While debug_dma_assert_idle() checks if a given *page* is actively
undergoing dma the valid granularity of a dma mapping is a *cacheline*.
Sander's testing shows that the warning message "DMA-API: exceeded 7
overlapping mappings of pfn..." is falsely triggering. The test is
simply mapping multiple cachelines in a given page.
Ultimately we want overlap tracking to be valid as it is a real api
violation, so we need to track active mappings by cachelines. Update
the active dma tracking to use the page-frame-relative cacheline of the
mapping as the key, and update debug_dma_assert_idle() to check for all
possible mapped cachelines for a given page.
However, the need to track active mappings is only relevant when the
dma-mapping is writable by the device. In fact it is fairly standard
for read-only mappings to have hundreds or thousands of overlapping
mappings at once. Limiting the overlap tracking to writable
(!DMA_TO_DEVICE) eliminates this class of false-positive overlap
reports.
Note, the radix gang lookup is sub-optimal. It would be best if it
stopped fetching entries once the search passed a page boundary.
Nevertheless, this implementation does not perturb the original net_dma
failing case. That is to say the extra overhead does not show up in
terms of making the failing case pass due to a timing change.
References:
http://marc.info/?l=linux-netdev&m=139232263419315&w=2http://marc.info/?l=linux-netdev&m=139217088107122&w=2
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: Sander Eikelenboom <linux@eikelenboom.it>
Reported-by: Dave Jones <davej@redhat.com>
Tested-by: Dave Jones <davej@redhat.com>
Tested-by: Sander Eikelenboom <linux@eikelenboom.it>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Francois Romieu <romieu@fr.zoreil.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 0abdd7a81b ("dma-debug: introduce debug_dma_assert_idle()") was
reworked to expand the overlap counter to the full range expressable by
3 tag bits, but it has a thinko in treating the overlap counter as a
pure reference count for the entry.
Instead of deleting when the reference-count drops to zero, we need to
delete when the overlap-count drops below zero. Also, when detecting
overflow we can just test the overlap-count > MAX rather than applying
special meaning to 0.
Regression report available here:
http://marc.info/?l=linux-netdev&m=139073373932386&w=2
This patch, now tested on the original net_dma case, sees the expected
handful of reports before the eventual data corruption occurs.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reported-by: Sander Eikelenboom <linux@eikelenboom.it>
Cc: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Record actively mapped pages and provide an api for asserting a given
page is dma inactive before execution proceeds. Placing
debug_dma_assert_idle() in cow_user_page() flagged the violation of the
dma-api in the NET_DMA implementation (see commit 7787380336 "net_dma:
mark broken").
The implementation includes the capability to count, in a limited way,
repeat mappings of the same page that occur without an intervening
unmap. This 'overlap' counter is limited to the few bits of tag space
in a radix tree. This mechanism is added to mitigate false negative
cases where, for example, a page is dma mapped twice and
debug_dma_assert_idle() is called after the page is un-mapped once.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Vinod Koul <vinod.koul@intel.com>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There were reports of the igb driver unmapping buffers without calling
dma_mapping_error. On closer inspection issues were found in the DMA
debug API and how it handled multiple mappings of the same buffer.
The issue I found is the fact that the debug_dma_mapping_error would
only set the map_err_type to MAP_ERR_CHECKED in the case that the was
only one match for device and device address. However in the case of
non-IOMMU, multiple addresses existed and as a result it was not setting
this field once a second mapping was instantiated. I have resolved this
by changing the search so that it instead will now set MAP_ERR_CHECKED
on the first buffer that matches the device and DMA address that is
currently in the state MAP_ERR_NOT_CHECKED.
