Technically declaring a prototype with an empty declaration has been deprecated since the early days of C, but we never got a warning for it. C2x will apparently be introducing a breaking change if you are using this type of declarator, so Clang 15 has started issuing a warning with -pedantic. Although not apparently a problem for any of the compiler we build on, if feels like the right thing is to properly adhere to the C standard and use (void).
This PR is to fix the compilation warnings and errors generated by the latest
complier toolchain, and to add a new runner of the latest toolchain for daily CI.
## Fix various compilation warnings and errors
1) jemalloc.c
COMPILER: clang-14 with FORTIFY_SOURCE
WARNING:
```
src/jemalloc.c:1028:7: warning: suspicious concatenation of string literals in an array initialization; did you mean to separate the elements with a comma? [-Wstring-concatenation]
"/etc/malloc.conf",
^
src/jemalloc.c:1027:3: note: place parentheses around the string literal to silence warning
"\"name\" of the file referenced by the symbolic link named "
^
```
REASON: the compiler to alert developers to potential issues with string concatenation
that may miss a comma,
just like #9534 which misses a comma.
SOLUTION: use `()` to tell the compiler that these two line strings are continuous.
2) config.h
COMPILER: clang-14 with FORTIFY_SOURCE
WARNING:
```
In file included from quicklist.c:36:
./config.h:319:76: warning: attribute declaration must precede definition [-Wignored-attributes]
char *strcat(char *restrict dest, const char *restrict src) __attribute__((deprecated("please avoid use of unsafe C functions. prefer use of redis_strlcat instead")));
```
REASON: Enabling _FORTIFY_SOURCE will cause the compiler to use `strcpy()` with check,
it results in a deprecated attribute declaration after including <features.h>.
SOLUTION: move the deprecated attribute declaration from config.h to fmacro.h before "#include <features.h>".
3) networking.c
COMPILER: GCC-12
WARNING:
```
networking.c: In function ‘addReplyDouble.part.0’:
networking.c:876:21: warning: writing 1 byte into a region of size 0 [-Wstringop-overflow=]
876 | dbuf[start] = '$';
| ^
networking.c:868:14: note: at offset -5 into destination object ‘dbuf’ of size 5152
868 | char dbuf[MAX_LONG_DOUBLE_CHARS+32];
| ^
networking.c:876:21: warning: writing 1 byte into a region of size 0 [-Wstringop-overflow=]
876 | dbuf[start] = '$';
| ^
networking.c:868:14: note: at offset -6 into destination object ‘dbuf’ of size 5152
868 | char dbuf[MAX_LONG_DOUBLE_CHARS+32];
```
REASON: GCC-12 predicts that digits10() may return 9 or 10 through `return 9 + (v >= 1000000000UL)`.
SOLUTION: add an assert to let the compiler know the possible length;
4) redis-cli.c & redis-benchmark.c
COMPILER: clang-14 with FORTIFY_SOURCE
WARNING:
```
redis-benchmark.c:1621:2: warning: embedding a directive within macro arguments has undefined behavior [-Wembedded-directive] #ifdef USE_OPENSSL
redis-cli.c:3015:2: warning: embedding a directive within macro arguments has undefined behavior [-Wembedded-directive] #ifdef USE_OPENSSL
```
REASON: when _FORTIFY_SOURCE is enabled, the compiler will use the print() with
check, which is a macro. this may result in the use of directives within the macro, which
is undefined behavior.
SOLUTION: move the directives-related code out of `print()`.
5) server.c
COMPILER: gcc-13 with FORTIFY_SOURCE
WARNING:
```
In function 'lookupCommandLogic',
inlined from 'lookupCommandBySdsLogic' at server.c:3139:32:
server.c:3102:66: error: '*(robj **)argv' may be used uninitialized [-Werror=maybe-uninitialized]
3102 | struct redisCommand *base_cmd = dictFetchValue(commands, argv[0]->ptr);
| ~~~~^~~
```
REASON: The compiler thinks that the `argc` returned by `sdssplitlen()` could be 0,
resulting in an empty array of size 0 being passed to lookupCommandLogic.
this should be a false positive, `argc` can't be 0 when strings are not NULL.
