mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: steiner@sgi.com
To: akpm@osdl.org, linux-kernel@vger.kernel.org
Cc: mingo@elte.hu, tglx@linutronix.de, holt@sgi.com, andrea@qumranet.com
Subject: [patch 06/11] GRU Driver - page faults & exceptions
Date: Mon, 09 Jun 2008 16:10:34 -0500	[thread overview]
Message-ID: <20080609211046.208712459@attica.americas.sgi.com> (raw)
In-Reply-To: <20080609211028.110089743@attica.americas.sgi.com>

[-- Attachment #1: gru_fault --]
[-- Type: text/plain, Size: 17900 bytes --]

This file contains the functions that manage GRU page faults and exceptions.


Signed-off-by: Jack Steiner <steiner@sgi.com>

---
 drivers/misc/gru/grufault.c |  628 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 628 insertions(+)

Index: linux/drivers/misc/gru/grufault.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ linux/drivers/misc/gru/grufault.c	2008-06-09 13:53:05.844783993 -0500
@@ -0,0 +1,628 @@
+/*
+ * SN Platform GRU Driver
+ *
+ *              FAULT HANDLER FOR GRU DETECTED TLB MISSES
+ *
+ * This file contains code that handles TLB misses within the GRU.
+ * These misses are reported either via interrupts or user polling of
+ * the user CB.
+ *
+ *  Copyright (c) 2008 Silicon Graphics, Inc.  All Rights Reserved.
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ */
+
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/spinlock.h>
+#include <linux/mm.h>
+#include <linux/hugetlb.h>
+#include <linux/device.h>
+#include <linux/io.h>
+#include <linux/uaccess.h>
+#include <asm/pgtable.h>
+#include "gru.h"
+#include "grutables.h"
+#include "grulib.h"
+#include "gru_instructions.h"
+#include <asm/uv/uv_hub.h>
+
+/*
+ * Test if a physical address is a valid GRU GSEG address
+ */
+static inline int is_gru_paddr(unsigned long paddr)
+{
+	return paddr >= gru_start_paddr && paddr < gru_end_paddr;
+}
+
+/*
+ * Find the vma of a GRU segment. Caller must hold mmap_sem.
+ */
+struct vm_area_struct *gru_find_vma(unsigned long vaddr)
+{
+	struct vm_area_struct *vma;
+
+	vma = find_vma(current->mm, vaddr);
+	if (vma && vma->vm_start <= vaddr && vma->vm_ops == &gru_vm_ops)
+		return vma;
+	return NULL;
+}
+
+/*
+ * Find and lock the gts that contains the specified user vaddr.
+ *
+ * Returns:
+ * 	- *gts with the mmap_sem locked for read and the GTS locked.
+ *	- NULL if vaddr invalid OR is not a valid GSEG vaddr.
+ */
+
+static struct gru_thread_state *gru_find_lock_gts(unsigned long vaddr)
+{
+	struct mm_struct *mm = current->mm;
+	struct vm_area_struct *vma;
+	struct gru_thread_state *gts = NULL;
+
+	down_read(&mm->mmap_sem);
+	vma = gru_find_vma(vaddr);
+	if (vma)
+		gts = gru_find_thread_state(vma, TSID(vaddr, vma));
+	if (gts)
+		mutex_lock(&gts->ts_ctxlock);
+	else
+		up_read(&mm->mmap_sem);
+	return gts;
+}
+
+static struct gru_thread_state *gru_alloc_locked_gts(unsigned long vaddr)
+{
+	struct mm_struct *mm = current->mm;
+	struct vm_area_struct *vma;
+	struct gru_thread_state *gts = NULL;
+
+	down_write(&mm->mmap_sem);
+	vma = gru_find_vma(vaddr);
+	if (vma)
+		gts = gru_alloc_thread_state(vma, TSID(vaddr, vma));
+	if (gts) {
+		mutex_lock(&gts->ts_ctxlock);
+		downgrade_write(&mm->mmap_sem);
+	} else {
+		up_write(&mm->mmap_sem);
+	}
+
+	return gts;
+}
+
+/*
+ * Unlock a GTS that was previously locked with gru_find_lock_gts().
+ */
+static void gru_unlock_gts(struct gru_thread_state *gts)
+{
+	mutex_unlock(&gts->ts_ctxlock);
+	up_read(&current->mm->mmap_sem);
+}
+
+/*
+ * Set a CB.istatus to active using a user virtual address. This must be done
+ * just prior to a TFH RESTART. The new cb.istatus is an in-cache status ONLY.