A secondary side effect of this patch is that in the case of multiple
buffers using the same address only the last mapping will have a valid
map_err_type. The previous mappings will all end up with map_err_type
set to MAP_ERR_CHECKED because of the dma_mapping_error call in
debug_dma_map_page. However this behavior may be preferable as it means
you will likely only see one real error per multi-mapped buffer, versus
the current behavior of multiple false errors mer multi-mapped buffer.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Cc: Joerg Roedel <joro@8bytes.org>
Reviewed-by: Shuah Khan <shuah.khan@hp.com>
Tested-by: Shuah Khan <shuah.khan@hp.com>
Cc: Jakub Kicinski <kubakici@wp.pl>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In check_unmap() it is possible to get into a dead-locked state if
dma_mapping_error is called. The problem is that the bucket is locked in
check_unmap, and locked again by debug_dma_mapping_error which is called
by dma_mapping_error. To resolve that we must release the lock on the
bucket before making the call to dma_mapping_error.
[akpm@linux-foundation.org: restore 80-col trickery to be consistent with the rest of the file]
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Cc: Joerg Roedel <joro@8bytes.org>
Reviewed-by: Shuah Khan <shuah.khan@hp.com>
Tested-by: Shuah Khan <shuah.khan@hp.com>
Cc: Jakub Kicinski <kubakici@wp.pl>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
dma-debug depends on get_dma_ops() interface. Several architectures
do not define dma_ops and get_dma_ops(). When dma debug interfaces are
used on an architecture (e.g: c6x) that doesn't define get_dmap_ops(),
compilation fails. Changing dma-debug to call dma_mapping_error() instead
of defining its own that calls get_dma_ops(), such that the internal use of
dma_mapping_error() doesn't interfere with the debug_dma_mapping_error()
interface's mapping error checks. Moving dma_mapping_error() checks in
check_unmap() under the dma debug entry not found is sufficient to fix the
problem.
Reference: https://lkml.org/lkml/2012/10/26/367
Signed-off-by: Shuah Khan <shuah.khan@hp.com>
Reported-by: Mark Salter <msalter@redhat.com>
Signed-off-by: Joerg Roedel <joro@8bytes.org>
Add dma-debug interface debug_dma_mapping_error() to debug
drivers that fail to check dma mapping errors on addresses
returned by dma_map_single() and dma_map_page() interfaces.
This interface clears a flag set by debug_dma_map_page() to
indicate that dma_mapping_error() has been called by the
driver. When driver does unmap, debug_dma_unmap() checks the
flag and if this flag is still set, prints warning message
that includes call trace that leads up to the unmap. This
interface can be called from dma_mapping_error() routines to
enable dma mapping error check debugging.
Tested: Intel iommu and swiotlb (iommu=soft) on x86-64 with
CONFIG_DMA_API_DEBUG enabled and disabled.
Signed-off-by: Shuah Khan <shuah.khan@hp.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
If there is only one match, the unique matched entry should be returned.
Without the fix, the upcoming dma debug interfaces ("dma-debug: new
interfaces to debug dma mapping errors") can't work reliably because
only device and dma_addr are passed to dma_mapping_error().
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Reported-by: Wu Fengguang <fengguang.wu@intel.com>
Cc: Joerg Roedel <joerg.roedel@amd.com>
Tested-by: Shuah Khan <shuah.khan@hp.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Jakub Kicinski <kubakici@wp.pl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Remove local BUS_NOTIFY_UNBOUND_DRIVER define. This is not used since
BUS_NOTIFY_UNBOUND_DRIVER is defined in include/linux/device.h
Signed-off-by: Shuah Khan <shuah.khan@hp.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Even though it has "bool" in the name, you have pass a u32 pointer to
debugfs_create_bool(). Otherwise you get memory corruption in
write_file_bool(). Fortunately in this case the corruption happens in
an alignment hole between variables so it doesn't cause any problems.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Saving stack trace can take a while and once the entry
is allocated free_entries_lock is no longer needed.
Signed-off-by: Jakub Kicinski <kubakici@wp.pl>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
As part of the removal of get_driver()/put_driver(), this patch
(as1512) gets rid of various useless and unnecessary calls in several
drivers. In some cases it may be desirable to pin the driver by
calling try_module_get(), but that can be done later.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
CC: "David S. Miller" <davem@davemloft.net>
CC: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
CC: Michael Buesch <m@bues.ch>
CC: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
cppcheck reported:
[lib/dma-debug.c:248] -> [lib/dma-debug.c:248]: (style) Same expression on both sides of '=='.
