dump_iterate() dumps blocks in a loop. Eventually, get_next_block()
returns "no more". We then call dump_completed(). But we neglect to
break the loop! Broken in commit 4c7e251a.
Because of that, we dump the last block again. This attempts to write
to s->fd, which fails if we're lucky. The error makes dump_iterate()
return failure. It's the only way it can ever return.
Theoretical: if we're not so lucky, something else has opened something
for writing and got the same fd. dump_iterate() then keeps looping,
messing up the something else's output, until a write fails, or the
process mercifully terminates.
The obvious fix is to restore the return lost in commit 4c7e251a. But
the root cause of the bug is needlessly opaque loop control. Replace it
by a clean do ... while loop.
This makes the badly chosen return values of get_next_block() more
visible. Cleaning that up is outside the scope of this bug fix.
Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Functions shouldn't return an error code and an Error object at the same time.
Turn all these functions that returning Error object to void.
We also judge if a function success or fail by reference to the local_err.
Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
The code calls dump_error() on error, and even passes it a suitable
message. However, the message is thrown away, and its callers pass
up only success/failure. All qmp_dump_guest_memory() can do is set
a generic error.
Propagate the errors properly, so qmp_dump_guest_memory() can return
a more useful error.
Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
In dump_init(), when failure occurs, need notice about 'fd' and memory
mapping. So call dump_cleanup() for it (need let all initializations at
front).
Also simplify dump_cleanup(): remove redundant 'ret' and redundant 'fd'
checking.
Signed-off-by: Chen Gang <gang.chen.5i5j@gmail.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Make DumpState and endian conversion routines available for arch-specific dump
code by moving into dump.h. DumpState will be needed by arch-specific dump
code to access target endian information from DumpState->ArchDumpInfo. Also
break the dependency of dump.h from stubs/dump.c by creating a separate
dump-arch.h.
This patch doesn't change any functionality.
Signed-off-by: Bharata B Rao <bharata@linux.ibm.com>
[ rebased on top of current master branch,
renamed endian helpers to cpu_to_dump{16,32,64},
pass a DumpState * argument to endian helpers,
Greg Kurz <gkurz@linux.vnet.ibm.com> ]
Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
[agraf: fix to apply]
Signed-off-by: Alexander Graf <agraf@suse.de>
We can (and should) rely on the fact that s->flag_compress is exactly one
of DUMP_DH_COMPRESSED_ZLIB, DUMP_DH_COMPRESSED_LZO, and
DUMP_DH_COMPRESSED_SNAPPY.
This is ensured by the QMP schema and dump_init() in combination.
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
qmp_dump_guest_memory()
dump_init()
lzo_init() <---------+
create_kdump_vmcore() |
write_dump_pages() |
get_len_buf_out() |
lzo_init() ------+
This patch doesn't change the fact that lzo_init() is called for every
LZO-compressed dump, but it makes get_len_buf_out() more focused (single
responsibility).
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
The specific ELF architecture (d_machine) carries Too Much Information
(TM) for deciding between create_header32() and create_header64(), use
"d_class" instead (ELFCLASS32 vs. ELFCLASS64).
This change adapts write_dump_header() to write_elf_loads(), dump_begin()
etc. that also rely on the ELF class of the target for bitness selection.
Considering the current targets that support dumping, cpu_get_dump_info()
works as follows:
- target-s390x/arch_dump.c: (EM_S390, ELFCLASS64) only
- target-ppc/arch_dump.c (EM_PPC64, ELFCLASS64) only
- target-i386/arch_dump.c: sets (EM_X86_64, ELFCLASS64) vs. (EM_386,
ELFCLASS32) keying off the same Long Mode Active flag.
Hence no observable change.
Approximately-suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Use TARGET_PAGE_SIZE and ~TARGET_PAGE_MASK instead.
"DumpState.page_size" has type "size_t", whereas TARGET_PAGE_SIZE has type
"int". TARGET_PAGE_MASK is of type "int" and has negative value. The patch
affects the implicit type conversions as follows:
- create_header32() and create_header64(): assigned to "block_size", which
has type "uint32_t". No change.
- get_next_page(): "block->target_start", "block->target_end" and "addr"
have type "hwaddr" (uint64_t).
Before the patch,
- if "size_t" was "uint64_t", then no additional conversion was done as
part of the usual arithmetic conversions,
- If "size_t" was "uint32_t", then it was widened to uint64_t as part of
the usual arithmetic conversions,
for the remainder and addition operators.
After the patch,
- "~TARGET_PAGE_MASK" expands to ~~((1 << TARGET_PAGE_BITS) - 1). It
has type "int" and positive value (only least significant bits set).
That's converted (widened) to "uint64_t" for the bit-ands. No visible
change.
- The same holds for the (addr + TARGET_PAGE_SIZE) addition.
- write_dump_pages():
- TARGET_PAGE_SIZE passed as argument to a bunch of functions that all
have prototypes. No change.
