Add the controller identifiers list CNS 0x13, available list of ctrls
in NVM Subsystem that may or may not be attached to namespaces.
In Identify Ctrl List of the CNS 0x12 and 0x13 no endian conversion
for the nsid field.
These two CNS values shows affect when there exists a Subsystem.
Added cond
In the documentation of the '-detached' param "be" and "not" has been
used side by side, fix that.
Signed-off-by: Gollu Appalanaidu
---
hw/nvme/ctrl.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/hw/nvme/ctrl.c b/hw/nvme/ctrl.c
index d513b022c4..21883e4b3c 100644
--- a/hw/
oVirt started to use now qemu:allocation-depth meta context.
In the past, we use base:allocation and reported NBD_STATE_HOLE
as a hole, and this broke in qemu 6.0.0.
Now we have NBD_STATE_HOLE from base:allocation, and flags == 0
from qemu:allocation-depth.
We map flags == 0 to EXTENT_BACKING (1
On machines with version > 6.0 replace a missing EUI-64 by a generated
value.
Signed-off-by: Heinrich Schuchardt
---
v3:
use 52-54-00-00-00-00-00-00 as starting values for generating
EUI-64s
v2:
new patch
---
docs/system/nvme.rst | 2 ++
hw/core/machine.c| 1 +
hw/nvm
The EUI-64 field is the only identifier for NVMe namespaces in UEFI device
paths. Add a new namespace property "eui64", that provides the user the
option to specify the EUI-64.
v3:
use 52-54-00-00-00-00-00-00 as starting values for generating
EUI-64s
Heinrich Schuchardt (2):
hw/
The EUI-64 field is the only identifier for NVMe namespaces in UEFI device
paths. Add a new namespace property "eui64", that provides the user the
option to specify the EUI-64.
Signed-off-by: Heinrich Schuchardt
Acked-by: Klaus Jensen
---
v3:
no change
v2:
fix typo %s/EUI64/EUI-6
On 6/14/21 9:23 PM, Heinrich Schuchardt wrote:
On 6/14/21 8:31 PM, Klaus Jensen wrote:
On Jun 12 01:46, Heinrich Schuchardt wrote:
On machines with version > 6.0 replace a missing EUI-64 by a generated
value.
Signed-off-by: Heinrich Schuchardt
---
v2:
new patch
---
docs/system/nvme.rst |
On 6/14/21 8:31 PM, Klaus Jensen wrote:
On Jun 12 01:46, Heinrich Schuchardt wrote:
On machines with version > 6.0 replace a missing EUI-64 by a generated
value.
Signed-off-by: Heinrich Schuchardt
---
v2:
new patch
---
docs/system/nvme.rst | 2 ++
hw/core/machine.c | 1 +
hw/nvme/ns.c
On Mon, 14 Jun 2021 at 15:18, Daniel P. Berrangé wrote:
>
> The following changes since commit 894fc4fd670aaf04a67dc7507739f914ff4bacf2:
>
> Merge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into
> staging (2021-06-11 09:21:48 +0100)
>
> are available in the Git repository a
On Sat, Jun 12, 2021 at 01:46:30AM +0200, Heinrich Schuchardt wrote:
> +``eui64``
> + Set the EUI-64 of the namespace. This will be reported as a "IEEE Extended
> + Unique Identifier" descriptor in the Namespace Identification Descriptor
> List.
This needs to match Identify Namespace's EUI64 fi
On Jun 14 11:48, Keith Busch wrote:
On Sat, Jun 12, 2021 at 01:46:30AM +0200, Heinrich Schuchardt wrote:
+``eui64``
+ Set the EUI-64 of the namespace. This will be reported as a "IEEE Extended
+ Unique Identifier" descriptor in the Namespace Identification Descriptor
List.
This needs to mat
On Sun, Jun 13, 2021 at 05:38:21PM +0800, Zhiwei Jiang wrote:
In the subject, s/waring/warning/
> when i compile this file with some error message
> ../block.c: In function ‘bdrv_replace_node_common’:
> ../block.c:4903:9: error: ‘to_cow_parent’ may be used
> uninitialized in this function [-Werro
On Jun 12 01:46, Heinrich Schuchardt wrote:
On machines with version > 6.0 replace a missing EUI-64 by a generated
value.