Signed-off-by: Thomas Jarosch <thomas.jarosch@intra2net.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
There are no modular calls here, so just the minimal header for
the EXPORT_SYMBOL macro will suffice.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Summary:
Users of the pci_dma_sync_single_* api allow users to sync address ranges within
the range of a mapped entry (i.e. you can dma map address X to dma_addr_t A and
then pci_dma_sync_single on dma_addr_t A+1. The dma-debug library however
assume dma syncs will always occur using the base address of a mapped region,
and uses that assumption to find entries in its hash table. Since thats often
(but not always the case), the dma debug library can give us false errors about
missing entries, which are reported as syncing of memory not allocated by the
driver. This was noted in the cxgb3 driver as this error:
WARNING: at lib/dma-debug.c:902 check_sync+0xdd/0x48c()
Hardware name: To be filled by O.E.M.
cxgb3 0000:01:00.0: DMA-API: device driver tries to sync DMA memory it has not
allocated [device address=0x00000000fff97800] [size=1984 bytes]
Modules linked in: autofs4 sunrpc cpufreq_ondemand acpi_cpufreq freq_table
mperf ip6t_REJECT nf_conntrack_ipv6 ip6table_filter ip6_tables ipv6 uinput
snd_hda_codec_intelhdmi snd_hda_codec_realtek snd_hda_intel snd_hda_codec
snd_hwdep snd_seq snd_seq_device snd_pcm snd_timer e1000e snd soundcore r8169
cxgb3 iTCO_wdt snd_page_alloc mii shpchp i2c_i801 iTCO_vendor_support mdio
microcode firewire_ohci firewire_core crc_itu_t ata_generic pata_acpi i915
drm_kms_helper drm i2c_algo_bit i2c_core video output [last unloaded:
scsi_wait_scan]
Pid: 1818, comm: ifconfig Not tainted 2.6.35-0.23.rc3.git6.fc14.x86_64 #1
Call Trace:
[<ffffffff81050f71>] warn_slowpath_common+0x85/0x9d
[<ffffffff8105102c>] warn_slowpath_fmt+0x46/0x48
[<ffffffff8124658e>] ? check_sync+0x39/0x48c
[<ffffffff8107c470>] ? trace_hardirqs_on+0xd/0xf
[<ffffffff81246632>] check_sync+0xdd/0x48c
[<ffffffff81246ca6>] debug_dma_sync_single_for_device+0x3f/0x41
[<ffffffffa011615c>] ? pci_map_page+0x84/0x97 [cxgb3]
[<ffffffffa0117bc3>] pci_dma_sync_single_for_device.clone.0+0x65/0x6e [cxgb3]
[<ffffffffa0117ed1>] refill_fl+0x305/0x30a [cxgb3]
[<ffffffffa011857d>] t3_sge_alloc_qset+0x6a7/0x821 [cxgb3]
[<ffffffffa010a07b>] cxgb_up+0x4d0/0xe62 [cxgb3]
[<ffffffff81086037>] ? __module_text_address+0x12/0x58
[<ffffffffa010aa4c>] cxgb_open+0x3f/0x309 [cxgb3]
[<ffffffff813e9f6c>] __dev_open+0x8e/0xbc
[<ffffffff813e7ca5>] __dev_change_flags+0xbe/0x142
[<ffffffff813e9ea8>] dev_change_flags+0x21/0x57
[<ffffffff81445937>] devinet_ioctl+0x29a/0x54b
[<ffffffff811f9a87>] ? inode_has_perm+0xaa/0xce
[<ffffffff81446ed2>] inet_ioctl+0x8f/0xa7
[<ffffffff813d683a>] sock_do_ioctl+0x29/0x48
[<ffffffff813d6c83>] sock_ioctl+0x213/0x222
[<ffffffff81137f78>] vfs_ioctl+0x32/0xa6
[<ffffffff811384e2>] do_vfs_ioctl+0x47a/0x4b3
[<ffffffff81138571>] sys_ioctl+0x56/0x79
[<ffffffff81009c32>] system_call_fastpath+0x16/0x1b
---[ end trace 69a4d4cc77b58004 ]---
(some edits by Joerg Roedel)
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Reported-by: Jay Fenalson <fenlason@redhat.com>
CC: Divy LeRay <divy@chelsio.com>
CC: Stanislaw Gruszka <sgruszka@redhat.com>
CC: Joerg Roedel <joerg.roedel@amd.com>
CC: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
When driver leak dma mapping, print additional information about one of
leaked entries, to to help investigate problem. Patch should be useful
for debugging drivers, which maps many different class of buffers.
