| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Change snprintf to scnprintf. There are generally two cases where using
snprintf causes problems.
1) Uses of size += snprintf(buf, SIZE - size, fmt, ...)
In this case, if snprintf would have written more characters than what the
buffer size (SIZE) is, then size will end up larger than SIZE. In later
uses of snprintf, SIZE - size will result in a negative number, leading
to problems. Note that size might already be too large by using
size = snprintf before the code reaches a case of size += snprintf.
2) If size is ultimately used as a length parameter for a copy back to user
space, then it will potentially allow for a buffer overflow and information
disclosure when size is greater than SIZE. When the size is used to index
the buffer directly, we can have memory corruption. This also means when
size = snprintf... is used, it may also cause problems since size may become
large. Copying to userspace is mitigated by the HARDENED_USERCOPY kernel
configuration.
The solution to these issues is to use scnprintf which returns the number of
characters actually written to the buffer, so the size variable will never
exceed SIZE.
Signed-off-by: Silvio Cesare <[email protected]>
Cc: Timur Tabi <[email protected]>
Cc: Nicolin Chen <[email protected]>
Cc: Mark Brown <[email protected]>
Cc: Xiubo Li <[email protected]>
Cc: Fabio Estevam <[email protected]>
Cc: Dan Carpenter <[email protected]>
Cc: Kees Cook <[email protected]>
Cc: Will Deacon <[email protected]>
Cc: Greg KH <[email protected]>
Signed-off-by: Willy Tarreau <[email protected]>
Acked-by: Nicolin Chen <[email protected]>
Reviewed-by: Kees Cook <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
There is a potential NULL pointer dereference in case devm_kzalloc()
fails and returns NULL.
Fix this by adding a NULL check on rt5514_dsp.
This issue was detected with the help of Coccinelle.
Fixes: 6eebf35b0e4a ("ASoC: rt5514: add rt5514 SPI driver")
Cc: [email protected]
Signed-off-by: Gustavo A. R. Silva <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
Change snprintf to scnprintf. There are generally two cases where using
snprintf causes problems.
1) Uses of size += snprintf(buf, SIZE - size, fmt, ...)
In this case, if snprintf would have written more characters than what the
buffer size (SIZE) is, then size will end up larger than SIZE. In later
uses of snprintf, SIZE - size will result in a negative number, leading
to problems. Note that size might already be too large by using
size = snprintf before the code reaches a case of size += snprintf.
2) If size is ultimately used as a length parameter for a copy back to user
space, then it will potentially allow for a buffer overflow and information
disclosure when size is greater than SIZE. When the size is used to index
the buffer directly, we can have memory corruption. This also means when
size = snprintf... is used, it may also cause problems since size may become
large. Copying to userspace is mitigated by the HARDENED_USERCOPY kernel
configuration.
The solution to these issues is to use scnprintf which returns the number of
characters actually written to the buffer, so the size variable will never
exceed SIZE.
Signed-off-by: Silvio Cesare <[email protected]>
Cc: Liam Girdwood <[email protected]>
Cc: Mark Brown <[email protected]>
Cc: Dan Carpenter <[email protected]>
Cc: Kees Cook <[email protected]>
Cc: Will Deacon <[email protected]>
Cc: Greg KH <[email protected]>
Signed-off-by: Willy Tarreau <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
Fix typo which causes headphone no sound while using BCLK
as PLL source.
Signed-off-by: Shuming Fan <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
cpu and platform are optional components in DAI links. For example
codec-codec links usually have no platform set.
Call snd_soc_find_component only if the name or of_node of
a cpu or platform is set. Otherwise it will return NULL and
soc_init_dai_link bails out immediately with -EPROBE_DEFER,
meaning registering a card with NULL cpu or platform in DAI links
can never succeed.
Fixes: 8780cf1142a5 ("ASoC: soc-core: defer card probe until all component is added to list")
Signed-off-by: Matthias Reichl <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
There are some use cases where you're checking for a lot of things on a
card and it makes sense that you might end up trying to call
snd_soc_find_component() without either a name or an of_node. Currently
in that case we try to dereference the name and crash but it's more
useful to allow the caller to just treat that as a case where we don't
find anything, that error handling will already exist.
Inspired by a patch from Ajit Pandey fixing some callers.