+ * If the line is evicted, the status may be lost. The in-cache update
+ * is necessary to prevent the user from seeing a stale cb.istatus that will
+ * change as soon as the TFH restart is complete. Races may cause an
+ * occasional failure to clear the cb.istatus, but that is ok.
+ *
+ * If the cb address is not valid (should not happen, but...), nothing
+ * bad will happen.. The get_user()/put_user() will fail but there
+ * are no bad side-effects.
+ */
+static void gru_cb_set_istatus_active(unsigned long __user *cb)
+{
+	union {
+		struct gru_instruction_bits bits;
+		unsigned long dw;
+	} u;
+
+	if (cb) {
+		get_user(u.dw, cb);
+		u.bits.istatus = CBS_ACTIVE;
+		put_user(u.dw, cb);
+	}
+}
+
+/*
+ * Convert a interrupt IRQ to a pointer to the GRU GTS that caused the
+ * interrupt. Interrupts are always sent to a cpu on the blade that contains the
+ * GRU (except for headless blades which are not currently supported). A blade
+ * has N grus; a block of N consecutive IRQs is assigned to the GRUs. The IRQ
+ * number uniquely identifies the GRU chiplet on the local blade that caused the
+ * interrupt. Always called in interrupt context.
+ */
+static inline struct gru_state *irq_to_gru(int irq)
+{
+	return &gru_base[uv_numa_blade_id()]->bs_grus[irq - IRQ_GRU];
+}
+
+/*
+ * Read & clear a TFM
+ *
+ * The GRU has an array of fault maps. A map is private to a cpu
+ * Only one cpu will be accessing a cpu's fault map.
+ *
+ * This function scans the cpu-private fault map & clears all bits that
+ * are set. The function returns a bitmap that indicates the bits that
+ * were cleared. Note that sense the maps may be updated asynchronously by
+ * the GRU, atomic operations must be used to clear bits.
+ */
+static void get_clear_fault_map(struct gru_state *gru,
+				struct gru_tlb_fault_map *map)
+{
+	unsigned long i, k;
+	struct gru_tlb_fault_map *tfm;
+
+	tfm = get_tfm_for_cpu(gru, gru_cpu_fault_map_id());
+	prefetchw(tfm);		/* Helps on hardware, required for emulator */
+	for (i = 0; i < BITS_TO_LONGS(GRU_NUM_CBE); i++) {
+		k = tfm->fault_bits[i];
+		if (k)
+			k = xchg(&tfm->fault_bits[i], 0UL);
+		map->fault_bits[i] = k;
+	}
+
+	/*
+	 * Not functionally required but helps performance. (Required
+	 * on emulator)
+	 */
+	gru_flush_cache(tfm);
+}
+
+/*
+ * Atomic (interrupt context) & non-atomic (user context) functions to
+ * convert a vaddr into a physical address. The sise of the page
+ * is returned in pageshift.
+ * 	returns:
+ * 		  0 - successful
+ * 		< 0 - error code
+ * 		  1 - (atomic only) try again in non-atomic context
+ */
+static int non_atomic_pte_lookup(struct vm_area_struct *vma,
+				 unsigned long vaddr, int write,
+				 unsigned long *paddr, int *pageshift)
+{
+	struct page *page;
+
+	if (get_user_pages
+	    (current, current->mm, vaddr, 1, write, 1, &page, NULL) <= 0)
+		return -EFAULT;
+	*paddr = page_to_phys(page);
+	*pageshift = is_vm_hugetlb_page(vma) ? HPAGE_SHIFT : PAGE_SHIFT;
+	put_page(page);
+	return 0;
+}
+
+static int atomic_pte_lookup(struct vm_area_struct *vma, unsigned long vaddr,
+			     int write, unsigned long *paddr, int *pageshift)
+{
+	struct page *page;
+
+	page = follow_page(vma, vaddr, (write ? FOLL_WRITE : 0));
+	if (!page)
+		return 1;
+	*paddr = page_to_phys(page);
+	*pageshift = is_vm_hugetlb_page(vma) ? HPAGE_SHIFT : PAGE_SHIFT;
+	return 0;
+}
+
+/*
+ * Drop a TLB entry into the GRU. The fault is described by info in an TFH.