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
All file_operations should get a .llseek operation so we can make
nonseekable_open the default for future file operations without a
.llseek pointer.
The three cases that we can automatically detect are no_llseek, seq_lseek
and default_llseek. For cases where we can we can automatically prove that
the file offset is always ignored, we use noop_llseek, which maintains
the current behavior of not returning an error from a seek.
New drivers should normally not use noop_llseek but instead use no_llseek
and call nonseekable_open at open time. Existing drivers can be converted
to do the same when the maintainer knows for certain that no user code
relies on calling seek on the device file.
The generated code is often incorrectly indented and right now contains
comments that clarify for each added line why a specific variant was
chosen. In the version that gets submitted upstream, the comments will
be gone and I will manually fix the indentation, because there does not
seem to be a way to do that using coccinelle.
Some amount of new code is currently sitting in linux-next that should get
the same modifications, which I will do at the end of the merge window.
Many thanks to Julia Lawall for helping me learn to write a semantic
patch that does all this.
===== begin semantic patch =====
// This adds an llseek= method to all file operations,
// as a preparation for making no_llseek the default.
//
// The rules are
// - use no_llseek explicitly if we do nonseekable_open
// - use seq_lseek for sequential files
// - use default_llseek if we know we access f_pos
// - use noop_llseek if we know we don't access f_pos,
// but we still want to allow users to call lseek
//
@ open1 exists @
identifier nested_open;
@@
nested_open(...)
{
<+...
nonseekable_open(...)
...+>
}
@ open exists@
identifier open_f;
identifier i, f;
identifier open1.nested_open;
@@
int open_f(struct inode *i, struct file *f)
{
<+...
(
nonseekable_open(...)
|
nested_open(...)
)
...+>
}
@ read disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ read_no_fpos disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
... when != off
}
@ write @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ write_no_fpos @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
... when != off
}
@ fops0 @
identifier fops;
@@
struct file_operations fops = {
...
};
@ has_llseek depends on fops0 @
identifier fops0.fops;
identifier llseek_f;
@@
struct file_operations fops = {
...
.llseek = llseek_f,
...
};
@ has_read depends on fops0 @
identifier fops0.fops;
identifier read_f;
@@
struct file_operations fops = {
...
.read = read_f,
...
};
@ has_write depends on fops0 @
identifier fops0.fops;
identifier write_f;
@@
struct file_operations fops = {
...
.write = write_f,
...
};
@ has_open depends on fops0 @
identifier fops0.fops;
identifier open_f;
@@
struct file_operations fops = {
...
.open = open_f,
...
};
// use no_llseek if we call nonseekable_open
////////////////////////////////////////////
@ nonseekable1 depends on !has_llseek && has_open @
identifier fops0.fops;
identifier nso ~= "nonseekable_open";
@@
struct file_operations fops = {
... .open = nso, ...
+.llseek = no_llseek, /* nonseekable */
};
@ nonseekable2 depends on !has_llseek @
identifier fops0.fops;
identifier open.open_f;
@@
struct file_operations fops = {
... .open = open_f, ...
+.llseek = no_llseek, /* open uses nonseekable */
};
// use seq_lseek for sequential files
/////////////////////////////////////
@ seq depends on !has_llseek @
identifier fops0.fops;
identifier sr ~= "seq_read";
@@
struct file_operations fops = {
... .read = sr, ...