Fixes: 8780cf1142a5 ("ASoC: soc-core: defer card probe until all component is added to list")
Reported-by: Pierre-Louis Bossart <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
snd_soc_init_platform initializes pointers to snd_soc_dai_link which is
statically allocated and it does this by devm_kzalloc. In the event of
an EPROBE_DEFER the memory will be freed and the pointers are left
dangling. snd_soc_init_platform sees the dangling pointers and assumes
they are pointing to initialized memory and does not reallocate them on
the second probe attempt which results in a use after free bug since
devm has freed the memory from the first probe attempt.
Since the intention for snd_soc_dai_link->platform is that it can be set
statically by the machine driver we need to respect the pointer in the
event we did not set it but still catch dangling pointers. The solution
is to add a flag to track whether the pointer was dynamically allocated
or not.
Signed-off-by: Curtis Malainey <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
soc_init_dai_link() calls soc_find_component() which needs
to be within client_mutex lock. Add client_mutex lock around
soc_init_dai_link() in snd_soc_register_card() to avoid
lockdep warning.
Fixes: 8780cf1142a5 ("ASoC: soc-core: defer card probe until all component is added to list")
Reported-by: Kuninori Morimoto <[email protected]>
Signed-off-by: Rohit kumar <[email protected]>
Signed-off-by: Ajit Pandey <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
move the codec PLL to rt5682_codec_init, because codec only need to config the clock source/PLL once.
As the result, remove the platform_clock_controls since no need to control clock anymore.
Signed-off-by: Shuming Fan <[email protected]>
Signed-off-by: Mac Chiang <[email protected]>
Acked-by: Pierre-Louis Bossart <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
DAI component probe is not called if it is not present
in component list during sound card registration.
Check if component is available in component list for
platform and cpu dai before soundcard registration.
Signed-off-by: Ajit Pandey <[email protected]>
Signed-off-by: Rohit kumar <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
snd_pcm_lib_malloc_pages() may fail, so let's check its status and
return its error code upstream.
Signed-off-by: Kangjie Lu <[email protected]>
Acked-by: Pierre-Louis Bossart <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
Cc: [email protected]
|
|
During the bootup of the kernel, the DAPM bias level is in the OFF
state. As soon as the DAPM framework kicks in it pushes the codec
into STANDBY state.
The probe function doesn't prepare the clock, and STANDBY state
does a clk_disable_unprepare() without checking the previous state.
This leads to an OOPS.
Not transitioning from an OFF state to the STANDBY state fixes the
problem.
Signed-off-by: b-ak <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
Cc: [email protected]
|
|
McASP can loose it's context when runtime_pm is disabled.
Save and restore the context when suspending and resuming the device.
Signed-off-by: Peter Ujfalusi <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
Inside function rt274_i2c_probe(), if regmap_read() function
returns -EINVAL, then local variable "val" leaves uninitialized
but used in if statement. This is potentially unsafe.
Signed-off-by: Yizhuo <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
The ADC mixer setting needs to restore to default value
after calibration.
Signed-off-by: Shuming Fan <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
After 'commit 5d32a66541c4 ("PCI/ACPI: Allow ACPI to be built without
CONFIG_PCI set")' dependencies on CONFIG_PCI that previously were
satisfied implicitly through dependencies on CONFIG_ACPI have to be
specified directly. This code relies on IOSF_MBI and IOSF_MBI depends
on PCI. For this reason, add a direct dependency on CONFIG_PCI to the
IOSF_MBI driver.
Fixes: 5d32a66541c46 ("PCI/ACPI: Allow ACPI to be built without CONFIG_PCI set")
Signed-off-by: Sinan Kaya <[email protected]>
Acked-by: Pierre-Louis Bossart <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
The problem is seen in the q6asm_dai_compr_set_params() function:
ret = q6asm_map_memory_regions(dir, prtd->audio_client, prtd->phys,
(prtd->pcm_size / prtd->periods),
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
prtd->periods);
In this code prtd->pcm_size is the buffer_size and prtd->periods comes
from params->buffer.fragments. If we allow the number of fragments to
be zero then it results in a divide by zero bug. One possible fix would
be to use prtd->pcm_count directly instead of using the division to
re-calculate it. But I decided that it doesn't really make sense to
allow zero fragments.
Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
We can't return directly if snd_dma_alloc_pages() fails; we first need
to free prtd->audio_client and prtd.
Fixes: 22930c79ac5c ("ASoC: qdsp6: q6asm-dai: Add support to compress offload")
Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
The q6asm_audio_client_alloc() doesn't return NULL, it returns error
pointers.
Fixes: 22930c79ac5c ("ASoC: qdsp6: q6asm-dai: Add support to compress offload")
Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
The q6asm_fe_dais[] array has MAX_SESSIONS (8) elements so the >
comparison should be >= or we access one element beyond the end of the
array.
Fixes: 22930c79ac5c ("ASoC: qdsp6: q6asm-dai: Add support to compress offload")
Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
We accidentally call mutex_unlock(&pcm512x->mutex); twice in a row.
I re-wrote the error handling to use "goto unlock;" instead of returning
directly. Hopefully, it makes the code a little simpler.
Fixes: 3500f1c589e9 ("ASoC: pcm512x: Implement the digital_mute interface")
Signed-off-by: Dan Carpenter <[email protected]>
Reviwed-by: Dimitris Papavasiliou <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
Changed License header from C to C++ style comment block.
Signed-off-by: Maruthi Srinivas Bayyavarapu <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
For some reason this field was set to zero when all other drivers use
.dynamic = 1 for front-ends. This change was tested on Dell XPS13 and
has no impact with the existing legacy driver. The SOF driver also works
with this change which enables it to override the fixed topology.
Signed-off-by: Rander Wang <[email protected]>
Acked-by: Pierre-Louis Bossart <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
Fixes: 33f8db9a89200c18 ("ASoC: xlnx: enable i2s driver build")
Signed-off-by: Geert Uytterhoeven <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
The intent was to print the address as a hexadecimal but there is an
extra "u" in the "0x%08ulx" format specification so it is displayed as
decimal.
Fixes: aef3b06ac697 ("[ALSA] SH7760 ASoC support")
Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
Headset codec is connected over PRIMARY_MI2S interface. Call
set_jack for codec associated with Primary Mi2s interface.
Also, set_jack to NULL when jack is freed.
Signed-off-by: Rohit kumar <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
|
|
|
|
We no longer have these options used anywhere.
Signed-off-by: Peter Ujfalusi <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
Use the new options for default audio support.
Signed-off-by: Peter Ujfalusi <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
No audio support is enabled by default via ASoC, disable the code as well.
Signed-off-by: Peter Ujfalusi <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
Use the new options for default audio support and enable Nokia n900 audio
as well.
Signed-off-by: Peter Ujfalusi <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
The codec selection config options for DMA365 has been changed.
Signed-off-by: Peter Ujfalusi <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
The MCBSP config option has been changed.
Signed-off-by: Peter Ujfalusi <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
The MCBSP config option has been changed.
Signed-off-by: Peter Ujfalusi <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
The sound/soc/davinci and omap directories are merged under sound/soc/ti.
Add myself as maintainer and update the OMAP audio support section with the
new path and file names.
Signed-off-by: Peter Ujfalusi <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
Create new directory to contain all Texas Instruments specific DAI,
platform and machine drivers instead of scattering them under davinci and
omap directories.
There is already inter dependency between the two directories becasue of
McASP (on dra7x it is serviced by sDMA, not EDMA).
With the upcoming AM654 we will need to introduce new platform driver for
UDMA and it does not fit under davinci, nor under omap.
With the move I have restructured the Kconfig to be more usable in the era
of simple-sound-card:
CPU DAIs can be selected individually and they will select the platform
driver they can be served with.
To avoid breakage, I have moved over deprecated Kconfig options so
defconfig builds will work without regression.
