: Michael Ellerman @ 2008-04-24 5:13 UTC (permalink / raw)
To: linuxppc-dev
Because the udbg_console has CON_ENABLED set, it's possible that when we
register it with the console code the index won't be set. This leads to
slightly confusing boot messages like:
[ 0.000000] console [udbg-1] enabled
We could remove CON_ENABLED, but we don't want to do that, we always
want the udbg console to be activated, even if the user specified some
other console on the command line.
The simplest fix seems to be just to set the index to 0 by hand. There
is no issue with duplicate udbg consoles, as we guard against registering
multiple times in register_early_udbg_console().
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
---
arch/powerpc/kernel/udbg.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/arch/powerpc/kernel/udbg.c b/arch/powerpc/kernel/udbg.c
index 7aad620..9ac5f3a 100644
--- a/arch/powerpc/kernel/udbg.c
+++ b/arch/powerpc/kernel/udbg.c
@@ -155,7 +155,7 @@ static struct console udbg_console = {
.name = "udbg",
.write = udbg_console_write,
.flags = CON_PRINTBUFFER | CON_ENABLED | CON_BOOT,
- .index = -1,
+ .index = 0,
};
static int early_console_initialized;
--
1.5.5
^ permalink raw reply related
* Re: [U-Boot-Users] AMCC PPC440EPx/sequoia stability question...
From: Stefan Roese @ 2008-04-24 4:36 UTC (permalink / raw)
To: jwboyer; +Cc: u-boot-users, linuxppc-embedded
In-Reply-To: <1208999427.2946.8.camel@vader.jdub.homelinux.org>
On Thursday 24 April 2008, Josh Boyer wrote:
> > Please note that I recently introduced a CFG_MEM_TOP_HIDE option for the
> > 440EPx CHIP 11 errata. I suggest you take a look at this too and see if
> > this changes your behavior.
>
> Explain this a bit more please? Is a kernel change needed here?
This depends. When the bootwrapper version is used then yes, the kernel should
get changed. This is because the bootwrapper detects the SDRAM size from the
DDR2 controller and passes it to Linux.
Without bootwrapper no changes are needed, since U-Boot already passes the
corrected memory size to Linux (totalsize-4k currently).
Best regards,
Stefan
=====================================================================
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de
=====================================================================
^ permalink raw reply
* Re: [PATCH 2/2] Raise the upper limit of NR_CPUS and move the pacas into the BSS.
From: Michael Ellerman @ 2008-04-24 4:08 UTC (permalink / raw)
To: Tony Breeds; +Cc: linuxppc-dev, Paul Mackerras
In-Reply-To: <6bbf4d3637c9b72f9f0a4cf8a6f2c322c0784d7e.1209008469.git.tony@bakeyournoodle.com>
[-- Attachment #1: Type: text/plain, Size: 631 bytes --]
On Thu, 2008-04-24 at 13:43 +1000, Tony Breeds wrote:
> This patch adds the required functionality to fill in all pacas at runtime.
>
> With NR_CPUS=1024
> text data bss dec hex filename
> 137 1704032 0 1704169 1a00e9 arch/powerpc/kernel/paca.o :Before
> 121 1179744 524288 1704153 1a00d9 arch/powerpc/kernel/paca.o :After
Niiiiice.
cheers
--
Michael Ellerman
OzLabs, IBM Australia Development Lab
wwweb: http://michael.ellerman.id.au
phone: +61 2 6212 1183 (tie line 70 21183)
We do not inherit the earth from our ancestors,
we borrow it from our children. - S.M.A.R.T Person
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]
^ permalink raw reply
* [PATCH 2/2] Raise the upper limit of NR_CPUS and move the pacas into the BSS.
From: Tony Breeds @ 2008-04-24 3:43 UTC (permalink / raw)
To: Paul Mackerras, linuxppc-dev
In-Reply-To: <b863aa7fba7b2b00cb72d327fdd7a04b6cb81106.1209008469.git.tony@bakeyournoodle.com>
This patch adds the required functionality to fill in all pacas at runtime.