+.llseek = seq_lseek, /* we have seq_read */
};
// use default_llseek if there is a readdir
///////////////////////////////////////////
@ fops1 depends on !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier readdir_e;
@@
// any other fop is used that changes pos
struct file_operations fops = {
... .readdir = readdir_e, ...
+.llseek = default_llseek, /* readdir is present */
};
// use default_llseek if at least one of read/write touches f_pos
/////////////////////////////////////////////////////////////////
@ fops2 depends on !fops1 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read.read_f;
@@
// read fops use offset
struct file_operations fops = {
... .read = read_f, ...
+.llseek = default_llseek, /* read accesses f_pos */
};
@ fops3 depends on !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write.write_f;
@@
// write fops use offset
struct file_operations fops = {
... .write = write_f, ...
+ .llseek = default_llseek, /* write accesses f_pos */
};
// Use noop_llseek if neither read nor write accesses f_pos
///////////////////////////////////////////////////////////
@ fops4 depends on !fops1 && !fops2 && !fops3 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
identifier write_no_fpos.write_f;
@@
// write fops use offset
struct file_operations fops = {
...
.write = write_f,
.read = read_f,
...
+.llseek = noop_llseek, /* read and write both use no f_pos */
};
@ depends on has_write && !has_read && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write_no_fpos.write_f;
@@
struct file_operations fops = {
... .write = write_f, ...
+.llseek = noop_llseek, /* write uses no f_pos */
};
@ depends on has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
@@
struct file_operations fops = {
... .read = read_f, ...
+.llseek = noop_llseek, /* read uses no f_pos */
};
@ depends on !has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
@@
struct file_operations fops = {
...
+.llseek = noop_llseek, /* no read or write fn */
};
===== End semantic patch =====
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Julia Lawall <julia@diku.dk>
Cc: Christoph Hellwig <hch@infradead.org>
Earlier in this function we set the last byte of "buf" to NULL so we
always hit the break statement and "i" is never equal to NAME_MAX_LEN.
This patch doesn't change how the driver works but it silences a Smatch
warning and it makes it clearer that we don't write past the end of the
array.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
warning: symbol 'filter_fops' was not declared. Should it be static?
Signed-off-by: Thiago Farina <tfransosi@gmail.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
There is no need to perform full BIDIR sync (copying the buffers in case
of swiotlb and similar schemes) if we know that the owner (CPU or device)
hasn't altered the data.
Addresses the false-positive reported at
http://bugzilla.kernel.org/show_bug.cgi?id=14169
Signed-off-by: Krzysztof Halasa <khc@pm.waw.pl>
Cc: David Miller <davem@davemloft.net>
Cc: Joerg Roedel <joerg.roedel@amd.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Stephen Rothwell reported the following build warning:
lib/dma-debug.c: In function 'dma_debug_device_change':
lib/dma-debug.c:680: warning: 'return' with no value, in function returning non-void
Introduced by commit f797d9881b
("dma-debug: Do not add notifier when dma debugging is disabled").
Return 0 [notify-done] when disabled. (this is standard bus notifier behavior.)
Signed-off-by: Shaun Ruffell <sruffell@digium.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: <stable@kernel.org>
LKML-Reference: <20091231125624.GA14666@liondog.tnic>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
If CONFIG_HAVE_DMA_API_DEBUG is defined and "dma_debug=off" is
specified on the kernel command line, when you detach a driver from a
device you can cause the following NULL pointer dereference:
BUG: unable to handle kernel NULL pointer dereference at (null)
IP: [<c0580d35>] dma_debug_device_change+0x5d/0x117
The problem is that the dma_debug_device_change notifier function is
added to the bus notifier chain even though the dma_entry_hash array
was never initialized. If dma debugging is disabled, this patch both
prevents dma_debug_device_change notifiers from being added to the
chain, and additionally ensures that the dma_debug_device_change
notifier function is a no-op.