Signed-off-by: Heinrich Schuchardt
---
v2:
new patch
---
docs/system/nvme.rst | 2 ++
hw/core/machine.c| 1 +
hw/nvme/ns.c | 9 +
hw/nvme/nvme.h
On Jun 14 15:54, Jakub Jermář wrote:
An IRQ vector used by a completion queue cannot be deasserted without
first checking if the same vector does not need to stay asserted for
some other completion queue. To this end the controller structure is
extended by a counter of asserted completion queues.
On 6/14/21 6:03 PM, Philippe Mathieu-Daudé wrote:
> On 6/11/21 1:46 PM, Philippe Mathieu-Daudé wrote:
>> When the NVMe block driver was introduced (see commit bdd6a90a9e5,
>> January 2018), Linux VFIO_IOMMU_MAP_DMA ioctl was only returning
>> -ENOMEM in case of error. The driver was correctly handl
Gerd Hoffmann writes:
> Hi,
>
>> > +# @arch: module architecture.
>>
>> Semantics?
>>
>> Should this be enum SysEmuTarget?
>
> Probably, will check ...
>
>> > +# @opts: qemu opts implemented by module.
>>
>> Is this the name of a QemuOptsList?
>>
>> Since this isn't a list, a module can onl
On 6/12/21 1:46 AM, Heinrich Schuchardt wrote:
virt-6.0 must consider hw_compat_6_0.
Fixes: da7e13c00b59 ("hw: add compat machines for 6.1")
Signed-off-by: Heinrich Schuchardt
Reviewed-by: Cornelia Huck
This patch was applied to target-arm.next by Peter.
---
v2:
add missing Fixes:
Daniel P. Berrangé wrote:
> ping: anyone willing to give a review of this one
>
> On Wed, May 05, 2021 at 11:36:58AM +0100, Daniel P. Berrangé wrote:
>> The GDateTime APIs provided by GLib avoid portability pitfalls, such
>> as some platforms where 'struct timeval.tv_sec' field is still 'long'
>>
On 6/11/21 1:46 PM, Philippe Mathieu-Daudé wrote:
> When the NVMe block driver was introduced (see commit bdd6a90a9e5,
> January 2018), Linux VFIO_IOMMU_MAP_DMA ioctl was only returning
> -ENOMEM in case of error. The driver was correctly handling the
> error path to recycle its volatile IOVA mappi
* Daniel P. Berrangé (berra...@redhat.com) wrote:
> ping: anyone willing to review this
>
Reviewed-by: Dr. David Alan Gilbert
> On Wed, May 05, 2021 at 11:37:00AM +0100, Daniel P. Berrangé wrote:
> > The GDateTime APIs provided by GLib avoid portability pitfalls, such
> > as some platforms wher
On 6/11/21 1:46 PM, Philippe Mathieu-Daudé wrote:
> When the NVMe block driver was introduced (see commit bdd6a90a9e5,
> January 2018), Linux VFIO_IOMMU_MAP_DMA ioctl was only returning
> -ENOMEM in case of error. The driver was correctly handling the
> error path to recycle its volatile IOVA mappi
Hi,
> > +# @arch: module architecture.
>
> Semantics?
>
> Should this be enum SysEmuTarget?
Probably, will check ...
> > +# @opts: qemu opts implemented by module.
>
> Is this the name of a QemuOptsList?
>
> Since this isn't a list, a module can only ever provide one
> QemuOptsList. Sure
On 6/10/21 8:14 PM, Klaus Jensen wrote:
On Jun 10 13:46, Jakub Jermář wrote:
An IRQ vector used by a completion queue cannot be deasserted without
first checking if the same vector does not need to stay asserted for
some other completion queue.
Signed-off-by: Jakub Jermar
---
hw/nvme/ctrl.c |
We cannot reasonably test the main point of allow-other, which is to
allow users other than the current one to access the FUSE export,
because that would require access to sudo, which this test most likely
will not have. (Also, we would need to figure out some user/group that
is on the machine and
An IRQ vector used by a completion queue cannot be deasserted without
first checking if the same vector does not need to stay asserted for
some other completion queue. To this end the controller structure is
extended by a counter of asserted completion queues.