Signed-off-by: Peter Ujfalusi <[email protected]>
For sound/soc/{omap => ti}:
Tested-by: Jarkko Nikula <[email protected]>
Acked-by: Jarkko Nikula <[email protected]>
Acked-by: Tony Lindgren <[email protected]>
Signed-off-by: Mark Brown <[email protected]>
|
|
|
|
Merge misc fixes from Andrew Morton:
"11 fixes"
* emailed patches from Andrew Morton <[email protected]>:
scripts/spdxcheck.py: always open files in binary mode
checkstack.pl: fix for aarch64
userfaultfd: check VM_MAYWRITE was set after verifying the uffd is registered
fs/iomap.c: get/put the page in iomap_page_create/release()
hugetlbfs: call VM_BUG_ON_PAGE earlier in free_huge_page()
memblock: annotate memblock_is_reserved() with __init_memblock
psi: fix reference to kernel commandline enable
arch/sh/include/asm/io.h: provide prototypes for PCI I/O mapping in asm/io.h
mm/sparse: add common helper to mark all memblocks present
mm: introduce common STRUCT_PAGE_MAX_SHIFT define
alpha: fix hang caused by the bootmem removal
|
|
The spdxcheck script currently falls over when confronted with a binary
file (such as Documentation/logo.gif). To avoid that, always open files
in binary mode and decode line-by-line, ignoring encoding errors.
One tricky case is when piping data into the script and reading it from
standard input. By default, standard input will be opened in text mode,
so we need to reopen it in binary mode.
The breakage only happens with python3 and results in a
UnicodeDecodeError (according to Uwe).
Link: http://lkml.kernel.org/r/[email protected]
Fixes: 6f4d29df66ac ("scripts/spdxcheck.py: make python3 compliant")
Signed-off-by: Thierry Reding <[email protected]>
Reviewed-by: Jeremy Cline <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Jonathan Corbet <[email protected]>
Cc: Joe Perches <[email protected]>
Cc: Uwe Kleine-König <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
There is actually a space after "sp," like this,
ffff2000080813c8: a9bb7bfd stp x29, x30, [sp, #-80]!
Right now, checkstack.pl isn't able to print anything on aarch64,
because it won't be able to match the stating objdump line of a function
due to this missing space. Hence, it displays every stack as zero-size.
After this patch, checkpatch.pl is able to match the start of a
function's objdump, and is then able to calculate each function's stack
correctly.
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Qian Cai <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
Calling UFFDIO_UNREGISTER on virtual ranges not yet registered in uffd
could trigger an harmless false positive WARN_ON. Check the vma is
already registered before checking VM_MAYWRITE to shut off the false
positive warning.
Link: http://lkml.kernel.org/r/[email protected]
Cc: <[email protected]>
Fixes: 29ec90660d68 ("userfaultfd: shmem/hugetlbfs: only allow to register VM_MAYWRITE vmas")
Signed-off-by: Andrea Arcangeli <[email protected]>
Reported-by: [email protected]
Acked-by: Mike Rapoport <[email protected]>
Acked-by: Hugh Dickins <[email protected]>
Acked-by: Peter Xu <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
migrate_page_move_mapping() expects pages with private data set to have
a page_count elevated by 1. This is what used to happen for xfs through
the buffer_heads code before the switch to iomap in commit 82cb14175e7d
("xfs: add support for sub-pagesize writeback without buffer_heads").
Not having the count elevated causes move_pages() to fail on memory
mapped files coming from xfs.
Make iomap compatible with the migrate_page_move_mapping() assumption by
elevating the page count as part of iomap_page_create() and lowering it
in iomap_page_release().
It causes the move_pages() syscall to misbehave on memory mapped files
from xfs. It does not not move any pages, which I suppose is "just" a
perf issue, but it also ends up returning a positive number which is out
of spec for the syscall. Talking to Michal Hocko, it sounds like
returning positive numbers might be a necessary update to move_pages()
anyway though
(https://lkml.kernel.org/r/[email protected]).
I only hit this in tests that verify that move_pages() actually moved
the pages. The test also got confused by the positive return from
move_pages() (it got treated as a success as positive numbers were not
expected and not handled) making it a bit harder to track down what's
going on.
Link: http://lkml.kernel.org/r/[email protected]
Fixes: 82cb14175e7d ("xfs: add support for sub-pagesize writeback without buffer_heads")
Signed-off-by: Piotr Jaroszynski <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Cc: William Kucharski <[email protected]>
Cc: Darrick J. Wong <[email protected]>
Cc: Brian Foster <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
A stack trace was triggered by VM_BUG_ON_PAGE(page_mapcount(page), page)
in free_huge_page(). Unfortunately, the page->mapping field was set to
NULL before this test. This made it more difficult to determine the
root cause of the problem.