Cc: stable@kernel.org
Signed-off-by: Shaun Ruffell <sruffell@digium.com>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
When PAE is enabled in the kernel configuration the size of
phys_addr_t differs from the size of a void pointer. The gcc
prints a warning about that in dma-debug code.
This patch fixes the warning by converting the output to
unsigned long long instead of a pointer.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
While it's debatable whether or not a NULL device argument to
the DMA API functions is valid... since it certainly isn't
valid on devices with an IOMMU... dma-debug really shouldn't be
dereferencing null pointers either.
Guard against that in err_printk and the driver_filter
functions. A Fedora rawhide user was seeing this in one of the
dvb drivers resulting in an oops on boot.
[ A patch has been sent for testing to the driver, but I feel
the dma debugging support should be fixed as well. (There's
still a pile of legacy garbage in the kernel passing null
pointers to dma_{alloc,free}_*. :( ]
Signed-off-by: Kyle McMartin <kyle@redhat.com>
Cc: mchehab@infradead.org
Cc: Joerg Roedel <joerg.roedel@amd.com>
LKML-Reference: <20090820011708.GP25206@bombadil.infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Linus noticed how unclean and buggy the overlap() function is:
- It uses convoluted (and bug-causing) positive checks for
range overlap - instead of using a more natural negative
check.
- Even the positive checks are buggy: a positive intersection
check has four natural cases while we checked only for three,
missing the (addr < start && addr2 == end) case for example.
- The variables are mis-named, making it non-obvious how the
check was done.
- It needlessly uses u64 instead of unsigned long. Since these
are kernel memory pointers and we explicitly exclude highmem
ranges anyway we cannot ever overflow 32 bits, even if we
could. (and on 64-bit it doesnt matter anyway)
All in one, this function needs a total revamp. I used Linus's
suggestions minus the paranoid checks (we cannot overflow really
because if we get totally bad DMA ranges passed far more things
break in the systems than just DMA debugging). I also fixed a
few other small details i noticed.
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Alan Cox reported that lockdep runs out of its stack-trace entries
with certain configs:
BUG: MAX_STACK_TRACE_ENTRIES too low
This happens because there are 1024 hash buckets, each with a
separate lock. Lockdep puts each lock into a separate lock class and
tracks them independently.
But in reality we never take more than one of the buckets, so they
really belong into a single lock-class. Annotate the has bucket lock
init accordingly.
[ Impact: reduce the lockdep footprint of dma-debug ]
Reported-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
This patch fixes a bug in the overlap function which returned true if
one region ends exactly before the second region begins. This is no
overlap but the function returned true in that case.
Cc: stable@kernel.org
Reported-by: Andrew Randrianasulu <randrik@mail.ru>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
The current code is not very careful when it builds reference
dma_debug_entries which get passed to hash_bucket_find(). But since this
function changed to a best-fit algorithm these entries have to be more
acurate. This patch adds this higher level of accuracy.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
If we don't check for sg_call_ents the hash_bucket_find function might
still return the wrong dma_debug_entry for sg mappings.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
This patch makes the driver_filter function more readable by
reorganizing the code. The removal of a code code block to an upper
indentation level makes hard-to-read line-wraps unnecessary.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
There is no need to disable/enable irqs on each loop iteration. Just
disable irqs for the whole time the loop runs.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
The pr_* macros are shorter than the old printk(KERN_ ...) variant.
Change the dma-debug code to use the new macros and save a few
unnecessary line breaks. If lines don't break the source code can also
be grepped more easily.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
This patch changes the recent updates to dma-debug to conform with
coding style guidelines of Linux and the -tip tree.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Last patch series introduced some new comment which does not fit the
Kernel comment style guidelines. Fix it with this patch.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Some device drivers map the same physical address multiple times to a
dma address. Without an IOMMU this results in the same dma address being
put into the dma-debug hash multiple times. With a first-fit match in
hash_bucket_find() this function may return the wrong dma_debug_entry.