To prevent incrementing the counter f
Without the allow_other mount option, no user (not even root) but the
one who started qemu/the storage daemon can access the export. Allow
users to configure the export such that such accesses are possible.
When we do pass allow_other, we should also pass default_permissions,
because our export c
In order to support changing other attributes than the file size in
fuse_setattr(), we have to give each its own independent branch. This
also applies to the only attribute we do support right now.
Signed-off-by: Max Reitz
---
block/export/fuse.c | 20 +++-
1 file changed, 11 in
Allow changing the file mode, UID, and GID through SETATTR.
This only really makes sense with allow-other, though (because without
it, the effective access mode is fixed to be 0600 (u+rw) with qemu's
user being the file's owner), so changing these stat fields is not
allowed without allow-other.
S
Hi,
With the default mount options, FUSE mounts are not accessible to any
users but the one who did the mount, not even to root. To allow such
accesses, allow_other must be passed.
This is probably useful to some people (it certainly is to me, e.g. when
exporting some image as my normal user, an
The GDateTime APIs provided by GLib avoid portability pitfalls, such
as some platforms where 'struct timeval.tv_sec' field is still 'long'
instead of 'time_t'. When combined with automatic cleanup, GDateTime
often results in simpler code too.
Reviewed-by: Gerd Hoffmann
Signed-off-by: Daniel P. Be
The VM stop process has to flush outstanding I/O and this is a critical
failure scenario that is hard to diagnose. Add a probe point that
records the flush return code.
Reviewed-by: Connor Kuehl
Reviewed-by: Dr. David Alan Gilbert
Signed-off-by: Daniel P. Berrangé
---
softmmu/cpus.c | 7
This is a critical failure scenario for migration that is hard to
diagnose from existing probes. Most likely it is caused by an error
from bdrv_flush(), but we're not logging the errno anywhere, hence
this new probe.
Reviewed-by: Connor Kuehl
Reviewed-by: Dr. David Alan Gilbert
Signed-off-by: Da
The GDateTime APIs provided by GLib avoid portability pitfalls, such
as some platforms where 'struct timeval.tv_sec' field is still 'long'
instead of 'time_t'. When combined with automatic cleanup, GDateTime
often results in simpler code too.
Reviewed-by: Dr. David Alan Gilbert
Signed-off-by: Dan
The GDateTime APIs provided by GLib avoid portability pitfalls, such
as some platforms where 'struct timeval.tv_sec' field is still 'long'
instead of 'time_t'. When combined with automatic cleanup, GDateTime
often results in simpler code too.
Reviewed-by: Max Reitz
Signed-off-by: Daniel P. Berran
Reviewed-by: Connor Kuehl
Reviewed-by: Dr. David Alan Gilbert
Signed-off-by: Daniel P. Berrangé
---
block/file-posix.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/block/file-posix.c b/block/file-posix.c
index 4189b2bfa6..b3fbb9bd63 100644
--- a/block/file-posix.c
+++ b/block/file-posix
A flush failure is a critical failure scenario for some operations.
For example, it will prevent migration from completing, as it will
make vm_stop() report an error. Thus it is important to have a
trace point present for debugging.
Reviewed-by: Connor Kuehl
Reviewed-by: Dr. David Alan Gilbert
S
When fdatasync() fails on a file backend we set a flag that
short-circuits any future attempts to call fdatasync(). The
first failure returns the true errno, but the later short-
circuited calls return a generic EIO. The latter is unhelpful
because fdatasync() can return a variety of errnos, includ
The authorization framework provides a way to control access to network
services after a client has been authenticated. This documents how to
actually use it.
Reviewed-by: Marc-André Lureau
Signed-off-by: Daniel P. Berrangé
---
docs/system/authz.rst | 263 +++
Reviewed-by: Marc-André Lureau
Signed-off-by: Daniel P. Berrangé
---
docs/system/index.rst | 1 +
docs/system/secrets.rst | 162
2 files changed, 163 insertions(+)
create mode 100644 docs/system/secrets.rst
diff --git a/docs/system/index.rst b/docs/
This is not relevant to any OS distro that QEMU currently targets.
Reviewed-by: Marc-André Lureau
Signed-off-by: Daniel P. Berrangé
---
qemu.sasl | 4
1 file changed, 4 deletions(-)
diff --git a/qemu.sasl b/qemu.sasl
index abdfc686be..851acc7e8f 100644
--- a/qemu.sasl
+++ b/qemu.sasl
@@ -
The SHA-256 variant better meats modern security expectations.