+ *	Input:
+ *		cb    Address of user CBR. Null if not running in user context
+ * 	Return:
+ * 		  0 = dropin, exception, or switch to UPM successful
+ * 		  1 = range invalidate active
+ * 		< 0 = error code
+ *
+ */
+static int gru_try_dropin(struct gru_thread_state *gts,
+			  struct gru_tlb_fault_handle *tfh,
+			  unsigned long __user *cb)
+{
+	struct mm_struct *mm = gts->ts_mm;
+	struct vm_area_struct *vma;
+	int pageshift, asid, write, ret;
+	unsigned long paddr, gpa, vaddr;
+
+	/*
+	 * NOTE: The GRU contains magic hardware that eliminates races between
+	 * TLB invalidates and TLB dropins. If an invalidate occurs
+	 * in the window between reading the TFH and the subsequent TLB dropin,
+	 * the dropin is ignored. This eliminates the need for additional locks.
+	 */
+
+	/*
+	 * Error if TFH state is IDLE or FMM mode & the user issuing a UPM call.
+	 * Might be a hardware race OR a stupid user. Ignore FMM because FMM
+	 * is a transient state.
+	 */
+	if (tfh->state == TFHSTATE_IDLE)
+		goto failidle;
+	if (tfh->state == TFHSTATE_MISS_FMM && cb)
+		goto failfmm;
+
+	write = (tfh->cause & TFHCAUSE_TLB_MOD) != 0;
+	vaddr = tfh->missvaddr;
+	asid = tfh->missasid;
+	if (asid == 0)
+		goto failnoasid;
+
+	rmb();	/* TFH must be cache resident before reading ms_range_active */
+
+	/*
+	 * TFH is cache resident - at least briefly. Fail the dropin
+	 * if a range invalidate is active.
+	 */
+	if (atomic_read(&gts->ts_gms->ms_range_active))
+		goto failactive;
+
+	vma = find_vma(mm, vaddr);
+	if (!vma)
+		goto failinval;
+
+	/*
+	 * Atomic lookup is faster & usually works even if called in non-atomic
+	 * context.
+	 */
+	ret = atomic_pte_lookup(vma, vaddr, write, &paddr, &pageshift);
+	if (ret) {
+		if (!cb)
+			goto failupm;
+		if (non_atomic_pte_lookup(vma, vaddr, write, &paddr,
+					  &pageshift))
+			goto failinval;
+	}
+	if (is_gru_paddr(paddr))
+		goto failinval;
+
+	paddr = paddr & ~((1UL << pageshift) - 1);
+	gpa = uv_soc_phys_ram_to_gpa(paddr);
+	gru_cb_set_istatus_active(cb);
+	tfh_write_restart(tfh, gpa, GAA_RAM, vaddr, asid, write,
+			  GRU_PAGESIZE(pageshift));
+	STAT(tlb_dropin);
+	gru_dbg(grudev,
+		"%s: tfh 0x%p, vaddr 0x%lx, asid 0x%x, ps %d, gpa 0x%lx\n",
+		ret ? "non-atomic" : "atomic", tfh, vaddr, asid,
+		pageshift, gpa);
+	return 0;
+
+failnoasid:
+	/* No asid (delayed unload). */
+	STAT(tlb_dropin_fail_no_asid);
+	gru_dbg(grudev, "FAILED no_asid tfh: 0x%p, vaddr 0x%lx\n", tfh, vaddr);
+	if (!cb)
+		tfh_user_polling_mode(tfh);
+	else
+		gru_flush_cache(tfh);
+	return -EAGAIN;
+
+failupm:
+	/* Atomic failure switch CBR to UPM */
+	tfh_user_polling_mode(tfh);
+	STAT(tlb_dropin_fail_upm);
+	gru_dbg(grudev, "FAILED upm tfh: 0x%p, vaddr 0x%lx\n", tfh, vaddr);
+	return 1;
+
+failfmm:
+	/* FMM state on UPM call */
+	STAT(tlb_dropin_fail_fmm);
+	gru_dbg(grudev, "FAILED fmm tfh: 0x%p, state %d\n", tfh, tfh->state);
+	return 0;
+
+failidle:
+	/* TFH was idle  - no miss pending */
+	gru_flush_cache(tfh);
+	if (cb)
+		gru_flush_cache(cb);
+	STAT(tlb_dropin_fail_idle);
+	gru_dbg(grudev, "FAILED idle tfh: 0x%p, state %d\n", tfh, tfh->state);
+	return 0;
+
+failinval:
+	/* All errors (atomic & non-atomic) switch CBR to EXCEPTION state */
+	tfh_exception(tfh);
+	STAT(tlb_dropin_fail_invalid);
+	gru_dbg(grudev, "FAILED inval tfh: 0x%p, vaddr 0x%lx\n", tfh, vaddr);
+	return -EFAULT;
+
+failactive:
+	/* Range invalidate active. Switch to UPM iff atomic */
+	if (!cb)
+		tfh_user_polling_mode(tfh);
+	else
+		gru_flush_cache(tfh);
+	STAT(tlb_dropin_fail_range_active);
+	gru_dbg(grudev, "FAILED range active: tfh 0x%p, vaddr 0x%lx\n",
+		tfh, vaddr);
+	return 1;
+}
+
+/*
+ * Process an external interrupt from the GRU. This interrupt is
+ * caused by a TLB miss.