SOLUTION: add an assert to let the compiler know that `argc` is positive.
6) sha1.c
COMPILER: gcc-12
WARNING:
```
In function ‘SHA1Update’,
inlined from ‘SHA1Final’ at sha1.c:195:5:
sha1.c:152:13: warning: ‘SHA1Transform’ reading 64 bytes from a region of size 0 [-Wstringop-overread]
152 | SHA1Transform(context->state, &data[i]);
| ^
sha1.c:152:13: note: referencing argument 2 of type ‘const unsigned char[64]’
sha1.c: In function ‘SHA1Final’:
sha1.c:56:6: note: in a call to function ‘SHA1Transform’
56 | void SHA1Transform(uint32_t state[5], const unsigned char buffer[64])
| ^
In function ‘SHA1Update’,
inlined from ‘SHA1Final’ at sha1.c:198:9:
sha1.c:152:13: warning: ‘SHA1Transform’ reading 64 bytes from a region of size 0 [-Wstringop-overread]
152 | SHA1Transform(context->state, &data[i]);
| ^
sha1.c:152:13: note: referencing argument 2 of type ‘const unsigned char[64]’
sha1.c: In function ‘SHA1Final’:
sha1.c:56:6: note: in a call to function ‘SHA1Transform’
56 | void SHA1Transform(uint32_t state[5], const unsigned char buffer[64])
```
REASON: due to the bug[https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80922], when
enable LTO, gcc-12 will not see `diagnostic ignored "-Wstringop-overread"`, resulting in a warning.
SOLUTION: temporarily set SHA1Update to noinline to avoid compiler warnings due
to LTO being enabled until the above gcc bug is fixed.
7) zmalloc.h
COMPILER: GCC-12
WARNING:
```
In function ‘memset’,
inlined from ‘moduleCreateContext’ at module.c:877:5,
inlined from ‘RM_GetDetachedThreadSafeContext’ at module.c:8410:5:
/usr/include/x86_64-linux-gnu/bits/string_fortified.h:59:10: warning: ‘__builtin_memset’ writing 104 bytes into a region of size 0 overflows the destination [-Wstringop-overflow=]
59 | return __builtin___memset_chk (__dest, __ch, __len,
```
REASON: due to the GCC-12 bug [https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96503],
GCC-12 cannot see alloc_size, which causes GCC to think that the actual size of memory
is 0 when checking with __glibc_objsize0().
SOLUTION: temporarily set malloc-related interfaces to `noinline` to avoid compiler warnings
due to LTO being enabled until the above gcc bug is fixed.
## Other changes
1) Fixed `ps -p [pid]` doesn't output `<defunct>` when using procps 4.x causing `replication
child dies when parent is killed - diskless` test to fail.
2) Add a new fortify CI with GCC-13 and ubuntu-lunar docker image.