Also warn that the password file is storing entries in clear
text.
Reviewed-by: Marc-André Lureau
Signed-off-by: Daniel P. Berrangé
---
docs/system/vnc-security.rst | 7 ---
qemu.sasl| 11 ++-
2 file
The QAPI reference docs for the guest agent, storage daemon and QMP are
all rather long and hard to navigate unless you already know the name of
the command and can do full text search for it.
A table of contents in each doc will help people locate stuff much more
easily.
Reviewed-by: Connor Kueh
The following changes since commit 894fc4fd670aaf04a67dc7507739f914ff4bacf2:
Merge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into
staging (2021-06-11 09:21:48 +0100)
are available in the Git repository at:
https://gitlab.com/berrange/qemu tags/misc-fixes-pull-request
Hi,
> > I'm not particularly attached to using json for that, it is just that
> > we already have infrastructure to parse/serialize structs from/to json
> > because we need that for qapi anyway.
>
> If we can generate json, we could generate .c code which has all the
> data statically declared
Hi,
> Based on the information you provided, the reason this whole series
> exists seems to be to remove the need to update modules in multiple
> places.
Well, I'm trying to improve the way we handle module meta-data (see
other mail just send for details).
> Is there a real need to copy over th
On Mon, Jun 14, 2021 at 4:56 PM Eric Blake wrote:
>
> On Sat, Jun 12, 2021 at 02:39:44AM +0300, Nir Soffer wrote:
> > Since this change is not simple, and the chance that we also get the dirty
> > bitmap included in the result seems to be very low, I decided to check the
> > direction of merging m
On Sat, Jun 12, 2021 at 02:39:44AM +0300, Nir Soffer wrote:
> Since this change is not simple, and the chance that we also get the dirty
> bitmap included in the result seems to be very low, I decided to check the
> direction of merging multiple extents.
>
> I started with merging "base:allocation
On Mon, Jun 14, 2021 at 03:44:53PM +0200, Gerd Hoffmann wrote:
> Hi,
>
> > > We can do both. The file is generated and installed as part of the
> > > build/install process, and it can be simply used as-is even if some of
> > > the modules are missing.
> > >
> > > It's also possible to update t
Hi,
> > We can do both. The file is generated and installed as part of the
> > build/install process, and it can be simply used as-is even if some of
> > the modules are missing.
> >
> > It's also possible to update the modinfo.json file in postinstall /
> > postuninstall by simply running qem
gcc version 8.3.1 20190311 (Red Hat 8.3.1-3) (GCC)
On 11/06/21 13:13, Paolo Bonzini wrote:
The commit message doesn't specify, but in the buggy case, are we
talking about calling aio_co_wake() for a coroutine in the main context
specifically, right? Could we have a unit test for this scenario?
Yes, that's the scenario. I will try to write a un
On 6/10/21 11:54 AM, Gerd Hoffmann wrote:
> On Thu, Jun 10, 2021 at 10:32:39AM +0200, Claudio Fontana wrote:
>> On 6/10/21 7:57 AM, Gerd Hoffmann wrote:
>>> This patch series adds support for module metadata. Here are the pieces
>>> of the puzzle:
>>>
>>> (1) Macros are added to store metadata i
On 6/10/21 4:03 PM, Gerd Hoffmann wrote:
> Hi,
>
>> Is the JSON file completely static, listing all modules that were built
>> regardless of whether they are currently installed, or would it need to
>> be refreshed when installing/uninstalling RPMs with modules ? I would
>> think we can do the f
On 6/10/21 2:35 PM, Daniel P. Berrangé wrote:
> On Thu, Jun 10, 2021 at 07:57:54AM +0200, Gerd Hoffmann wrote:
>> Add module_allow_arch() to set the target architecture.
>> In case a module is limited to some arch verify arches
>> match and ignore the module if not.
>>
>> Signed-off-by: Gerd Hoffma
On 04/06/2021 11:17, Emanuele Giuseppe Esposito wrote:
Attaching gdbserver implies that the qmp socket
should wait indefinitely for an answer from QEMU.