+ * Note that this is the interrupt handler that is registered with linux
+ * interrupt handlers.
+ */
+irqreturn_t gru_intr(int irq, void *dev_id)
+{
+	struct gru_state *gru;
+	struct gru_tlb_fault_map map;
+	struct gru_thread_state *gts;
+	struct gru_tlb_fault_handle *tfh = NULL;
+	int cbrnum, ctxnum;
+
+	STAT(intr);
+
+	gru = irq_to_gru(irq);
+	if (!gru) {
+		dev_err(grudev, "GRU: invalid interrupt: cpu %d, irq %d\n",
+			raw_smp_processor_id(), irq);
+		return IRQ_NONE;
+	}
+	get_clear_fault_map(gru, &map);
+	gru_dbg(grudev, "irq %d, gru %x, map 0x%lx\n", irq, gru->gs_gid,
+		map.fault_bits[0]);
+
+	for_each_cbr_in_tfm(cbrnum, map.fault_bits) {
+		tfh = get_tfh_by_index(gru, cbrnum);
+		prefetchw(tfh);	/* Helps on hdw, required for emulator */
+
+		/*
+		 * When hardware sets a bit in the faultmap, it implicitly
+		 * locks the GRU context so that it cannot be unloaded.
+		 * The gts cannot change until a TFH start/writestart command
+		 * is issued.
+		 */
+		ctxnum = tfh->ctxnum;
+		gts = gru->gs_gts[ctxnum];
+
+		/*
+		 * This is running in interrupt context. Trylock the mmap_sem.
+		 * If it fails, retry the fault in user context.
+		 */
+		if (down_read_trylock(&gts->ts_mm->mmap_sem)) {
+			gru_try_dropin(gts, tfh, NULL);
+			up_read(&gts->ts_mm->mmap_sem);
+		} else {
+			tfh_user_polling_mode(tfh);
+		}
+	}
+	return IRQ_HANDLED;
+}
+
+#ifdef ZZZ
+/*
+ * UPM call but nothing found in TFH. It _could_ be a race that was lost,
+ * a user bug, or a hardware bug. Try to determine which.
+ */
+static int gru_check_for_bug(unsigned long arg,
+			     struct gru_tlb_fault_handle *tfh)
+{
+	struct gru_instruction_bits ins, *cb = (void *)arg;
+
+	STAT(call_os_check_for_bug);
+	gru_dbg(grudev, "cb %p\n", cb);
+	if (copy_from_user(&ins, cb, sizeof(ins)))
+		return -EFAULT;
+	if (cb->istatus != CBS_CALL_OS)
+		return 0;
+	barrier();
+	gru_flush_cache(cb);
+	if (copy_from_user(&ins, cb, sizeof(ins)))
+		return -EFAULT;
+	if (cb->istatus != CBS_CALL_OS) {
+		dev_info(grudev, "cb %p: Possible coherency bug\n", cb);
+		return 0;
+	}
+
+	gru_flush_cache(tfh);
+	barrier();
+
+	if (tfh->state == TFHSTATE_MISS_UPM) {
+		dev_info(grudev, "tfh %p: Possible coherency bug\n", cb);
+		return -EAGAIN;
+	}
+	gru_dbg(grudev, "cb %p: CB in UPM state but no TFH fault\n", cb);
+	return -EIO;
+
+}
+#endif
+
+static int gru_user_dropin(struct gru_thread_state *gts,
+			   struct gru_tlb_fault_handle *tfh,
+			   unsigned long __user *cb)
+{
+	struct gru_mm_struct *gms = gts->ts_gms;
+	int ret;
+
+	while (1) {
+		wait_event(gms->ms_wait_queue,
+			   atomic_read(&gms->ms_range_active) == 0);
+		prefetchw(tfh);	/* Helps on hdw, required for emulator */
+		ret = gru_try_dropin(gts, tfh, cb);
+		if (ret <= 0)
+			return ret;
+		STAT(call_os_wait_queue);
+	}
+}
+
+/*
+ * This interface is called as a result of a user detecting a "call OS" bit
+ * in a user CB. Normally means that a TLB fault has occurred.