## Issue
When we use GCC-12 later or clang 9.0 later to build with `-D_FORTIFY_SOURCE=3`,
we can see the following buffer overflow:
```
=== REDIS BUG REPORT START: Cut & paste starting from here ===
6263:M 06 Apr 2023 08:59:12.915 # Redis 255.255.255 crashed by signal: 6, si_code: -6
6263:M 06 Apr 2023 08:59:12.915 # Crashed running the instruction at: 0x7f03d59efa7c
------ STACK TRACE ------
EIP:
/lib/x86_64-linux-gnu/libc.so.6(pthread_kill+0x12c)[0x7f03d59efa7c]
Backtrace:
/lib/x86_64-linux-gnu/libc.so.6(+0x42520)[0x7f03d599b520]
/lib/x86_64-linux-gnu/libc.so.6(pthread_kill+0x12c)[0x7f03d59efa7c]
/lib/x86_64-linux-gnu/libc.so.6(raise+0x16)[0x7f03d599b476]
/lib/x86_64-linux-gnu/libc.so.6(abort+0xd3)[0x7f03d59817f3]
/lib/x86_64-linux-gnu/libc.so.6(+0x896f6)[0x7f03d59e26f6]
/lib/x86_64-linux-gnu/libc.so.6(__fortify_fail+0x2a)[0x7f03d5a8f76a]
/lib/x86_64-linux-gnu/libc.so.6(+0x1350c6)[0x7f03d5a8e0c6]
src/redis-server 127.0.0.1:25111(+0xd5e80)[0x557cddd3be80]
src/redis-server 127.0.0.1:25111(feedReplicationBufferWithObject+0x78)[0x557cddd3c768]
src/redis-server 127.0.0.1:25111(replicationFeedSlaves+0x1a4)[0x557cddd3cbc4]
src/redis-server 127.0.0.1:25111(+0x8721a)[0x557cddced21a]
src/redis-server 127.0.0.1:25111(call+0x47a)[0x557cddcf38ea]
src/redis-server 127.0.0.1:25111(processCommand+0xbf4)[0x557cddcf4aa4]
src/redis-server 127.0.0.1:25111(processInputBuffer+0xe6)[0x557cddd22216]
src/redis-server 127.0.0.1:25111(readQueryFromClient+0x3a8)[0x557cddd22898]
src/redis-server 127.0.0.1:25111(+0x1b9134)[0x557cdde1f134]
src/redis-server 127.0.0.1:25111(aeMain+0x119)[0x557cddce5349]
src/redis-server 127.0.0.1:25111(main+0x466)[0x557cddcd6716]
/lib/x86_64-linux-gnu/libc.so.6(+0x29d90)[0x7f03d5982d90]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0x80)[0x7f03d5982e40]
src/redis-server 127.0.0.1:25111(_start+0x25)[0x557cddcd7025]
```
The main reason is that when FORTIFY_SOURCE is enabled, GCC or clang will enhance some
common functions, such as `strcpy`, `memcpy`, `fgets`, etc, so that they can detect buffer
overflow errors and stop program execution, thus improving the safety of the program.
We use `zmalloc_usable_size()` everywhere to use memory blocks, but that is an abuse since the
malloc_usable_size() isn't meant for this kind of use, it is for diagnostics only. That is also why the
behavior is flaky when built with _FORTIFY_SOURCE, the compiler can sense that we reach outside
the allocated block and SIGABRT.
### Solution
If we need to use the additional memory we got, we need to use a dummy realloc with `alloc_size` attribute
and no inlining, (see `extend_to_usable`) to let the compiler see the large of memory we need to use.
This can either be an implicit call inside `z*usable` that returns the size, so that the caller doesn't have any
other worry, or it can be a normal zmalloc call which means that if the caller wants to use
zmalloc_usable_size it must also use extend_to_usable.
### Changes
This PR does the following:
1) rename the current z[try]malloc_usable family to z[try]malloc_internal and don't expose them to users outside zmalloc.c,
2) expose a new set of `z[*]_usable` family that use z[*]_internal and `extend_to_usable()` implicitly, the caller gets the
size of the allocation and it is safe to use.
3) go over all the users of `zmalloc_usable_size` and convert them to use the `z[*]_usable` family if possible.
4) in the places where the caller can't use `z[*]_usable` and store the real size, and must still rely on zmalloc_usable_size,
we still make sure that the allocation used `z[*]_usable` (which has a call to `extend_to_usable()`) and ignores the
returning size, this way a later call to `zmalloc_usable_size` is still safe.
[4] was done for module.c and listpack.c, all the others places (sds, reply proto list, replication backlog, client->buf)
are using [3].
Co-authored-by: Oran Agra <oran@redislabs.com>
micro optimizations, giving the hints that the returned addresses
are guaranteed to be unique. The alloc_size attribute gives an extra hint
about the source of the size, useful mostly for calloc-like calls or when there
are extra arguments.
This PR does 2 main things:
1) Add warning for suspected slow system clocksource setting. This is Linux specific.
2) Add a `--check-system` argument to redis which runs all system checks and prints a report.
## System checks
Add a command line option `--check-system` which runs all known system checks and provides
a report to stdout of which systems checks have failed with details on how to reconfigure the
system for optimized redis performance.
