All of lore.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Arnd Bergmann <arnd@arndb.de>,
	Rob Herring <robherring2@gmail.com>,
	Randy Dunlap <rdunlap@infradead.org>,
	Anton Vorontsov <anton@enomsg.org>,
	Colin Cross <ccross@android.com>, Olof Johansson <olof@lixom.net>,
	Russell King <linux@arm.linux.org.uk>,
	Kees Cook <keescook@chromium.org>,
	Tony Lindgren <tony@atomide.com>, Tony Luck <tony.luck@intel.com>
Subject: [PATCH 3.10 13/44] pstore-ram: Allow optional mapping with pgprot_noncached
Date: Tue, 13 Jan 2015 23:23:34 -0800	[thread overview]
Message-ID: <20150114072228.243117366@linuxfoundation.org> (raw)
In-Reply-To: <20150114072227.419663002@linuxfoundation.org>

3.10-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Tony Lindgren <tony@atomide.com>

commit 027bc8b08242c59e19356b4b2c189f2d849ab660 upstream.

On some ARMs the memory can be mapped pgprot_noncached() and still
be working for atomic operations. As pointed out by Colin Cross
<ccross@android.com>, in some cases you do want to use
pgprot_noncached() if the SoC supports it to see a debug printk
just before a write hanging the system.

On ARMs, the atomic operations on strongly ordered memory are
implementation defined. So let's provide an optional kernel parameter
for configuring pgprot_noncached(), and use pgprot_writecombine() by
default.

Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Rob Herring <robherring2@gmail.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Anton Vorontsov <anton@enomsg.org>
Cc: Colin Cross <ccross@android.com>
Cc: Olof Johansson <olof@lixom.net>
Cc: Russell King <linux@arm.linux.org.uk>
Acked-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 Documentation/ramoops.txt  |   13 +++++++++++--
 fs/pstore/ram.c            |   13 +++++++++++--
 fs/pstore/ram_core.c       |   31 ++++++++++++++++++++++---------
 include/linux/pstore_ram.h |    4 +++-
 4 files changed, 47 insertions(+), 14 deletions(-)

--- a/Documentation/ramoops.txt
+++ b/Documentation/ramoops.txt
@@ -14,11 +14,19 @@ survive after a restart.
 
 1. Ramoops concepts
 
-Ramoops uses a predefined memory area to store the dump. The start and size of
-the memory area are set using two variables:
+Ramoops uses a predefined memory area to store the dump. The start and size
+and type of the memory area are set using three variables:
   * "mem_address" for the start
   * "mem_size" for the size. The memory size will be rounded down to a
   power of two.
+  * "mem_type" to specifiy if the memory type (default is pgprot_writecombine).
+
+Typically the default value of mem_type=0 should be used as that sets the pstore
+mapping to pgprot_writecombine. Setting mem_type=1 attempts to use
+pgprot_noncached, which only works on some platforms. This is because pstore
+depends on atomic operations. At least on ARM, pgprot_noncached causes the
+memory to be mapped strongly ordered, and atomic operations on strongly ordered
+memory are implementation defined, and won't work on many ARMs such as omaps.
 
 The memory area is divided into "record_size" chunks (also rounded down to
 power of two) and each oops/panic writes a "record_size" chunk of
@@ -55,6 +63,7 @@ Setting the ramoops parameters can be do
 static struct ramoops_platform_data ramoops_data = {
         .mem_size               = <...>,
         .mem_address            = <...>,
+        .mem_type               = <...>,
         .record_size            = <...>,
         .dump_oops              = <...>,
         .ecc                    = <...>,
--- a/fs/pstore/ram.c
+++ b/fs/pstore/ram.c
@@ -61,6 +61,11 @@ module_param(mem_size, ulong, 0400);
 MODULE_PARM_DESC(mem_size,
 		"size of reserved RAM used to store oops/panic logs");
 
+static unsigned int mem_type;
+module_param(mem_type, uint, 0600);
+MODULE_PARM_DESC(mem_type,
+		"set to 1 to try to use unbuffered memory (default 0)");
+
 static int dump_oops = 1;
 module_param(dump_oops, int, 0600);
 MODULE_PARM_DESC(dump_oops,
@@ -79,6 +84,7 @@ struct ramoops_context {
 	struct persistent_ram_zone *fprz;
 	phys_addr_t phys_addr;
 	unsigned long size;
+	unsigned int memtype;
 	size_t record_size;
 	size_t console_size;
 	size_t ftrace_size;
@@ -331,7 +337,8 @@ static int ramoops_init_przs(struct devi
 		size_t sz = cxt->record_size;
 
 		cxt->przs[i] = persistent_ram_new(*paddr, sz, 0,
-						  &cxt->ecc_info);
+						  &cxt->ecc_info,
+						  cxt->memtype);
 		if (IS_ERR(cxt->przs[i])) {
 			err = PTR_ERR(cxt->przs[i]);
 			dev_err(dev, "failed to request mem region (0x%zx@0x%llx): %d\n",
@@ -361,7 +368,7 @@ static int ramoops_init_prz(struct devic
 		return -ENOMEM;
 	}
 