+ * 	cb - user virtual address of the CB
+ */
+int gru_handle_user_call_os(unsigned long cb)
+{
+	struct gru_tlb_fault_handle *tfh;
+	struct gru_thread_state *gts;
+	unsigned long __user *cbp;
+	int ucbnum, cbrnum, ret = -EINVAL;
+
+	STAT(call_os);
+	gru_dbg(grudev, "address 0x%lx\n", cb);
+
+	/* sanity check the cb pointer */
+	ucbnum = CBNUM(cb);
+	if ((cb & (GRU_HANDLE_STRIDE - 1)) || ucbnum >= GRU_NUM_CB)
+		return -EINVAL;
+	cbp = (unsigned long *)cb;
+
+	gts = gru_find_lock_gts(cb);
+	if (!gts)
+		return -EINVAL;
+
+	if (ucbnum >= gts->ts_cbr_au_count * GRU_CBR_AU_SIZE) {
+		ret = -EINVAL;
+		goto exit;
+	}
+
+	/*
+	 * If force_unload is set, the UPM TLB fault is phony. The task
+	 * has migrated to another node and the GSEG must be moved. Just
+	 * unload the context. The task will page fault and assign a new
+	 * context.
+	 */
+	ret = -EAGAIN;
+	cbrnum = thread_cbr_number(gts, ucbnum);
+	if (gts->ts_force_unload) {
+		gru_unload_context(gts, 1);
+	} else if (gts->ts_gru) {
+		tfh = get_tfh_by_index(gts->ts_gru, cbrnum);
+		ret = gru_user_dropin(gts, tfh, cbp);
+	}
+exit:
+	gru_unlock_gts(gts);
+	return ret;
+}
+
+/*
+ * Fetch the exception detail information for a CB that terminated with
+ * an exception.
+ */
+int gru_get_exception_detail(unsigned long arg)
+{
+	struct control_block_extended_exc_detail excdet;
+	struct gru_control_block_extended *cbe;
+	struct gru_thread_state *gts;
+	int ucbnum, cbrnum, ret;
+
+	STAT(user_exception);
+	if (copy_from_user(&excdet, (void __user *)arg, sizeof(excdet)))
+		return -EFAULT;
+
+	gru_dbg(grudev, "address 0x%lx\n", excdet.cb);
+	gts = gru_find_lock_gts(excdet.cb);
+	if (!gts)
+		return -EINVAL;
+
+	if (gts->ts_gru) {
+		ucbnum = CBNUM(excdet.cb);
+		cbrnum = thread_cbr_number(gts, ucbnum);
+		cbe = get_cbe_by_index(gts->ts_gru, cbrnum);
+		excdet.opc = cbe->opccpy;
+		excdet.exopc = cbe->exopccpy;
+		excdet.ecause = cbe->ecause;
+		excdet.exceptdet0 = cbe->idef1upd;
+		excdet.exceptdet1 = cbe->idef3upd;
+		ret = 0;
+	} else {
+		ret = -EAGAIN;
+	}
+	gru_unlock_gts(gts);
+
+	gru_dbg(grudev, "address 0x%lx, ecause 0x%x\n", excdet.cb,
+		excdet.ecause);
+	if (!ret && copy_to_user((void __user *)arg, &excdet, sizeof(excdet)))
+		ret = -EFAULT;
+	return ret;
+}
+
+/*
+ * User request to unload a context. Content is saved for possible reload.