The `--system-check` mode exists with an appropriate error code after running all the checks.
## Slow clocksource details
We check the system's clocksource performance by running `clock_gettime()` in a loop and then
checking how much time was spent in a system call (via `getrusage()`). If we spend more than
10% of the time in the kernel then we print a warning. I verified that using the slow clock sources:
`acpi_pm` (~90% in the kernel on my laptop) and `xen` (~30% in the kernel on an ec2 `m4.large`)
we get this warning.
The check runs 5 system ticks so we can detect time spent in kernel at 20% jumps (0%,20%,40%...).
Anything more accurate will require the test to run longer. Typically 5 ticks are 50ms. This means
running the test on startup will delay startup by 50ms. To avoid this we make sure the test is only
executed in the `--check-system` mode.
For a quick startup check, we specifically warn if the we see the system is using the `xen` clocksource
which we know has bad performance and isn't recommended (at least on ec2). In such a case the
user should manually rung redis with `--check-system` to force the slower clocksource test described
above.
## Other changes in the PR
* All the system checks are now implemented as functions in _syscheck.c_.
They are implemented using a standard interface (see details in _syscheck.c_).
To do this I moved the checking functions `linuxOvercommitMemoryValue()`,
`THPIsEnabled()`, `linuxMadvFreeForkBugCheck()` out of _server.c_ and _latency.c_
and into the new _syscheck.c_. When moving these functions I made sure they don't
depend on other functionality provided in _server.c_ and made them use a standard
"check functions" interface. Specifically:
* I removed all logging out of `linuxMadvFreeForkBugCheck()`. In case there's some
unexpected error during the check aborts as before, but without any logging.
It returns an error code 0 meaning the check didn't not complete.
* All these functions now return 1 on success, -1 on failure, 0 in case the check itself
cannot be completed.
* The `linuxMadvFreeForkBugCheck()` function now internally calls `exit()` and not
`exitFromChild()` because the latter is only available in _server.c_ and I wanted to
remove that dependency. This isn't an because we don't need to worry about the
child process created by the test doing anything related to the rdb/aof files which
is why `exitFromChild()` was created.
* This also fixes parsing of other /proc/\<pid\>/stat fields to correctly handle spaces
in the process name and be more robust in general. Not that before this fix the rss
info in `INFO memory` was corrupt in case of spaces in the process name. To
recreate just rename `redis-server` to `redis server`, start it, and run `INFO memory`.
# Short description
The Redis extended latency stats track per command latencies and enables:
- exporting the per-command percentile distribution via the `INFO LATENCYSTATS` command.
**( percentile distribution is not mergeable between cluster nodes ).**
- exporting the per-command cumulative latency distributions via the `LATENCY HISTOGRAM` command.
Using the cumulative distribution of latencies we can merge several stats from different cluster nodes
to calculate aggregate metrics .
By default, the extended latency monitoring is enabled since the overhead of keeping track of the
command latency is very small.
If you don't want to track extended latency metrics, you can easily disable it at runtime using the command:
- `CONFIG SET latency-tracking no`
By default, the exported latency percentiles are the p50, p99, and p999.
You can alter them at runtime using the command:
- `CONFIG SET latency-tracking-info-percentiles "0.0 50.0 100.0"`
## Some details:
- The total size per histogram should sit around 40 KiB. We only allocate those 40KiB when a command
was called for the first time.
- With regards to the WRITE overhead As seen below, there is no measurable overhead on the achievable
ops/sec or full latency spectrum on the client. Including also the measured redis-benchmark for unstable
vs this branch.
- We track from 1 nanosecond to 1 second ( everything above 1 second is considered +Inf )
## `INFO LATENCYSTATS` exposition format
- Format: `latency_percentiles_usec_<CMDNAME>:p0=XX,p50....`
## `LATENCY HISTOGRAM [command ...]` exposition format
Return a cumulative distribution of latencies in the format of a histogram for the specified command names.
The histogram is composed of a map of time buckets:
- Each representing a latency range, between 1 nanosecond and roughly 1 second.