With NR_CPUS=1024
text data bss dec hex filename
137 1704032 0 1704169 1a00e9 arch/powerpc/kernel/paca.o :Before
121 1179744 524288 1704153 1a00d9 arch/powerpc/kernel/paca.o :After
Also remove unneeded #includes from arch/powerpc/kernel/paca.c
Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
---
Now merged, with patch 1 fomr previous series.
arch/powerpc/kernel/paca.c | 87 +++++++++----------------------
arch/powerpc/kernel/setup_64.c | 3 +
arch/powerpc/platforms/Kconfig.cputype | 4 +-
include/asm-powerpc/paca.h | 1 +
4 files changed, 31 insertions(+), 64 deletions(-)
diff --git a/arch/powerpc/kernel/paca.c b/arch/powerpc/kernel/paca.c
index ac163bd..c9bf17e 100644
--- a/arch/powerpc/kernel/paca.c
+++ b/arch/powerpc/kernel/paca.c
@@ -7,17 +7,11 @@
* 2 of the License, or (at your option) any later version.
*/
-#include <linux/types.h>
#include <linux/threads.h>
#include <linux/module.h>
-#include <asm/processor.h>
-#include <asm/ptrace.h>
-#include <asm/page.h>
#include <asm/lppaca.h>
#include <asm/paca.h>
-#include <asm/mmu.h>
-
/* This symbol is provided by the linker - let it fill in the paca
* field correctly */
@@ -65,60 +59,29 @@ struct slb_shadow slb_shadow[] __cacheline_aligned = {
* processors. The processor VPD array needs one entry per physical
* processor (not thread).
*/
-#define PACA_INIT(number) \
-{ \
- .lppaca_ptr = &lppaca[number], \
- .lock_token = 0x8000, \
- .paca_index = (number), /* Paca Index */ \
- .kernel_toc = (unsigned long)(&__toc_start) + 0x8000UL, \
- .hw_cpu_id = 0xffff, \
- .slb_shadow_ptr = &slb_shadow[number], \
- .__current = &init_task, \
-}
-
-struct paca_struct paca[] = {
- PACA_INIT(0),
-#if NR_CPUS > 1
- PACA_INIT( 1), PACA_INIT( 2), PACA_INIT( 3),
-#if NR_CPUS > 4
- PACA_INIT( 4), PACA_INIT( 5), PACA_INIT( 6), PACA_INIT( 7),
-#if NR_CPUS > 8
- PACA_INIT( 8), PACA_INIT( 9), PACA_INIT( 10), PACA_INIT( 11),
- PACA_INIT( 12), PACA_INIT( 13), PACA_INIT( 14), PACA_INIT( 15),
- PACA_INIT( 16), PACA_INIT( 17), PACA_INIT( 18), PACA_INIT( 19),
- PACA_INIT( 20), PACA_INIT( 21), PACA_INIT( 22), PACA_INIT( 23),
- PACA_INIT( 24), PACA_INIT( 25), PACA_INIT( 26), PACA_INIT( 27),
- PACA_INIT( 28), PACA_INIT( 29), PACA_INIT( 30), PACA_INIT( 31),
-#if NR_CPUS > 32
- PACA_INIT( 32), PACA_INIT( 33), PACA_INIT( 34), PACA_INIT( 35),
- PACA_INIT( 36), PACA_INIT( 37), PACA_INIT( 38), PACA_INIT( 39),
- PACA_INIT( 40), PACA_INIT( 41), PACA_INIT( 42), PACA_INIT( 43),
- PACA_INIT( 44), PACA_INIT( 45), PACA_INIT( 46), PACA_INIT( 47),
- PACA_INIT( 48), PACA_INIT( 49), PACA_INIT( 50), PACA_INIT( 51),
- PACA_INIT( 52), PACA_INIT( 53), PACA_INIT( 54), PACA_INIT( 55),
- PACA_INIT( 56), PACA_INIT( 57), PACA_INIT( 58), PACA_INIT( 59),
- PACA_INIT( 60), PACA_INIT( 61), PACA_INIT( 62), PACA_INIT( 63),
-#if NR_CPUS > 64
- PACA_INIT( 64), PACA_INIT( 65), PACA_INIT( 66), PACA_INIT( 67),
- PACA_INIT( 68), PACA_INIT( 69), PACA_INIT( 70), PACA_INIT( 71),