+ */
+int gru_user_unload_context(unsigned long arg)
+{
+	struct gru_thread_state *gts;
+	struct gru_unload_context_req req;
+
+	STAT(user_unload_context);
+	if (copy_from_user(&req, (void __user *)arg, sizeof(req)))
+		return -EFAULT;
+
+	gru_dbg(grudev, "gseg 0x%lx\n", req.gseg);
+
+	gts = gru_find_lock_gts(req.gseg);
+	if (!gts)
+		return -EINVAL;
+
+	if (gts->ts_gru)
+		gru_unload_context(gts, 1);
+	gru_unlock_gts(gts);
+
+	return 0;
+}
+
+/*
+ * User request to flush a range of virtual addresses from the GRU TLB
+ * (Mainly for testing).
+ */
+int gru_user_flush_tlb(unsigned long arg)
+{
+	struct gru_thread_state *gts;
+	struct gru_flush_tlb_req req;
+
+	STAT(user_flush_tlb);
+	if (copy_from_user(&req, (void __user *)arg, sizeof(req)))
+		return -EFAULT;
+
+	gru_dbg(grudev, "gseg 0x%lx, vaddr 0x%lx, len 0x%lx\n", req.gseg,
+		req.vaddr, req.len);
+
+	gts = gru_find_lock_gts(req.gseg);
+	if (!gts)
+		return -EINVAL;
+
+	gru_flush_tlb_range(gts->ts_gms, req.vaddr, req.vaddr + req.len);
+	gru_unlock_gts(gts);
+
+	return 0;
+}
+
+/*
+ * Register the current task as the user of the GSEG slice.
+ * Needed for TLB fault interrupt targeting.
+ */
+int gru_set_task_slice(long address)
+{
+	struct gru_thread_state *gts;
+
+	STAT(set_task_slice);
+	gru_dbg(grudev, "address 0x%lx\n", address);
+	gts = gru_alloc_locked_gts(address);
+	if (!gts)
+		return -EINVAL;
+
+	gts->ts_tgid_owner = current->tgid;
+	gru_unlock_gts(gts);
+
+	return 0;
+}

-- 
Thanks

Jack Steiner (steiner@sgi.com)          651-683-5302
Sr Principal Engineer                   SGI - Silicon Graphics, Inc.



  parent reply	other threads:[~2008-06-09 21:12 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-06-09 21:10 [patch 00/11] GRU Driver steiner
2008-06-09 21:10 ` [patch 01/11] GRU Driver - hardware data structures steiner
2008-06-09 22:52   ` Andrew Morton
2008-06-10  4:07     ` Andi Kleen
2008-06-11 18:57     ` Jack Steiner
2008-06-09 21:10 ` [patch 02/11] GRU Driver - GRU instructions & macros steiner
2008-06-09 21:10 ` [patch 03/11] GRU Driver - driver internal header files steiner
2008-06-09 21:38   ` Roland Dreier
2008-06-10 14:57     ` Jack Steiner
2008-06-09 21:10 ` [patch 04/11] GRU Driver - kernel services " steiner
2008-06-09 21:10 ` [patch 05/11] GRU Driver - driver initialization, file & vma ops steiner
2008-06-09 21:10 ` steiner [this message]
2008-06-09 21:10 ` [patch 07/11] GRU Driver - kernel services provide by driver steiner
2008-06-09 21:10 ` [patch 08/11] GRU Driver - resource management steiner
2008-06-09 21:10 ` [patch 09/11] GRU Driver - /proc interfaces steiner
2008-06-09 21:32   ` Roland Dreier
2008-06-09 22:11     ` Jack Steiner
2008-06-10 14:20       ` Roland Dreier
2008-06-09 21:10 ` [patch 10/11] GRU Driver - TLB flushing, MMUOPS callouts steiner
2008-06-09 21:10 ` [patch 11/11] GRU Driver - makefile & Kconfig file changes steiner
2008-06-09 21:35   ` Roland Dreier
2008-06-10 14:41     ` Jack Steiner
2008-06-12 13:27 ` [patch 00/11] GRU Driver Ingo Molnar
2008-06-12 14:05   ` Jack Steiner
2008-06-12 18:03     ` Andrew Morton
2008-06-12 20:52       ` Jack Steiner

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20080609211046.208712459@attica.americas.sgi.com \
    --to=steiner@sgi.com \
    --cc=akpm@osdl.org \
    --cc=andrea@qumranet.com \
    --cc=holt@sgi.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=tglx@linutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®