For Timeout class, create a @contextmanager that
switches Timeout with NoTimeout (empty context manager)
so that if --gdb is set, no timeout
On Jun 13 17:29, Gollu Appalanaidu wrote:
On Wed, Jun 09, 2021 at 10:22:49PM +0200, Klaus Jensen wrote:
On Jun 1 20:32, Gollu Appalanaidu wrote:
Add the controller identifiers list CNS 0x13, available list of ctrls
in NVM Subsystem that may or may not be attached to namespaces.
In Identify Ct
On 04/06/2021 11:17, Emanuele Giuseppe Esposito wrote:
This series adds the option to attach gdbserver and valgrind
to the QEMU binary running in qemu_iotests.
It also allows to redirect QEMU binaries output of the python tests
to the stdout, instead of a log file.
Patches 1-9 introduce the -
First, categorize the structure fields to identify what needs
to be protected and what doesn't.
We essentially need to protect only .state, and the 3 lists in
BDRVBlkdebugState.
Then, add the lock and mark the functions accordingly.
Co-developed-by: Paolo Bonzini
Signed-off-by: Emanuele Giusepp
That would be unsafe in case a rule other than the current one
is removed while the coroutine has yielded.
Keep FOREACH_SAFE because suspend_request deletes the current rule.
After this patch, *all* matching rules are deleted before suspending
the coroutine, rather than just one.
This doesn't affe
Extract to a separate function. Do not rely on FOREACH_SAFE, which is
only "safe" if the *current* node is removed---not if another node is
removed. Instead, just walk the entire list from the beginning when
asked to resume all suspended requests with a given tag.
Co-developed-by: Paolo Bonzini
There seems to be no benefit in using a field. Replace it with a local
variable, and move the state update before the yields.
The state update has do be done before the yields because now using
a local variable does not allow the new updated state to be visible
by the other yields.
Signed-off-by:
Add a counter for each action that a rule can trigger.
This is mainly used to keep track of how many coroutine_yield()
we need to perform after processing all rules in the list.
Co-developed-by: Paolo Bonzini
Signed-off-by: Emanuele Giuseppe Esposito
Reviewed-by: Vladimir Sementsov-Ogievskiy
--
We want to move qemu_coroutine_yield() after the loop on rules,
because QLIST_FOREACH_SAFE is wrong if the rule list is modified
while the coroutine has yielded. Therefore move the suspended
request to the heap and clean it up from the remove side.
All that is left is for blkdebug_debug_event to h
When qemu_coroutine_enter is executed in a loop
(even QEMU_FOREACH_SAFE), the new routine can modify the list,
for example removing an element, causing problem when control
is given back to the caller that continues iterating on the same list.
Patch 1 solves the issue in blkdebug_debug_resume by
Please discard this thread, I had an issue with git send-email and patch
3-5 are missing.
Thank you,
Emanuele
On 14/06/2021 10:08, Emanuele Giuseppe Esposito wrote:
This serie of patches bring thread safety to the smaller APIs used by
block-copy, namely ratelimit, progressmeter, co-shared-reso
This serie of patches bring thread safety to the smaller APIs used by
block-copy, namely ratelimit, progressmeter, co-shared-resource
and aiotask.
The end goal is to reduce the usage of AioContexlock in block-copy,
by introducing smaller granularity locks thus on making the block layer
thread safe.
From: Paolo Bonzini
Reviewed-by: Vladimir Sementsov-Ogievskiy
Signed-off-by: Paolo Bonzini
Signed-off-by: Emanuele Giuseppe Esposito
---
block/block-copy.c | 28 +++-
1 file changed, 11 insertions(+), 17 deletions(-)
diff --git a/block/block-copy.c b/block/block-copy.
From: Paolo Bonzini
Both users of RateLimit, block-copy.c and blockjob.c, treat
a speed of zero as unlimited, while RateLimit treats it as
"as slow as possible". The latter is nicer from the code
point of view but pretty useless, so disable rate limiting
if a speed of zero is provided.
Reviewed
On 14/06/2021 10:11, Emanuele Giuseppe Esposito wrote:
This serie of patches bring thread safety to the smaller APIs used by
block-copy, namely ratelimit, progressmeter, co-shared-resource
and aiotask.
