Summary:
This is a preparatory phase to make the refcount usuable on Windows. For more
details, see D24716801 (e50725d2cb)
Reviewed By: chadaustin
Differential Revision: D24764568
fbshipit-source-id: 1e8c6ab00d4c1ec79c347fd5ae7167b2ce1dff68
Summary:
As of right now, opendir is the expensive callbacks due to fetching the sizes
for all the files in a directory. This strategy however breaks down when
timeouts are added as very large directories would trigger the timeout, not
because EdenFS is having a hard time reaching Mononoke, but because of
bandwidth limitation.
To avoid this issue, we need to have a per-file timeout and thus makes opendir
just trigger the futures, but not wait on them. The waiting bit will happen
at readdir time which will also enable having a timeout per file.
The one drawback to this is that the ObjectFetchContext that is passed in by
opendir cannot live long enough to be used in the size future. For now, we can
use a null context, but a proper context will need to be passed in, in the
future.
Reviewed By: wez
Differential Revision: D24895089
fbshipit-source-id: e10ceae2f7c49b4a006b15a34f85d06a2863ae3a
Summary:
The documentation for PrjFillDirEntryBuffer states that if no entries could be
added, then the ERROR_INSUFFICIENT_BUFFER errors need to be returned as is, the
code didn't do that.
Reviewed By: chadaustin
Differential Revision: D24764566
fbshipit-source-id: d6411822eac71b2f9aa7cf07858d09115767cc59
Summary:
The EdenFS codebase uses folly/logging/xlog to log, but we were still relying
on glog for the various CHECK macros. Since xlog also contains equivalent CHECK
macros, let's just rely on them instead.
This is mostly codemodded + arc lint + various fixes to get it compile.
Reviewed By: chadaustin
Differential Revision: D24871174
fbshipit-source-id: 4d2a691df235d6dbd0fbd8f7c19d5a956e86b31c
Summary:
One of the issue that EdenFS on Windows is currently facing is around
invalidation during an update. In effect, EdenFS is over invalidating, which
causes update to be slower than it should be, as well as EdenFS recursively
triggering ProjectedFS callbacks during invalidation. Both of these are a
sub-par UX.
The reason this issue exist is multi-faceted. First, the update code follows
the "kPreciseInodeNumberMemory" path which enforces that a directory that is
present in the overlay needs to be invalidated, even if it isn't materialized.
The second reason is that no reclamation is done for the overlay, combine the
two and you get an update that gets both slower over time and will issue
significantly more invalidation that is needed.
Solving this is a bit involved. We could for instance start by reclaiming
inodes from the overlay, but this wouldn't be effective as we use the fact that
an inode is present in the overlay as a way to know that the file is cached in
the overlay. If we reclaim from the overlay we simply won't be invalidating
enough and some files will be out of date.
It turns out that we already have a mechanism to track what is cached by the
kernel: the fuse refcount. On Linux/macOS, everytime an inode is returned to
the kernel, this refcount incremented, and the kernel then notifies us when it
forgot about it, at which point the refcount can be decremented. On Windows,
the rules are a bit different, and a simple flag is sufficient: set when we
write a placeholder on disk (either during a directory listing, or when
ProjectedFS asks for it), and unset at invalidation time during update. There
is however a small snag in this plan. On Linux, the refcount starts at 0 when
EdenFS starts as a mount/unmount will clear all the kernel references on the
inodes. On Windows, the placeholder aren't disappearing when EdenFS dies or is
stopped, so we need a way to scan the working copy when EdenFS starts to know
which inodes should be loaded (an UnloadedInode really).
The astute reader will have noticed that this last part is effectively a
O(materialized) operation that needs to happen at startup, which would be
fairly expensive in itself. It turns out that we really don't have choice and
we need to do it regardless due to Windows not disallowing writes to the
working copy when EdenFS is stopped, and thus for EdenFS to be aware of the
actual state of the working copy, it needs to scan it at startup...
The first step in doing all of this is to simply rename the various places that
uses "fuse refcount" to "fs refcount" which is what this diff does.
Reviewed By: chadaustin
Differential Revision: D24716801
fbshipit-source-id: e9e6ccff14c454e9f2626fab23daeb3930554b1a
Summary:
Futures are by default run inline, meaning that when the previous future is
completed, the future will run in the same context as the previous one. In the
case where the previous one is completed by another thread setting up its
promise, the future will be completed in the context of that other thread.
In most cases, this is OK, in others, this can cause a deadlock. And this is
exactly what we're seeing here. When a file is renamed concurrently to an `hg
update`, the inode the rename operates on might not be loaded, and thus both
update and the rename callback will race to load that inode. When update wins
that race, the rename callback will wait on a promise that update will then
set. When that happens, the rest of the rename callback will be run in the
update context, but that will in turn cause update to try to re-acquire the
rename lock that it already holds...