-	*prz = persistent_ram_new(*paddr, sz, sig, &cxt->ecc_info);
+	*prz = persistent_ram_new(*paddr, sz, sig, &cxt->ecc_info, cxt->memtype);
 	if (IS_ERR(*prz)) {
 		int err = PTR_ERR(*prz);
 
@@ -411,6 +418,7 @@ static int ramoops_probe(struct platform
 	cxt->dump_read_cnt = 0;
 	cxt->size = pdata->mem_size;
 	cxt->phys_addr = pdata->mem_address;
+	cxt->memtype = pdata->mem_type;
 	cxt->record_size = pdata->record_size;
 	cxt->console_size = pdata->console_size;
 	cxt->ftrace_size = pdata->ftrace_size;
@@ -541,6 +549,7 @@ static void ramoops_register_dummy(void)
 
 	dummy_data->mem_size = mem_size;
 	dummy_data->mem_address = mem_address;
+	dummy_data->mem_type = 0;
 	dummy_data->record_size = record_size;
 	dummy_data->console_size = ramoops_console_size;
 	dummy_data->ftrace_size = ramoops_ftrace_size;
--- a/fs/pstore/ram_core.c
+++ b/fs/pstore/ram_core.c
@@ -333,7 +333,8 @@ void persistent_ram_zap(struct persisten
 	persistent_ram_update_header_ecc(prz);
 }
 
-static void *persistent_ram_vmap(phys_addr_t start, size_t size)
+static void *persistent_ram_vmap(phys_addr_t start, size_t size,
+		unsigned int memtype)
 {
 	struct page **pages;
 	phys_addr_t page_start;
@@ -345,7 +346,10 @@ static void *persistent_ram_vmap(phys_ad
 	page_start = start - offset_in_page(start);
 	page_count = DIV_ROUND_UP(size + offset_in_page(start), PAGE_SIZE);
 
-	prot = pgprot_writecombine(PAGE_KERNEL);
+	if (memtype)
+		prot = pgprot_noncached(PAGE_KERNEL);
+	else
+		prot = pgprot_writecombine(PAGE_KERNEL);
 
 	pages = kmalloc(sizeof(struct page *) * page_count, GFP_KERNEL);
 	if (!pages) {
@@ -364,27 +368,35 @@ static void *persistent_ram_vmap(phys_ad
 	return vaddr;
 }
 
-static void *persistent_ram_iomap(phys_addr_t start, size_t size)
+static void *persistent_ram_iomap(phys_addr_t start, size_t size,
+		unsigned int memtype)
 {
+	void *va;
+
 	if (!request_mem_region(start, size, "persistent_ram")) {
 		pr_err("request mem region (0x%llx@0x%llx) failed\n",
 			(unsigned long long)size, (unsigned long long)start);
 		return NULL;
 	}
 
-	return ioremap_wc(start, size);
+	if (memtype)
+		va = ioremap(start, size);
+	else
+		va = ioremap_wc(start, size);
+
+	return va;
 }
 
 static int persistent_ram_buffer_map(phys_addr_t start, phys_addr_t size,
-		struct persistent_ram_zone *prz)
+		struct persistent_ram_zone *prz, int memtype)
 {
 	prz->paddr = start;
 	prz->size = size;
 
 	if (pfn_valid(start >> PAGE_SHIFT))
-		prz->vaddr = persistent_ram_vmap(start, size);
+		prz->vaddr = persistent_ram_vmap(start, size, memtype);
 	else
-		prz->vaddr = persistent_ram_iomap(start, size);
+		prz->vaddr = persistent_ram_iomap(start, size, memtype);
 
 	if (!prz->vaddr) {
 		pr_err("%s: Failed to map 0x%llx pages at 0x%llx\n", __func__,
@@ -452,7 +464,8 @@ void persistent_ram_free(struct persiste
 }
 
 struct persistent_ram_zone *persistent_ram_new(phys_addr_t start, size_t size,
-			u32 sig, struct persistent_ram_ecc_info *ecc_info)
+			u32 sig, struct persistent_ram_ecc_info *ecc_info,
+			unsigned int memtype)
 {
 	struct persistent_ram_zone *prz;
 	int ret = -ENOMEM;
@@ -463,7 +476,7 @@ struct persistent_ram_zone *persistent_r
 		goto err;
 	}
 
-	ret = persistent_ram_buffer_map(start, size, prz);
+	ret = persistent_ram_buffer_map(start, size, prz, memtype);
 	if (ret)
 		goto err;
 
--- a/include/linux/pstore_ram.h
+++ b/include/linux/pstore_ram.h
@@ -53,7 +53,8 @@ struct persistent_ram_zone {
 };
 
 struct persistent_ram_zone *persistent_ram_new(phys_addr_t start, size_t size,
-			u32 sig, struct persistent_ram_ecc_info *ecc_info);
+			u32 sig, struct persistent_ram_ecc_info *ecc_info,
+			unsigned int memtype);
 void persistent_ram_free(struct persistent_ram_zone *prz);
 void persistent_ram_zap(struct persistent_ram_zone *prz);
 
