| Age | Commit message (Collapse) | Author |
|---|
|
|
|
When listen(2) is called on an unbound socket, the socket is
automatically bound to a random free port with the local address
set to INADDR_ANY.
PiperOrigin-RevId: 286305906
|
|
|
|
PiperOrigin-RevId: 286299056
|
|
|
|
It can take more than 10 seconds when running under --race.
PiperOrigin-RevId: 286296060
|
|
|
|
PiperOrigin-RevId: 286281274
|
|
|
|
This change removes the requirement that a new routing table be provided when a
router or prefix discovery event happens so that an updated routing table may
be provided to the stack at a later time from the event.
This change is to address the use case where the netstack integrator may need to
obtain a lock before providing updated routes in response to the events above.
As an example, say we have an integrator that performs the below two operations
operations as described:
A. Normal route update:
1. Obtain integrator lock
2. Update routes in the integrator
3. Call Stack.SetRouteTable with the updated routes
3.1. Obtain Stack lock
3.2. Update routes in Stack
3.3. Release Stack lock
4. Release integrator lock
B. NDP event triggered route update:
1. Obtain Stack lock
2. Call event handler
2.1. Obtain integrator lock
2.2. Update routes in the integrator
2.3. Release integrator lock
2.4. Return updated routes to update Stack
3. Update routes in Stack
4. Release Stack lock
A deadlock may occur if a Normal route update was attemped at the same time an
NDP event triggered route update was attempted. With threads T1 and T2:
1) T1 -> A.1, A.2
2) T2 -> B.1
3) T1 -> A.3 (hangs at A.3.1 since Stack lock is taken in step 2)
4) T2 -> B.2 (hangs at B.2.1 since integrator lock is taken in step 1)
Test: Existing tests were modified to not provide or expect routing table
changes in response to Router and Prefix discovery events.
PiperOrigin-RevId: 286274712
|
|
|
|
PiperOrigin-RevId: 286259750
|
|
|
|
This change makes sure that test variables are captured before running tests
in parallel, and removes unneeded buffered channel allocations. This change also
removes unnecessary timeouts.
PiperOrigin-RevId: 286255066
|
|
|
|
PiperOrigin-RevId: 286249699
|
|
|
|
PiperOrigin-RevId: 286248378
|
|
|
|
This avoids conflicting definitions of GetSocketPairs() in outer namespace when
multiple such cc files are complied for one binary.
PiperOrigin-RevId: 286243045
|
|
|
|
Python 3 tools must be listed in exec_tools for genrules.
PiperOrigin-RevId: 286241702
|
|
Several jobs were finished in this patch:
1, provide functions to get/set fpcr/fpsr/vregs
2, support lazy-fpsimd-context-switch in el1
Signed-off-by: Bin Lu <bin.lu@arm.com>
|
|
|
|
PiperOrigin-RevId: 286083614
|
|
|
|
PiperOrigin-RevId: 286068605
|
|
|
|
This is called after fork, so it must be nosplit.
Updates #1408
PiperOrigin-RevId: 286053054
|
|
|
|
PiperOrigin-RevId: 286051631
|
|
|
|
PiperOrigin-RevId: 286042427
|
|
|
|
PiperOrigin-RevId: 286003946
|
|
Remove introduced CPUNumMin config and hard-code it as 2.
|
|
|
|
PiperOrigin-RevId: 285968611
|
|
* Add `--cpu-num-min` flag to control minimum CPUs
* Only lower CPU count
* Fix comments
|
|
|
|
These comments provided nothing, and have been copy-pasted into all
implementations. The code is clear without them.
I considered also removing the "handle implements handler.handle" comments, but
will let those stay for now.
PiperOrigin-RevId: 285876428
|
|
|
|
PiperOrigin-RevId: 285874181
|
|
|
|
Bitshift operators with signed int is supported in Go 1.13.
PiperOrigin-RevId: 285853622
|
|
|
|
Add checks for input arguments, file type, permissions, etc. that match
the Linux implementation. A call to get/setxattr that passes all the
checks will still currently return EOPNOTSUPP. Actual support will be
added in following commits.
Only allow user.* extended attributes for the time being.
PiperOrigin-RevId: 285835159
|
|
Copy up parent when binding UDS on overlayfs is supported in commit
02ab1f187cd24c67b754b004229421d189cee264.
But the using of copyUp in overlayBind will cause sentry stuck, reason
is dead lock in renameMu.
1 [Process A] Invoke a Unix socket bind operation
renameMu is hold in fs.(*Dirent).genericCreate by process A
2 [Process B] Invoke a read syscall on /proc/task/mounts
waitng on Lock of renameMu in fs.(*MountNamespace).FindMount
3 [Process A] Continue Unix socket bind operation
wating on RLock of renameMu in fs.copyUp
Root cause is recursive reading lock of reanmeMu in bind call trace,
if there are writing lock between the two reading lock, then deadlock
occured.
Fixes #1397
|
|
After the finalizer optimize in 76039f895995c3fe0deef5958f843868685ecc38
commit, clientFile needs to closed before finalizer release it.
The clientFile is not closed if it is created via
gofer.(*inodeOperations).Bind, this will cause fd leak which is hold
by gofer process.
Fixes #1396
Signed-off-by: Yong He <chenglang.hy@antfin.com>
Signed-off-by: Jianfeng Tan <henry.tjf@antfin.com>
|
|
When application is not cgroups-aware, it can spawn excessive threads
which often defaults to CPU number.
Introduce a opt-in flag that will set CPU number accordingly to CPU
quota (if available).
Fixes #1391
|