This can result in false positive warnings. This patch fixes it by
changing the first-fit behavior of hash_bucket_find() into a best-fit
algorithm.
Reported-by: Torsten Kaiser <just.for.lkml@googlemail.com>
Reported-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Cc: lethal@linux-sh.org
Cc: just.for.lkml@googlemail.com
Cc: hancockrwd@gmail.com
Cc: jens.axboe@oracle.com
Cc: bharrosh@panasas.com
Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: <stable@kernel.org>
LKML-Reference: <20090605104132.GE24836@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
This patch adds the dma-api/driver_filter file to debugfs. The root user
can write a driver name into this file to see only dma-api errors for
that particular driver in the kernel log. Writing an empty string to
that file disables the driver filter.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
This patch adds the state variables for the driver filter and a function
to check if the filter is enabled and matches to the current device. The
check is built into the err_printk function.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
DMA-mapping.txt says that debug_dma_sync_sg family must be called with
the _same_ one you passed into the dma_map_sg call, it should _NOT_ be
the 'count' value _returned_ from the dma_map_sg call.
debug_dma_sync_sg_for_cpu and debug_dma_sync_sg_for_device can't
handle this properly; they need to use the sg_mapped_ents in struct
dma_debug_entry as debug_dma_unmap_sg() does.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
debug_dma_map_sg() and debug_dma_unmap_sg() use length in struct
scatterlist while debug_dma_sync_sg_for_cpu() and
debug_dma_sync_sg_for_device() use dma_length. This causes bugs
warnings on some IOMMU implementations since these values are not
same; the length doesn't represent the dma length.
We always need to use sg_dma_len() accessor to get the dma length of a
scatterlist entry.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Architectures might not have dma_address in struct scatterlist (PARISC
doesn't). Directly accessing to dma_address in struct scatterlist is
wrong; we need to use sg_dma_address() accesssor instead.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
The feature needs some more work because the notfier which is used to
check for pending allocations is called before the device drivers
->remove() function. Therefore this feature reports false positives.
A real fix for this issue is to introduce a new notifier event which sent
_after_ the driver has deinitialized itself. That will done for the next
kernel version.
[ Impact: reduce the scope of CONFIG_DMA_API_DEBUG=y checks ]
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Cc: iommu@lists.linux-foundation.org
LKML-Reference: <1240576557-22442-1-git-send-email-joerg.roedel@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
We use a static value for the number of dma_debug_entries. It can be
overwritten by a kernel command line option.
Some IOMMUs (e.g. GART) can't set an appropriate value by a kernel
command line option because they can't know such value until they
finish initializing up their hardware.
This patch adds dma_debug_resize_entries() enables IOMMUs to adjust
the number of dma_debug_entries anytime.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Acked-by: Joerg Roedel <joerg.roedel@amd.com>
Cc: fujita.tomonori@lab.ntt.co.jp
Cc: akpm@linux-foundation.org
LKML-Reference: <20090415182234R.fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Fix printk format warnings in dma-debug:
lib/dma-debug.c:645: warning: format '%016llx' expects type 'long long unsigned int', but argument 6 has type 'dma_addr_t'
lib/dma-debug.c:662: warning: format '%016llx' expects type 'long long unsigned int', but argument 6 has type 'dma_addr_t'
lib/dma-debug.c:676: warning: format '%016llx' expects type 'long long unsigned int', but argument 6 has type 'dma_addr_t'
lib/dma-debug.c:686: warning: format '%016llx' expects type 'long long unsigned int', but argument 6 has type 'dma_addr_t'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Dan Williams
Alexander Duyck
Shuah Khan
Ming Lei
Dan Carpenter
Jakub Kicinski
Alan Stern
Thomas Jarosch
Paul Gortmaker
Neil Horman
Stanislaw Gruszka
Arnd Bergmann
Dan Carpenter
Joerg Roedel
Thiago Farina
Krzysztof Halasa
Ingo Molnar
Shaun Ruffell
André Goddard Rosa
Kyle McMartin
FUJITA Tomonori
Randy Dunlap