To fix this, we need to make sure that the rename callback doesn't run inline.
Reviewed By: chadaustin
Differential Revision: D24657422
fbshipit-source-id: 23b08765afae7bda4a628f0c23675bff9f486b6b
Summary:
Dealing with non-owned path in futures is mind bending and can lead to use
after free fairly easily if not careful. It turns out that this code is
actually a bit buggy and in some cases will attempt to use a path that is
already freed. Since the callsite doesn't need to hold onto the paths, let's
just move them, which resolves the issuue.
Reviewed By: chadaustin
Differential Revision: D24657423
fbshipit-source-id: 47bbaccf18cd86e53860491e3cbfeadb4363499c
Summary:
Instead of logging in the Dispatcher, move strace logging to
FuseChannel where it can be standardized for all FUSE request types.
Reviewed By: wez
Differential Revision: D24035838
fbshipit-source-id: c84d8c27b62f9944e2d26a35a7ed7bbbeeb5bf0e
Summary:
Now that prjfs/EdenDispatcher is no longer directly tied to ProjectedFS (sort
of, this is still a bit WIP), we can move it out of the prjfs directory onto
the inodes one. This allows breaking the circular dependency cycle mentioned in
the previous diff where prjfs and inodes depend on each other.
For now, this is merely a copy/paste of the code enclosed in big #ifdef _WIN32,
we might be able to do better later, but for now this is properly good enough.
Reviewed By: wez
Differential Revision: D23857539
fbshipit-source-id: 77c620bac1656d01d7daee4dbf8b10694a589751
Summary:
None of these were used, let's remove them.
ps: I thought we had a system to detect unused headers and lint about them?
Reviewed By: chadaustin
Differential Revision: D23465783
fbshipit-source-id: c21a34c9838db29f4fd0057d3be4e0fcb527cd6d
Summary:
As previous diffs in the stack show there were at least one place in the
codebase which used incorrect object context logger and that resulted in "blind
spots" in undesired file fetches logging i.e. undesired file fetches were
logged, but neither pid nor cmd-line was logged.
There are quite a few places in the codebase that use null
object fetch context, and threading the correct object fetch context to all of
them might be hard. Threading the context is a bit annoying, so it would be good to know something like "EdenDispatcher code is responsible for most of the blind spots, so let's thread the correct context there first". Or it would be equally good to know that none of the null object context are responsible for blind spots.
This diff might help us decide where we need to thread real object fetch context
first. Instead of passing null object fetch context let's pass null object
fetch context with causeDetail field. This field will be logged to scuba (see
BackingStoreLogger::logImport code), and instead of getting "Unknown" interface
we'll get e.g. "Unknown - EdenDispatcher::create", and that would highlight
where we need to thread the context.
A note about implementation - getNullContextWithCauseDetail returns a raw pointer
which is expected to be static i.e. it should work similarly to current
getNullContext implementation. It's quite a hack, but allows us to get rid of
memory allocations (we'd have one memory allocation per place in the code where
getNullContextWithCauseDetail). Let me know if you are ok with this hack.
Reviewed By: kmancini
Differential Revision: D23422526
fbshipit-source-id: e576bba9fc09e160fc42771c7589cdd1694d93c0
Summary:
Scuba logging that tracks undesired file fetches has some blind spots i.e. a
lot of fetches have null pid and null cmd line. This diff tries to fix another
part of the problem.
TreeInode::getOrLoadChild() has TODO `pass a fetch context down through
getOrLoadChild to track this load`. This diff fixes this TODO, and also starts
to pass context from EdenDispatcher:lookup method.
Note that it adds quite a lot of new `ObjectFetchContext::getNullContext()`
calls, and potentially those might be responsible for blind spots in logging.
I'll try to address this problem in the next diffs.
Reviewed By: kmancini
Differential Revision: D23418218
fbshipit-source-id: 319d7436494d8dce3580289aae9963aa13bfc191
Summary:
Setting up, tearing down, and querying RequestContext has some
overhead that I would like to avoid in the inner FUSE loop, so replace
RequestData with a single class that's heap-allocated at the start of
a request and is guaranteed to survive until the request ends, and is
otherwise explicitly passed where it's needed.
Reviewed By: kmancini
Differential Revision: D22712310
fbshipit-source-id: fc30d0b0f7e22b39306b857194ea07a913110b0f
Summary:
Avoid the cost of dynamically querying whether we are in a FUSE
request handler or not by passing a flag.
Reviewed By: kmancini
Differential Revision: D22710480
fbshipit-source-id: 010bb8efee8074441aa20aab0eb12277452c5252
Summary:
Avoid the cost of dynamically querying whether we are in a FUSE
request handler or not by passing a flag.