- PACA_INIT( 72), PACA_INIT( 73), PACA_INIT( 74), PACA_INIT( 75),
- PACA_INIT( 76), PACA_INIT( 77), PACA_INIT( 78), PACA_INIT( 79),
- PACA_INIT( 80), PACA_INIT( 81), PACA_INIT( 82), PACA_INIT( 83),
- PACA_INIT( 84), PACA_INIT( 85), PACA_INIT( 86), PACA_INIT( 87),
- PACA_INIT( 88), PACA_INIT( 89), PACA_INIT( 90), PACA_INIT( 91),
- PACA_INIT( 92), PACA_INIT( 93), PACA_INIT( 94), PACA_INIT( 95),
- PACA_INIT( 96), PACA_INIT( 97), PACA_INIT( 98), PACA_INIT( 99),
- PACA_INIT(100), PACA_INIT(101), PACA_INIT(102), PACA_INIT(103),
- PACA_INIT(104), PACA_INIT(105), PACA_INIT(106), PACA_INIT(107),
- PACA_INIT(108), PACA_INIT(109), PACA_INIT(110), PACA_INIT(111),
- PACA_INIT(112), PACA_INIT(113), PACA_INIT(114), PACA_INIT(115),
- PACA_INIT(116), PACA_INIT(117), PACA_INIT(118), PACA_INIT(119),
- PACA_INIT(120), PACA_INIT(121), PACA_INIT(122), PACA_INIT(123),
- PACA_INIT(124), PACA_INIT(125), PACA_INIT(126), PACA_INIT(127),
-#endif
-#endif
-#endif
-#endif
-#endif
-};
+struct paca_struct paca[NR_CPUS];
EXPORT_SYMBOL(paca);
+
+void __init initialise_pacas(void)
+{
+ int cpu;
+
+ /* The TOC register (GPR2) points 32kB into the TOC, so that 64kB
+ * of the TOC can be addressed using a single machine instruction.
+ */
+ unsigned long kernel_toc = (unsigned long)(&__toc_start) + 0x8000UL;
+
+ /* Can't use for_each_*_cpu, as they aren't functional yet */
+ for (cpu = 0; cpu < NR_CPUS; cpu++) {
+ struct paca_struct *new_paca = &paca[cpu];
+
+ new_paca->lppaca_ptr = &lppaca[cpu];
+ new_paca->lock_token = 0x8000;
+ new_paca->paca_index = cpu;
+ new_paca->kernel_toc = kernel_toc;
+ new_paca->hw_cpu_id = 0xffff;
+ new_paca->slb_shadow_ptr = &slb_shadow[cpu];
+ new_paca->__current = &init_task;
+
+ }
+}
diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index 31ada9f..5e382ac 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -170,6 +170,9 @@ void __init setup_paca(int cpu)
void __init early_setup(unsigned long dt_ptr)
{
+ /* Fill in any unititialised pacas */
+ initialise_pacas();
+
/* Identify CPU type */
identify_cpu(0, mfspr(SPRN_PVR));
diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
index 5fc7fac..f7efaa9 100644
--- a/arch/powerpc/platforms/Kconfig.cputype
+++ b/arch/powerpc/platforms/Kconfig.cputype
@@ -220,8 +220,8 @@ config SMP
If you don't know what to do here, say N.
config NR_CPUS
- int "Maximum number of CPUs (2-128)"
- range 2 128
+ int "Maximum number of CPUs (2-1024)"
+ range 2 1024
depends on SMP
default "32" if PPC64
default "4"
diff --git a/include/asm-powerpc/paca.h b/include/asm-powerpc/paca.h
index eb61b9c..7b56444 100644
--- a/include/asm-powerpc/paca.h
+++ b/include/asm-powerpc/paca.h
@@ -108,6 +108,7 @@ struct paca_struct {
};
extern struct paca_struct paca[];
+extern void initialise_pacas(void);
#endif /* __KERNEL__ */
#endif /* _ASM_POWERPC_PACA_H */
--
1.5.5.1
^ permalink raw reply related
* [PATCH 1/2 v4] Make iSeries spin on __secondary_hold_spinloop, like pSeries.