- When incrementing "offset_data" (of type "off_t"): given that we never
build for ILP32_OFF32 (see "-D_FILE_OFFSET_BITS=64" in configure),
"off_t" is always "int64_t", and we only need to consider:
- ILP32_OFFBIG: "size_t" is "uint32_t".
- before: int64_t += uint32_t. Page size converted to int64_t for
the addition.
- after: int64_t += int32_t. No change.
- LP64_OFF64: "size_t" is "uint64_t".
- before: int64_t += uint64_t. Offset converted to uint64_t for the
addition, then the uint64_t result is converted to int64_t for
storage.
- after: int64_t += int32_t. Same as the ILP32_OFFBIG/after case.
No visible change.
- (size_out < s->page_size) comparisons, and (size_out = s->page_size)
assignment:
- before: "size_out" is of type "size_t", no implicit conversion for
either operator.
- after: TARGET_PAGE_SIZE (of type "int" and positive value) is
converted to "size_t" (for the relop because the latter is
one of "uint32_t" and "uint64_t"). No visible change.
- dump_init():
- DIV_ROUND_UP(DIV_ROUND_UP(s->max_mapnr, CHAR_BIT), s->page_size): The
innermost "DumpState.max_mapnr" field has type uint64_t, which
propagates through all implicit conversions at hand:
#define DIV_ROUND_UP(n,d) (((n) + (d) - 1) / (d))
regardless of the page size macro argument's type. In the outer macro
replacement, the page size is converted from uint32_t and int32_t
alike to uint64_t.
- (tmp * s->page_size) multiplication: "tmp" has size "uint64_t"; the
RHS is converted to that type from uint32_t and int32_t just the same
if it's not uint64_t to begin with.
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Just use TARGET_PAGE_BITS.
"DumpState.page_shift" used to have type "uint32_t", while the replacement
TARGET_PAGE_BITS has type "int". Since "DumpState.page_shift" was only
used as bit shift counts in the paddr_to_pfn() and pfn_to_paddr() macros,
this is safe.
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Currently, the function
- defines and populates an auto variable of type MakedumpfileHeader
- allocates and zeroes a buffer of size MAX_SIZE_MDF_HEADER (4096)
- copies the former into the latter (covering an initial portion of the
latter)
Fill in the MakedumpfileHeader structure in its final place (the alignment
is OK because the structure lives at the address returned by g_malloc0()).
Approximately-suggested-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
The "mh.signature" array field has size 16, and is zeroed by the preceding
memset(). MAKEDUMPFILE_SIGNATURE expands to a string literal with string
length 12 (size 13). There's no need to measure the length of
MAKEDUMPFILE_SIGNATURE at runtime, nor for the extra zero-filling of
"mh.signature" with strncpy().
Use memcpy() with MIN(sizeof, sizeof) for robustness (which is an integer
constant expression, evaluable at compile time.)
Approximately-suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
In qmp_dump_guest_memory(), the error must be clear on entry, and we
always bail out after setting it, directly or via dump_init().
Therefore, both error_is_set() are always false. Drop them.
DumpState member errp is now write-only. Drop it, too.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
'query-dump-guest-memory-capability' is used to query the available formats for
'dump-guest-memory'. The output of the command will be like:
-> { "execute": "query-dump-guest-memory-capability" }
<- { "return": { "formats":
["elf", "kdump-zlib", "kdump-lzo", "kdump-snappy"] }
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Make monitor command 'dump-guest-memory' be able to dump in kdump-compressed
format. The command's usage:
dump [-p] protocol [begin] [length] [format]
'format' is used to specified the format of vmcore and can be:
1. 'elf': ELF format, without compression
2. 'kdump-zlib': kdump-compressed format, with zlib-compressed
3. 'kdump-lzo': kdump-compressed format, with lzo-compressed
4. 'kdump-snappy': kdump-compressed format, with snappy-compressed
Without 'format' being set, it is same as 'elf'. And if non-elf format is
specified, paging and filter is not allowed.
Note:
1. The kdump-compressed format is readable only with the crash utility and
makedumpfile, and it can be smaller than the ELF format because of the
compression support.
2. The kdump-compressed format is the 6th edition.
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
functions are used to write page to vmcore. vmcore is written page by page.
page desc is used to store the information of a page, including a page's size,
offset, compression format, etc.
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
DataCache is used to store data temporarily, then the data will be written to
vmcore. These functions will be called later when writing data of page to
vmcore.
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
functions are used to write 1st and 2nd dump_bitmap of kdump-compressed format,
which is used to indicate whether the corresponded page is existed in vmcore.
1st and 2nd dump_bitmap are same, because dump level is specified to 1 here.
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
the functions are used to write header of kdump-compressed format to vmcore.
Header of kdump-compressed format includes:
1. common header: DiskDumpHeader32 / DiskDumpHeader64
2. sub header: KdumpSubHeader32 / KdumpSubHeader64
3. extra information: only elf notes here
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
add some members to DumpState that will be used in writing vmcore in
kdump-compressed format. some of them, like page_size, will be initialized
in the patch.