- Each bucket covers twice the previous bucket's range.
- Empty buckets are not printed.
- Everything above 1 sec is considered +Inf.
- At max there will be log2(1000000000)=30 buckets
We reply a map for each command in the format:
`<command name> : { `calls`: <total command calls> , `histogram` : { <bucket 1> : latency , < bucket 2> : latency, ... } }`
Co-authored-by: Oran Agra <oran@redislabs.com>
## Backgroud
As we know, after `fork`, one process will copy pages when writing data to these
pages(CoW), and another process still keep old pages, they totally cost more memory.
For redis, we suffered that redis consumed much memory when the fork child is serializing
key/values, even that maybe cause OOM.
But actually we find, in redis fork child process, the child process don't need to keep some
memory and parent process may write or update that, for example, child process will never
access the key-value that is serialized but users may update it in parent process.
So we think it may reduce COW if the child process release memory that it is not needed.
## Implementation
For releasing key value in child process, we may think we call `decrRefCount` to free memory,
but i find the fork child process still use much memory when we don't write any data to redis,
and it costs much more time that slows down bgsave. Maybe because memory allocator doesn't
really release memory to OS, and it may modify some inner data for this free operation, especially
when we free small objects.
Moreover, CoW is based on pages, so it is a easy way that we only free the memory bulk that is
not less than kernel page size. madvise(MADV_DONTNEED) can quickly release specified region
pages to OS bypassing memory allocator, and allocator still consider that this memory still is used
and don't change its inner data.
There are some buffers we can release in the fork child process:
- **Serialized key-values**
the fork child process never access serialized key-values, so we try to free them.
Because we only can release big bulk memory, and it is time consumed to iterate all
items/members/fields/entries of complex data type. So we decide to iterate them and
try to release them only when their average size of item/member/field/entry is more
than page size of OS.
- **Replication backlog**
Because replication backlog is a cycle buffer, it will be changed quickly if redis has heavy
write traffic, but in fork child process, we don't need to access that.
- **Client buffers**
If clients have requests during having the fork child process, clients' buffer also be changed
frequently. The memory includes client query buffer, output buffer, and client struct used memory.
To get child process peak private dirty memory, we need to count peak memory instead
of last used memory, because the child process may continue to release memory (since
COW used to only grow till now, the last was equivalent to the peak).
Also we're adding a new `current_cow_peak` info variable (to complement the existing
`current_cow_size`)
Co-authored-by: Oran Agra <oran@redislabs.com>
1. Add `redis-server test all` support to run all tests.
2. Add redis test to daily ci.
3. Add `--accurate` option to run slow tests for more iterations (so that
by default we run less cycles (shorter time, and less prints).
4. Move dict benchmark to REDIS_TEST.
5. fix some leaks in tests
6. make quicklist tests run on a specific fill set of options rather than huge ranges
7. move some prints in quicklist test outside their loops to reduce prints
8. removing sds.h from dict.c since it is now used in both redis-server and
redis-cli (uses hiredis sds)
* Add better control of malloc_usable_size() usage.
* Use malloc_usable_size on alpine libc daily job.
* Add no-malloc-usable-size daily jobs.
* Fix zmalloc(0) when HAVE_MALLOC_SIZE is undefined.
In order to align with the jemalloc behavior, this should never return
NULL or OOM panic.
This commit has two aspects:
1) improve memory reporting for all the places that use sdsAllocSize to compute
memory used by a string, in this case it'll include the internal fragmentation.
2) reduce the need for realloc calls by making the sds implicitly take over
the internal fragmentation of the block it allocated.
jemalloc 5 doesn't immediately release memory back to the OS, instead there's a decaying
mechanism, which doesn't work when there's no traffic (no allocations).
this is most evident if there's no traffic after flushdb, the RSS will remain high.
1) enable jemalloc background purging
2) explicitly purge in flushdb
When HAVE_MALLOC_SIZE is false, each call to zrealloc causes used_memory
to increase by PREFIX_SIZE more than it should, due to mis-matched
accounting between the original zmalloc (which includes PREFIX size in
its increment) and zrealloc (which misses it from its decrement).