@@ -76,6 +77,7 @@ ssize_t persistent_ram_ecc_string(struct
 struct ramoops_platform_data {
 	unsigned long	mem_size;
 	unsigned long	mem_address;
+	unsigned int	mem_type;
 	unsigned long	record_size;
 	unsigned long	console_size;
 	unsigned long	ftrace_size;



  parent reply	other threads:[~2015-01-14  7:48 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-01-14  7:23 [PATCH 3.10 00/44] 3.10.65-stable review Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 01/44] ocfs2: fix journal commit deadlock Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 02/44] ath9k_hw: fix hardware queue allocation Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 03/44] ath9k: fix BE/BK queue order Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 04/44] can: peak_usb: fix cleanup sequence order in case of error during init Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 05/44] can: peak_usb: fix memset() usage Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 06/44] swiotlb-xen: pass dev_addr to swiotlb_tbl_unmap_single Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 07/44] ath5k: fix hardware queue index assignment Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 08/44] ASoC: sigmadsp: Refuse to load firmware files with a non-supported version Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 09/44] ASoC: max98090: Fix ill-defined sidetone route Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 10/44] ASoC: dwc: Ensure FIFOs are flushed to prevent channel swap Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 11/44] PCI: Restore detection of read-only BARs Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 12/44] pstore-ram: Fix hangs by using write-combine mappings Greg Kroah-Hartman
2015-01-14  7:23 ` Greg Kroah-Hartman [this message]
2015-01-14  7:23 ` [PATCH 3.10 14/44] UBI: Fix invalid vfree() Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 15/44] UBI: Fix double free after do_sync_erase() Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 16/44] iommu/vt-d: Fix an off-by-one bug in __domain_mapping() Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 17/44] HID: i2c-hid: fix race condition reading reports Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 18/44] HID: i2c-hid: prevent buffer overflow in early IRQ Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 19/44] HID: roccat: potential out of bounds in pyra_sysfs_write_settings() Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 20/44] HID: add battery quirk for USB_DEVICE_ID_APPLE_ALU_WIRELESS_2011_ISO keyboard Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 22/44] x86_64, vdso: Fix the vdso address randomization algorithm Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 23/44] x86, vdso: Use asm volatile in __getcpu Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 24/44] driver core: Fix unbalanced device reference in drivers_probe Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 25/44] ALSA: usb-audio: extend KEF X300A FU 10 tweak to Arcam rPAC Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 26/44] ALSA: hda - using uninitialized data Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 27/44] ALSA: hda - Fix wrong gpio_dir & gpio_mask hint setups for IDT/STAC codecs Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 28/44] USB: cdc-acm: check for valid interfaces Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 29/44] genhd: check for int overflow in disk_expand_part_tbl() Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 30/44] cdc-acm: memory leak in error case Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 31/44] writeback: fix a subtle race condition in I_DIRTY clearing Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 32/44] serial: samsung: wait for transfer completion before clock disable Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 33/44] fs: nfsd: Fix signedness bug in compare_blob Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 34/44] nfsd4: fix xdr4 inclusion of escaped char Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 35/44] nilfs2: fix the nilfs_iget() vs. nilfs_new_inode() races Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 36/44] scripts/kernel-doc: dont eat struct members with __aligned Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 37/44] ARM: mvebu: disable I/O coherency on non-SMP situations on Armada 370/375/38x/XP Greg Kroah-Hartman
2015-01-14  7:23 ` [PATCH 3.10 38/44] Btrfs: dont delay inode ref updates during log replay Greg Kroah-Hartman
2015-01-14  7:24 ` [PATCH 3.10 39/44] perf/x86/intel/uncore: Make sure only uncore events are collected Greg Kroah-Hartman
2015-01-14  7:24 ` [PATCH 3.10 40/44] perf: Fix events installation during moving group Greg Kroah-Hartman
2015-01-14  7:24 ` [PATCH 3.10 41/44] perf session: Do not fail on processing out of order event Greg Kroah-Hartman
2015-01-14  7:24 ` [PATCH 3.10 42/44] mm, vmscan: prevent kswapd livelock due to pfmemalloc-throttled process being killed Greg Kroah-Hartman
2015-01-14  7:24 ` [PATCH 3.10 43/44] mm: propagate error from stack expansion even for guard page Greg Kroah-Hartman
2015-01-14  7:24 ` [PATCH 3.10 44/44] mm: Dont count the stack guard page towards RLIMIT_STACK Greg Kroah-Hartman
2015-01-14 22:49 ` [PATCH 3.10 00/44] 3.10.65-stable review Shuah Khan
2015-01-15  0:43 ` Guenter Roeck

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=20150114072228.243117366@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=anton@enomsg.org \
    --cc=arnd@arndb.de \
    --cc=ccross@android.com \
    --cc=keescook@chromium.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux@arm.linux.org.uk \
    --cc=olof@lixom.net \
    --cc=rdunlap@infradead.org \
    --cc=robherring2@gmail.com \
    --cc=stable@vger.kernel.org \
    --cc=tony.luck@intel.com \
    --cc=tony@atomide.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.