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
the function can be used by write_elf32_notes/write_elf64_notes to write notes
to a buffer. If fd_write_vmcore is used, write_elf32_notes/write_elf64_notes
will write elf notes to vmcore directly. Instead, if buf_write_note is used,
elf notes will be written to opaque->note_buf at first.
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Function is used to write vmcore in flatten format. In flatten format, data is
written block by block, and in front of each block, a struct
MakedumpfileDataHeader is stored there to indicate the offset and size of the
data block.
struct MakedumpfileDataHeader {
int64_t offset;
int64_t buf_size;
};
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
flatten format will be used when writing kdump-compressed format. The format is
also used by makedumpfile, you can refer to the following URL to get more
detailed information about flatten format of kdump-compressed format:
http://sourceforge.net/projects/makedumpfile/
The two functions here are used to write start flat header and end flat header
to vmcore, and they will be called later when flatten format is used.
struct MakedumpfileHeader stored at the head of vmcore is used to indicate the
vmcore is in flatten format.
struct MakedumpfileHeader {
char signature[16]; /* = "makedumpfile" */
int64_t type; /* = 1 */
int64_t version; /* = 1 */
};
And struct MakedumpfileDataHeader, with offset and buf_size set to -1, is used
to indicate the end of vmcore in flatten format.
struct MakedumpfileDataHeader {
int64_t offset; /* = -1 */
int64_t buf_size; /* = -1 */
};
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
write_elf32_notes/wirte_elf64_notes use fd_write_vmcore to write elf notes to
vmcore. Adding parameter "WriteCoreDumpFunction f" makes it available to choose
the method of writing elf notes
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
WriteCoreDumpFunction is a function pointer that points to the function used to
write content in "buf" into core file, so "buf" should be const-qualify.
Signed-off-by: Qiao Nuohan <qiaonuohan@cn.fujitsu.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
We should check for error with s->note_size
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
While the machine is paused, in guest_phys_blocks_append() we register a
one-shot MemoryListener, solely for the initial collection of the valid
guest-physical memory ranges that happens at listener registration time.
For each range that is reported to guest_phys_blocks_region_add(), we
attempt to merge the range with the preceding one.
Ranges can only be joined if they are contiguous in both guest-physical
address space, and contiguous in host virtual address space.
The "maximal" ranges that remain in the end constitute the guest-physical
memory map that the dump will be based on.
Related RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=981582
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
The vmcore must use physical addresses that are visible to the guest, not
addresses that point into linear RAMBlocks. As first step, introduce the
list type into which we'll collect the physical mappings in effect at the
time of the dump.
Related RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=981582
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Even a trusted & clean-state guest can map more memory than what it was
given. Since the vmcore contains RAMBlocks, mapping sizes should be
clamped to RAMBlock sizes. Otherwise such oversized mappings can exceed
the entire file size, and ELF parsers might refuse even the valid portion
of the PT_LOAD entry.
Related RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=981582
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Move next_cpu from CPU_COMMON to CPUState.
Move first_cpu variable to qom/cpu.h.
gdbstub needs to use CPUState::env_ptr for now.
cpu_copy() no longer needs to save and restore cpu_next.
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
[AF: Rebased, simplified cpu_copy()]
Signed-off-by: Andreas Färber <afaerber@suse.de>
Pass any Error out into dump_init() and have it actually stop on errors.
Whether it is unsupported on a certain CPU can be checked by looking for
a NULL CPUClass::get_memory_mapping field.
Reviewed-by: Luiz Capitulino <lcapitulino@redhat.com>
[AF: Reverted changes to CPU loops]
Signed-off-by: Andreas Färber <afaerber@suse.de>
Instead of calling cpu_synchronize_state() for each CPU, call the
existing cpu_synchronize_all_states() helper.
Reviewed-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Headers in include/exec/ are for the deepest innards of QEMU,
they should almost never be included directly.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
target_phys_addr_t is unwieldly, violates the C standard (_t suffixes are
reserved) and its purpose doesn't match the name (most target_phys_addr_t
addresses are not target specific). Replace it with a finger-friendly,
standards conformant hwaddr.
Outstanding patchsets can be fixed up with the command
git rebase -i --exec 'find -name "*.[ch]"
| xargs s/target_phys_addr_t/hwaddr/g' origin
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
fd_write_vmcore() will indefinitely spin for a non-blocking
file-descriptor that would block. However, if the fd is non-blocking,
how does it make sense to spin?
Change this behavior to return an error instead.
Note that this can only happen with an fd provided by a management
application. The fd opened internally by dump-guest-memory is blocking.
While there, also fix 'writen_size' variable name.
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
dump.c was recently added to the code. It unconditionally
includes sys/procfs which is not available with MinGW (w32, w64).
It looks like this file is not needed at all (tested on Linux),
so I removed it completely.
Some other include statements are also redundant because they are
already included in qemu-common, therefore they were removed, too.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
This simplifies things, because they will only be included for softmmu
targets and because the stubs are taken out-of-line in separate files,
which in the future could even be compiled only once.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