I've also supplied a command-line test to easily demonstrate the
problem. It's not wired into the test framework, because I don't know
TCL so I'm not sure how to automate it.
A) slave buffers didn't count internal fragmentation and sds unused space,
this caused them to induce eviction although we didn't mean for it.
B) slave buffers were consuming about twice the memory of what they actually needed.
- this was mainly due to sdsMakeRoomFor growing to twice as much as needed each time
but networking.c not storing more than 16k (partially fixed recently in 237a38737).
- besides it wasn't able to store half of the new string into one buffer and the
other half into the next (so the above mentioned fix helped mainly for small items).
- lastly, the sds buffers had up to 30% internal fragmentation that was wasted,
consumed but not used.
C) inefficient performance due to starting from a small string and reallocing many times.
what i changed:
- creating dedicated buffers for reply list, counting their size with zmalloc_size
- when creating a new reply node from, preallocate it to at least 16k.
- when appending a new reply to the buffer, first fill all the unused space of the
previous node before starting a new one.
other changes:
- expose mem_not_counted_for_evict info field for the benefit of the test suite
- add a test to make sure slave buffers are counted correctly and that they don't cause eviction
this reduces the extra 8 bytes we save before each pointer.
but more importantly maybe, it makes the valgrind runs to be more similiar
to our normal runs.
note: the change in malloc_stats struct in server.h is to eliminate an name conflict.
structs that are not typedefed are resolved from a separate name space.
other fixes / improvements:
- LUA script memory isn't taken from zmalloc (taken from libc malloc)
so it can cause high fragmentation ratio to be displayed (which is false)
- there was a problem with "fragmentation" info being calculated from
RSS and used_memory sampled at different times (now sampling them together)
other details:
- adding a few more allocator info fields to INFO and MEMORY commands
- improve defrag test to measure defrag latency of big keys
- increasing the accuracy of the defrag test (by looking at real grag info)
this way we can use an even lower threshold and still avoid false positives
- keep the old (total) "fragmentation" field unchanged, but add new ones for spcific things
- add these the MEMORY DOCTOR command
- deduct LUA memory from the rss in case of non jemalloc allocator (one for which we don't "allocator active/used")
- reduce sampling rate of the rss and allocator info
Obtaining the RSS (Resident Set Size) info is slow in Linux and OSX.
This slowed down the generation of the INFO 'memory' section.
Since the RSS does not require to be a real-time measurement, we
now sample it with server.hz frequency (10 times per second by default)
and use this value both to show the INFO rss field and to compute the
fragmentation ratio.
Practically this does not make any difference for memory profiling of
Redis but speeds up the INFO call significantly.
The previous implementation of zmalloc.c was not able to handle out of
memory in an application-specific way. It just logged an error on
standard error, and aborted.
The result was that in the case of an actual out of memory in Redis
where malloc returned NULL (In Linux this actually happens under
specific overcommit policy settings and/or with no or little swap
configured) the error was not properly logged in the Redis log.
This commit fixes this problem, fixing issue #509.
Now the out of memory is properly reported in the Redis log and a stack
trace is generated.
The approach used is to provide a configurable out of memory handler
to zmalloc (otherwise the default one logging the event on the
standard output is used).
calloc is more effecient than malloc+memset when the system uses mmap to
allocate memory. mmap always returns zeroed memory so the memset can be
avoided. The threshold to use mmap is 16k in osx libc and 128k in bsd
libc and glibc. The kernel can lazily allocate the pages, this reduces
memory usage when we have a page table or hash table that is mostly
empty.
This change is most visible when you start a new redis instance with vm
enabled. You'll see no increased memory usage no matter how big your
page table is.
networking related stuff moved into networking.c
moved more code
more work on layout of source code
SDS instantaneuos memory saving. By Pieter and Salvatore at VMware ;)
cleanly compiling again after the first split, now splitting it in more C files
moving more things around... work in progress
split replication code
splitting more
Sets split
Hash split
replication split
even more splitting
more splitting
minor change