Reviewed By: kmancini
Differential Revision: D22710452
fbshipit-source-id: 818035b72b793fa895147d9df3bb668d5b9c55f3
Summary:
Avoid the cost of dynamically querying whether we are in a FUSE
request handler or not by passing a flag.
Reviewed By: kmancini
Differential Revision: D22710422
fbshipit-source-id: 65b0737ad5f8ca74d12f2c657691d3751df4aa54
Summary:
Avoid the cost of dynamically querying whether we are in a FUSE
request handler or not by passing a flag.
Reviewed By: genevievehelsel
Differential Revision: D22710397
fbshipit-source-id: 7c62f45dfc227416c91070842a349b9d0c626cba
Summary: This commit makes `readdir()` to correct send its `RequestContext` to `EdenDispatcher` method so underlying code can know the current request being processed is from FUSE.
Reviewed By: xavierd
Differential Revision: D21821949
fbshipit-source-id: f41ba912fedbfc040e3c9267aad25e7f33f8e912
Summary: The method name is symlink, not rmdir.
Reviewed By: genevievehelsel
Differential Revision: D22291297
fbshipit-source-id: 5dc37b053e06c965fd47df79990fc40adc097f87
Summary:
In following changes I will be threading ObjectFetchContext into the backing
store importing process, since this will start to be used more outside of the
ObjectStore, I am moving this class into its own files.
Reviewed By: chadaustin
Differential Revision: D22022488
fbshipit-source-id: 1a291fea6e0fd56855936962363dfc9f6de8533d
Summary:
This diff makes `lookup` to use `RequestData` as `ObjectFetchContext` in `getattr` calls.
This will make sure we correctly record backing store fetches in `eden top`
Reviewed By: chadaustin
Differential Revision: D21780969
fbshipit-source-id: 468e2fadcebf4a00477bc5de434e6c658b99d1ce
Summary: This diff makes `InodeBase::stat()` to be able to accept `ObjectFetchContext` as a parameter.
Reviewed By: chadaustin
Differential Revision: D21780883
fbshipit-source-id: 9b1db2e2268cb98663bcf902ea61897da593ea05
Summary: This diff starts to use `RequestData` as `ObjectFetchContext` in our `read()` methods. This ensures EdenFS could track backing store fetches happened in FUSE requests.
Reviewed By: chadaustin
Differential Revision: D21792503
fbshipit-source-id: 9509a1bc8f28100a0dfe196e312c4785c7842345
Summary:
FUSE_HANDLE_KILLPRIV expects that any write() call is handled by
clearing the setuid and setgid bits in the userspace. To avoid
implementing that behavior, disallow setting setuid or setgid in the
first place.
Reviewed By: xavierd
Differential Revision: D21333703
fbshipit-source-id: eb084ee8b00afe74c0da26e41c32c2cb742723da
Summary:
The operation originally wanted to operate on the fuse `Attr`
structure which we don't have on Windows, so I repurposed the
`InodeBase::getattr` into `InodeBase::stat` and moved the conversion
of `struct stat` to `Dispatcher::Attr` to the `EdenDispatcher::getattr`
method (and a couple of other adhoc places that were doing a similar
conversion).
Reviewed By: chadaustin
Differential Revision: D20562459
fbshipit-source-id: 6b538110038352e9b5590fcb5ff5c33fe84ac1d8
Summary: Pass a valid ObjectFetchContext down into certain untracked requests.
Reviewed By: chadaustin
Differential Revision: D20243575
fbshipit-source-id: e7112c3bab1265803a26130c4d72905c25f2e729
Summary:
When importing an object from the backing store failed, we would not
fail lookup requests (e.g. lstat) so that the file could still be
removed with `rm`. Unfortunately that data would get cached, even if
the error was transient. This diff returns a correct inode number for
failed requests and limits the amount of time these fake stat()
results are cached.
Reviewed By: wez
Differential Revision: D19655425
fbshipit-source-id: 1a44ace9217f1ab3f69a2c5a6edc89d0f9c4a941
Summary:
When we implemented FUSE_NO_OPEN_SUPPORT and FUSE_NO_OPENDIR_SUPPORT,
we forgot to remove the FileHandle and FileHandleBase code.
Reviewed By: pkaush
Differential Revision: D17991710
fbshipit-source-id: dfeb26d512f017cef7710929ccff1f6940cf8641
Summary:
This diff passes the free space information through from
the overlay lock file descriptor back up to the kernel when the
filesystem stats are requested.
This makes the user experience with eg: Finder on macOS nicer.
Reviewed By: chadaustin
Differential Revision: D17255859
fbshipit-source-id: ed9f3b9fd386c5706539879513854a9cd1550d8a
Summary:
Update the copyright & license headers in C++ files to reflect the
relicensing to GPLv2
Reviewed By: wez
Differential Revision: D15487078
fbshipit-source-id: 19f24c933a64ecad0d3a692d0f8d2a38b4194b1d
Summary:
Eden requires no state in its directory handles, so tell the kernel it
doesn't need to send opendir() and releasedir() requests, provided it
has FUSE_NO_OPENDIR_SUPPORT.