The end goal is to reduce the usage of AioContexlock in block-copy,
by introducing smaller g
co-shared-resource is currently not thread-safe, as also reported
in co-shared-resource.h. Add a QemuMutex because co_try_get_from_shres
can also be invoked from non-coroutine context.
Reviewed-by: Vladimir Sementsov-Ogievskiy
Signed-off-by: Emanuele Giuseppe Esposito
---
include/qemu/co-shared
This serie of patches bring thread safety to the smaller APIs used by
block-copy, namely ratelimit, progressmeter, co-shared-resource
and aiotask.
The end goal is to reduce the usage of AioContexlock in block-copy,
by introducing smaller granularity locks thus on making the block layer
thread safe.
From: Paolo Bonzini
Reviewed-by: Vladimir Sementsov-Ogievskiy
Signed-off-by: Paolo Bonzini
Signed-off-by: Emanuele Giuseppe Esposito
---
blockjob.c | 13 +++--
1 file changed, 3 insertions(+), 10 deletions(-)
diff --git a/blockjob.c b/blockjob.c
index dc1d9e0e46..22e5bb9b1f 100644
--
Progressmeter is protected by the AioContext mutex, which
is taken by the block jobs and their caller (like blockdev).
We would like to remove the dependency of block layer code on the
AioContext mutex, since most drivers and the core I/O code are already
not relying on it.
Create a new C file to
From: Paolo Bonzini
Both users of RateLimit, block-copy.c and blockjob.c, treat
a speed of zero as unlimited, while RateLimit treats it as
"as slow as possible". The latter is nicer from the code
point of view but pretty useless, so disable rate limiting
if a speed of zero is provided.
Reviewed
From: Paolo Bonzini
Reviewed-by: Vladimir Sementsov-Ogievskiy
Signed-off-by: Paolo Bonzini
Signed-off-by: Emanuele Giuseppe Esposito
---
block/block-copy.c | 28 +++-
1 file changed, 11 insertions(+), 17 deletions(-)
diff --git a/block/block-copy.c b/block/block-copy.
Gerd Hoffmann writes:
> Add QAPI schema for the module info database.
>
> Signed-off-by: Gerd Hoffmann
> ---
> qapi/meson.build | 1 +
> qapi/modules.json | 36
> qapi/qapi-schema.json | 1 +
> 3 files changed, 38 insertions(+)
> create mode 1006
Moving this function in task_end ensures to update the progress
anyways, even if there is an error.
It also helps in next patch, allowing task_end to have only
one critical section.
Reviewed-by: Vladimir Sementsov-Ogievskiy
Signed-off-by: Emanuele Giuseppe Esposito
---
block/block-copy.c | 6 +
By adding acquire/release pairs, we ensure that .ret and .error_is_read
fields are written by block_copy_dirty_clusters before .finished is true.
The atomic here are necessary because the fields are concurrently modified
also outside coroutines.
Signed-off-by: Emanuele Giuseppe Esposito
---
blo
Add a CoMutex to protect concurrent access of block-copy
data structures.
This mutex also protects .copy_bitmap, because its thread-safe
API does not prevent it from assigning two tasks to the same
bitmap region.
.finished, .cancelled and reads to .ret and .error_is_read will be
protected in the
From: Paolo Bonzini
Put the logic to determine the copy size in a separate function, so
that there is a simple state machine for the possible methods of
copying data from one BlockDriverState to the other.
Use .method instead of .copy_range as in-out argument, and
include also .zeroes as an addi
As done in BlockCopyCallState, categorize BlockCopyTask
and BlockCopyState in IN, State and OUT fields.
This is just to understand which field has to be protected with a lock.
.sleep_state is handled in the series "coroutine: new sleep/wake API"
and thus here left as TODO.
Signed-off-by: Emanuele
Use a local variable instead of referencing BlockCopyState through a
BlockCopyCallState or BlockCopyTask every time.
This is in preparation for next patches.
No functional change intended.
Signed-off-by: Emanuele Giuseppe Esposito
---
block/block-copy.c | 14 --
1 file changed, 8 in
This serie of patches aims to reduce the usage of the
AioContexlock in block-copy, by introducing smaller granularity
locks thus on making the block layer thread safe.
This serie depends on my previous serie that brings thread safety
to the smaller API used by block-copy, like ratelimit, progress
83 matches
Mail list logo