From: Tony Breeds @ 2008-04-24 3:43 UTC (permalink / raw)
To: Paul Mackerras, linuxppc-dev; +Cc: Stephen Rothwell
Currently all iSeries secondary CPU's spin directly on the cpu_start in thier
paca. Make them spin on the global __secondary_hold_spinloop, until after the
pacas have been initialised.
As sfr points out this works because __secondary_hold_spinloop is being set
already, but iSeries isn't looking at it :)
Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
Acked-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
Changes since v1:
- Updated commit message.
Changes since v2:
- Fixed my brain damage, and split out the code that should be in 3/3
Changes since v3:
- moved in patch series, no funcational changes.
arch/powerpc/platforms/iseries/exception.S | 27 ++++++++++++++++++++++++++-
1 files changed, 26 insertions(+), 1 deletions(-)
diff --git a/arch/powerpc/platforms/iseries/exception.S b/arch/powerpc/platforms/iseries/exception.S
index c775cd4..8ff330d 100644
--- a/arch/powerpc/platforms/iseries/exception.S
+++ b/arch/powerpc/platforms/iseries/exception.S
@@ -59,8 +59,33 @@ system_reset_iSeries:
andc r4,r4,r5
mtspr SPRN_CTRLT,r4
+/* Spin on __secondary_hold_spinloop until it is updated by the boot cpu. */
+/* In the UP case we'll yeild() later, and we will not access the paca anyway */
+#ifdef CONFIG_SMP
1:
HMT_LOW
+ LOAD_REG_IMMEDIATE(r23, __secondary_hold_spinloop)
+ ld r23,0(r23)
+ sync
+ LOAD_REG_IMMEDIATE(r3,current_set)
+ sldi r28,r24,3 /* get current_set[cpu#] */
+ ldx r3,r3,r28
+ addi r1,r3,THREAD_SIZE
+ subi r1,r1,STACK_FRAME_OVERHEAD
+
+ cmpwi 0,r23,0 /* Keep poking the Hypervisor until */
+ bne 2f /* we're released */
+ /* Let the Hypervisor know we are alive */
+ /* 8002 is a call to HvCallCfg::getLps, a harmless Hypervisor function */
+ lis r3,0x8002
+ rldicr r3,r3,32,15 /* r0 = (r3 << 32) & 0xffff000000000000 */
+ li r0,-1 /* r0=-1 indicates a Hypervisor call */
+ sc /* Invoke the hypervisor via a system call */
+ b 1b
+#endif
+
+2:
+ HMT_LOW
#ifdef CONFIG_SMP
lbz r23,PACAPROCSTART(r13) /* Test if this processor
* should start */
@@ -91,7 +116,7 @@ iSeries_secondary_smp_loop:
li r0,-1 /* r0=-1 indicates a Hypervisor call */
sc /* Invoke the hypervisor via a system call */
mfspr r13,SPRN_SPRG3 /* Put r13 back ???? */
- b 1b /* If SMP not configured, secondaries
+ b 2b /* If SMP not configured, secondaries
* loop forever */
/*** ISeries-LPAR interrupt handlers ***/
--
1.5.5.1
^ permalink raw reply related
* Re: [PATCH] Discourage people from fiddling with kernel data from prom_init
From: Benjamin Herrenschmidt @ 2008-04-24 3:37 UTC (permalink / raw)
To: Michael Ellerman; +Cc: linuxppc-dev
In-Reply-To: <d3f56e3d99619e642b6f8cd73b35dad4597934ed.1209002897.git.michael@ellerman.id.au>
On Thu, 2008-04-24 at 12:08 +1000, Michael Ellerman wrote:
> As BenH said the other day, it is an "accident" that prom_init.o is linked
> with the rest of the kernel. The truth is a little more subtle, prom_init
> isn't truly bootloader, it does fiddle with kernel data in a few places.
>
> What we can do is discourage people from adding new code that accesses
> data outside of prom_init. And hence this patch, from the script:
>
> # This script checks prom_init.o to see what external symbols it
> # is using, if it finds symbols not in the whitelist it returns
> # an error. The point of this is to discourage people from
> # intentionally or accidentally adding new code to prom_init.c
> # which has side effects on other parts of the kernel.