Reviewed By: strager
Differential Revision: D13594734
fbshipit-source-id: ebd4b69f4efcd1428a69024c4bdffb1ae455fa40
Summary:
Implement `flush()` and `fsyncdir()` in `EdenDispatcher` and explicitly return
ENOSYS. We intentionally do not need to do any work for these methods.
Providing our an implementation that returns ENOSYS avoids the error log
message in the default `Dispatcher` implementation about the these calls being
unimplemented.
Reviewed By: chadaustin
Differential Revision: D14453245
fbshipit-source-id: 71efe6de6af73a5d705dace0f3439ba2466a50a8
Summary:
Replace Future::onError with Future::thenError:
* to remove ambiguous typing
* to ensure that the executor is not lost and the returned Future is still bound to an executor
See:
https://fb.workplace.com/groups/fbcode/permalink/2002251863144976/
for details.
Reviewed By: yfeldblum
Differential Revision: D13784772
fbshipit-source-id: 1d3ede848b7d31c7a197a21b4ff2b31e840040a5
Summary:
We think that it shouldn't really be needed to perform
the prefetch call during lookup; for file inodes it doesn't buy
us much, and it should only really help for readdir.
This removes the prefetch call from lookup, instead prefetching
upon the first readdir() of a loaded TreeInode.
Reviewed By: simpkins
Differential Revision: D12896022
fbshipit-source-id: 0209eb64bd522daf5f7461dffccd1312d32a1554
Summary:
The fuse opcodes are defined as an enum so we have to use
the relatively coarse and indirect apple vs linux preprocessor
checks in the maps for the opcode names.
The osxfuse implementation branched off from the 7.19 fuse
implementation, so add a light dusting of some preprocessor
checks around enabling the performance optimization features
we desire on Linux.
We also need to relax the compile time check for the min
fuse version; I've constrained this to be apple specific,
although I suppose it wouldn't hurt to make it more broadly
applicable.
Reviewed By: chadaustin
Differential Revision: D13480145
fbshipit-source-id: 010ac114e22ea942dfcebf1105cb1f01b766f297
Summary:
Eden no longer tracks any state in file handles, and has no plans to in the future.
Therefore, remove all related code.
Reviewed By: strager
Differential Revision: D13354307
fbshipit-source-id: 341d081f64c6c8fb2b4b1b5a5ff42f2cc7d38039
Summary:
Now that all file access in Eden is stateless, we no longer need to handle open() or release().
If the kernel advertises FUSE_NO_OPEN_SUPPORT, return ENOSYS from open().
Reviewed By: simpkins
Differential Revision: D13325759
fbshipit-source-id: 38486848f27ffeb005f74407888e94d891496f98
Summary:
Previously, a file handle must have been held for the entirety of a write operation. That is no
longer true. Stop looking up file handles on write.
Reviewed By: strager
Differential Revision: D13325662
fbshipit-source-id: 9ae31b467d17d633c388917d18098e6e5a620b89
Summary:
Stop holding a reference count to the TreeInode while a directory
handle is open. This allows eden to shut down while a directory handle
is open.
Reviewed By: strager
Differential Revision: D13287701
fbshipit-source-id: a24f32a1ac40b6c19bc5864aa5f5785f3016361b
Summary:
Send readdir requests to TreeInode. This may not sound like a good
idea: the FUSE documentation suggests that stateful directory handles
are required to implement correct readdir semantics under concurrent
deletes and renames. However, the 63-bit offset value is treated as a
cookie that is passed from one readdir call into the next, and 63 bits
should be sufficient to implement readdir concurrent with
rename/unlink. So move readdir's implementation into TreeInode in
preparation for the complete removal of TreeInodeDirHandle.
Reviewed By: strager
Differential Revision: D13287664
fbshipit-source-id: c0d615675edd9b83353534468a69b89068bba923
Summary:
FUSE_NO_OPEN_SUPPORT is better than ATOMIC_O_TRUNC for Eden's use
case. Remove the code that pretended we might support ATOMIC_O_TRUNC
again someday.
(Note: this ignores all push blocking failures!)
Reviewed By: strager
Differential Revision: D13163382
fbshipit-source-id: 948d701571a8d2977da3d2532fdc9538c5011636
Summary:
The new blob cache wants to know, given a request, whether the blob is
expected to be needed or not. The answer, in general, is yes if the
request came from Thrift and no if it came from FUSE, because the kernel
will cache the result of the request in its own page and dentry caches.
Propagate this information through FileInode.
Reviewed By: strager
Differential Revision: D12813838
fbshipit-source-id: 7a359686149cd4daff41630c94085b680c448c4f