Move the VM_BUG_ON_PAGE tests earlier in the function so that if they do
trigger more information is present in the page struct.
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Yongkai Wu <[email protected]>
Acked-by: Michal Hocko <[email protected]>
Acked-by: Mike Kravetz <[email protected]>
Reviewed-by: William Kucharski <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
Found warning:
WARNING: EXPORT symbol "gsi_write_channel_scratch" [vmlinux] version generation failed, symbol will not be versioned.
WARNING: vmlinux.o(.text+0x1e0a0): Section mismatch in reference from the function valid_phys_addr_range() to the function .init.text:memblock_is_reserved()
The function valid_phys_addr_range() references
the function __init memblock_is_reserved().
This is often because valid_phys_addr_range lacks a __init
annotation or the annotation of memblock_is_reserved is wrong.
Use __init_memblock instead of __init.
Link: http://lkml.kernel.org/r/BLUPR13MB02893411BF12EACB61888E80DFAE0@BLUPR13MB0289.namprd13.prod.outlook.com
Signed-off-by: Yueyi Li <[email protected]>
Acked-by: Michal Hocko <[email protected]>
Reviewed-by: David Hildenbrand <[email protected]>
Acked-by: Mike Rapoport <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
The kernel commandline parameter named in CONFIG_PSI_DEFAULT_DISABLED
help text contradicts the documentation in kernel-parameters.txt, and
the code. Fix that.
Link: http://lkml.kernel.org/r/[email protected]
Fixes: e0c274472d ("psi: make disabling/enabling easier for vendor kernels")
Signed-off-by: Baruch Siach <[email protected]>
Acked-by: Johannes Weiner <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
Most architectures provide prototypes for the PCI I/O mapping operations
when asm/io.h is included but SH doesn't currently do that, leading to
for example warnings in sound/pci/hda/patch_ca0132.c when pci_iomap() is
used on current -next. Make SH more consistent with other architectures
by including asm-generic/pci_iomap.h in asm/io.h.
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Mark Brown <[email protected]>
Reported-by: kbuild test robot <[email protected]>
Cc: Geert Uytterhoeven <[email protected]>
Cc: Yoshinori Sato <[email protected]>
Cc: Rich Felker <[email protected]>
Cc: Arnd Bergmann <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
Presently the arches arm64, arm and sh have a function which loops
through each memblock and calls memory present. riscv will require a
similar function.
Introduce a common memblocks_present() function that can be used by all
the arches. Subsequent patches will cleanup the arches that make use of
this.
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Logan Gunthorpe <[email protected]>
Acked-by: Andrew Morton <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: Oscar Salvador <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
This define is used by arm64 to calculate the size of the vmemmap
region. It is defined as the log2 of the upper bound on the size of a
struct page.
We move it into mm_types.h so it can be defined properly instead of set
and checked with a build bug. This also allows us to use the same
define for riscv.
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Logan Gunthorpe <[email protected]>
Acked-by: Will Deacon <[email protected]>
Acked-by: Andrew Morton <[email protected]>
Acked-by: Ard Biesheuvel <[email protected]>
Acked-by: Catalin Marinas <[email protected]>
Cc: Arnd Bergmann <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|
|
The conversion of alpha to memblock as the early memory manager caused
boot to hang as described at [1].
The issue is caused because for CONFIG_DISCTONTIGMEM=y case,
memblock_add() is called using memory start PFN that had been rounded
down to the nearest 8Mb and it caused memblock to see more memory that
is actually present in the system.
Besides, memblock allocates memory from high addresses while bootmem was
using low memory, which broke the assumption that early allocations are
always accessible by the hardware.
This patch ensures that memblock_add() is using the correct PFN for the
memory start and forces memblock to use bottom-up allocations.
[1] https://lkml.org/lkml/2018/11/22/1032
Link: http://lkml.kernel.org/r/[email protected]
Reported-by: Meelis Roos <[email protected]>
Signed-off-by: Mike Rapoport <[email protected]>
Tested-by: Meelis Roos <[email protected]>
Cc: Richard Henderson <[email protected]>
Cc: Ivan Kokshaysky <[email protected]>
Cc: Matt Turner <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
|