>
> Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> ---
> arch/powerpc/kernel/Makefile | 9 +++++
> arch/powerpc/kernel/prom_init_check.sh | 58 ++++++++++++++++++++++++++++++++
> 2 files changed, 67 insertions(+), 0 deletions(-)
>
> diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
> index 5183a90..562bb02 100644
> --- a/arch/powerpc/kernel/Makefile
> +++ b/arch/powerpc/kernel/Makefile
> @@ -106,4 +106,13 @@ PHONY += systbl_chk
> systbl_chk: $(src)/systbl_chk.sh $(obj)/systbl_chk.i
> $(call cmd,systbl_chk)
>
> +$(obj)/built-in.o: prom_init_check
> +
> +quiet_cmd_prom_init_check = CALL $<
> + cmd_prom_init_check = $(CONFIG_SHELL) $< "$(NM)" "$(obj)/prom_init.o"
> +
> +PHONY += prom_init_check
> +prom_init_check: $(src)/prom_init_check.sh $(obj)/prom_init.o
> + $(call cmd,prom_init_check)
> +
> clean-files := vmlinux.lds
> diff --git a/arch/powerpc/kernel/prom_init_check.sh b/arch/powerpc/kernel/prom_init_check.sh
> new file mode 100644
> index 0000000..8e24fc1
> --- /dev/null
> +++ b/arch/powerpc/kernel/prom_init_check.sh
> @@ -0,0 +1,58 @@
> +#!/bin/sh
> +#
> +# Copyright © 2008 IBM Corporation
> +#
> +# This program is free software; you can redistribute it and/or
> +# modify it under the terms of the GNU General Public License
> +# as published by the Free Software Foundation; either version
> +# 2 of the License, or (at your option) any later version.
> +
> +# This script checks prom_init.o to see what external symbols it
> +# is using, if it finds symbols not in the whitelist it returns
> +# an error. The point of this is to discourage people from
> +# intentionally or accidentally adding new code to prom_init.c
> +# which has side effects on other parts of the kernel.
> +
> +# If you really need to reference something from prom_init.o add
> +# it to the list below:
> +
> +WHITELIST="add_reloc_offset __bss_start __bss_stop copy_and_flush
> +_end enter_prom memcpy memset reloc_offset __secondary_hold
> +__secondary_hold_acknowledge __secondary_hold_spinloop __start
> +strcmp strcpy strlcpy strlen strncmp strstr logo_linux_clut224
> +reloc_got2"
> +
> +NM="$1"
> +OBJ="$2"
> +
> +ERROR=0
> +
> +for UNDEF in $($NM -u $OBJ | awk '{print $2}')
> +do
> + # On 64-bit nm gives us the function descriptors, which have
> + # a leading . on the name, so strip it off here.
> + UNDEF="${UNDEF#.}"
> +
> + if [ $KBUILD_VERBOSE ]; then
> + if [ $KBUILD_VERBOSE -ne 0 ]; then
> + echo "Checking prom_init.o symbol '$UNDEF'"
> + fi
> + fi
> +
> + OK=0
> + for WHITE in $WHITELIST
> + do
> + if [ "$UNDEF" = "$WHITE" ]; then
> + OK=1
> + break
> + fi
> + done
> +
> + if [ $OK -eq 0 ]; then
> + ERROR=1
> + echo "Error: External symbol '$UNDEF' referenced" \
> + "from prom_init.c" >&2
> + fi
> +done
> +
> +exit $ERROR
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-dev
^ permalink raw reply
* Re: [PATCH 2/11] cell: generalize io-workarounds code
From: Benjamin Herrenschmidt @ 2008-04-24 3:30 UTC (permalink / raw)
To: Ishizaki Kou; +Cc: linuxppc-dev, paulus
In-Reply-To: <20080424.120722.-1300531374.kouish@swc.toshiba.co.jp>
On Thu, 2008-04-24 at 12:07 +0900, Ishizaki Kou wrote:
>
> I'm sorry to have kept you waiting for my response.
>
> I just finished reviewing your patch. Your patch works well on
> Celleb, and I found I also should do the same thing for Celleb as you
> pointed. I will send a new patch which includes your fix.
Thanks. Please do so ASAP as tomorrow is non-working day and we are
getting close to -rc1.
Cheers,
Ben.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox