On Sat, Dec 8, 2012 at 10:59 AM, Andreas Färber wrote:
> Generally we encourage people to upstream their devices, given they are
> sufficiently isolated and/or maintainable.
Yes, that makes sense and have been doing that. We maintain an
upstream repo, but needed to merge with a more recent qemu r
If the temperature threshold warning is set below the controller temperature,
enqueue the async event warning.
Signed-off-by: Keith Busch
---
hw/nvme.c |9 +
hw/nvme.h |1 +
2 files changed, 10 insertions(+), 0 deletions(-)
diff --git a/hw/nvme.c b/hw/nvme.c
index c65c179
This allows a driver to set and retrieve the various nvme features defined
in the spec. Where applicable, setting the features changes the device
behavior to reflect the desired feature.
Signed-off-by: Keith Busch
---
hw/nvme.c | 148
For controller capabilities and features visibile to a host driver,
add command line parameters to allow them to be any value the spec
allows. Useful for verifying a driver against a device with various
constraints.
Signed-off-by: Keith Busch
---
hw/nvme.c | 161
This seems like a silly feature, but the spec allows it, so this adds
support for queues that are not physically contiguous.
Signed-off-by: Keith Busch
---
hw/nvme.c | 143 +---
hw/nvme.h |2 +
2 files changed, 118 insertions(+), 27
tches add optional command support, or implement additional features.
Keith Busch (11):
NVMe: Initial commit for NVM Express device
QEMU NVMe: Add command line options
QEMU NVMe: Implement NVMe features
QEMU NVMe: Implement additional admin commands
QEMU NVMe: Add DSM command support
QEMU
Adds logging at various levels of detail throughout the code so that
a developer may see what is occuring within a device interacting with
their nvme drivers and utilities.
Signed-off-by: Keith Busch
---
hw/nvme.c | 266 +++--
1 files
required for an nvme device to work with existing block drivers.
Cc: Keith Busch
Signed-off-by: Keith Busch
---
MAINTAINERS |5 +
blockdev.c|6 +
default-configs/pci.mak |1 +
hw/Makefile.objs |2 +
hw/nvme.c | 851
When errors are detected in commands, set the error log page so that
the host driver may see relevent error data.
Signed-off-by: Keith Busch
---
hw/nvme.c | 25 +
hw/nvme.h |2 ++
2 files changed, 27 insertions(+), 0 deletions(-)
diff --git a/hw/nvme.c b/hw/nvme.c
rom a host driver perspected, except for perhaps controller
PRACT bit set to 1.
Signed-off-by: Keith Busch
---
hw/nvme.c | 72 +---
hw/nvme.h |3 ++
2 files changed, 61 insertions(+), 14 deletions(-)
diff --git a/hw/nvme.c b/hw/nvme.c
This adds support for format, asynch event, abort, and format admin
commands, and helper routines to make it easier to put interesting data
in the log pages.
Signed-off-by: Keith Busch
---
hw/nvme.c | 420 -
hw/nvme.h | 21 +++
2
This adds support for the optional data-set-management command. Action
is only done for the deallocate option.
Signed-off-by: Keith Busch
---
hw/nvme.c | 46 ++
1 files changed, 46 insertions(+), 0 deletions(-)
diff --git a/hw/nvme.c b/hw/nvme.c
Signed-off-by: Keith Busch
---
hw/nvme.c | 72 +++--
hw/nvme.h |2 +
2 files changed, 67 insertions(+), 7 deletions(-)
diff --git a/hw/nvme.c b/hw/nvme.c
index 69136e0..087fce9 100644
--- a/hw/nvme.c
+++ b/hw/nvme.c
@@ -327,7 +327,6
On Wed, 27 Feb 2013, Stefan Hajnoczi wrote:
Hi Keith,
Any pointers on how to test this? For example NVMe-specific userspace
tools that poke the more exotic commands/features?
I use internal tools not externally available, but I know some folks use
the compliance suite here:
http://github.com/
On Wed, 27 Feb 2013, Paolo Bonzini wrote:
Il 27/02/2013 10:30, Andreas Färber ha scritto:
+id->vwc = 1;
+snprintf((char *)id->mn, sizeof(id->mn), "QEMU NVMe Ctrl");
+snprintf((char *)id->fr, sizeof(id->fr), "1.0");
+snprintf((char *)id->sn, sizeof(id->sn), "NVMeQx10%02x", n->
Initial commit for emulated nvme pci storage device. Implements the
minimum from the specification to work with existing drivers.
Cc: Keith Busch
Signed-off-by: Keith Busch
A lot of people are very interested in this, so I should have more time
to dedicate working on submitting this upstream
On Fri, 19 Apr 2013, Stefan Hajnoczi wrote:
I don't see bdrv_aio_flush() in this patch. How does the guest ensure
that data is safely on persistent storage (e.g. protected against data
loss by power failure).
Great point! This is to be consistent with the NVMe spec. This commit
is the bare min
On Thu, 27 Mar 2014, Andreas Färber wrote:
Am 12.03.2014 21:13, schrieb Keith Busch:
On Wed, 12 Mar 2014, Andreas Färber wrote:
Tonight is Hard Freeze for 2.0. If you could also review the matching
test case http://patchwork.ozlabs.org/patch/329627/ I would put both in
my pull.
Looks great
On Wed, 12 Mar 2014, Andreas Färber wrote:
It may not be sensible for normal use cases, but it allows to use
/dev/null in QTest.
Thanks! Sounds useful despite not being usable as storage. :)
I applied to my nvme tree here:
http://git.infradead.org/users/kbusch/qemu-nvme.git
It's quite a bit
8afc13e460a63cac22:
nvme: Permit zero-length block devices (2014-03-12 13:17:58 -0600)
Andreas Färber (1):
nvme: Permit zero-length block devices
Keith Busch (4):
NVMe: Implement remainder from spec
NVMe:
On Wed, 7 Jan 2015, Stefan Hajnoczi wrote:
On Fri, Dec 05, 2014 at 02:40:24PM +0200, a...@e8storage.com wrote:
From: Alex Friedman
According to the specification, the low 16 bits should contain the number of
I/O submission queues, and the high 16 bits should contain the number of
I/O completio
the fix!
Acked-by: Keith Busch
---
hw/block/nvme.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/hw/block/nvme.c b/hw/block/nvme.c
index b6263dc..405cf4f 100644
--- a/hw/block/nvme.c
+++ b/hw/block/nvme.c
@@ -476,7 +476,8 @@ static uint16_t nvme_get_feature(NvmeCtrl
e the configuration data and then handle enable and
NVMe driver works on Windows 8.1.
Hm, Microsoft's driver must be issuing MMIO reads to mask in the enable
bit rather than keep the state known. Sounds odd, but thanks for the fix.
Acked-by: Keith Busch
On Mon, 12 Oct 2015, Kevin Wolf wrote:
Am 12.10.2015 um 10:05 hat Igor Mammedov geschrieben:
On Sat, 10 Oct 2015 00:39:43 -0700
sahlot arvind wrote:
Hello,
Does qemu emulate a proper NVMe controller? Can someone please tell me how
s far as I'm aware, QEMU doesn't emulate any NVMe controlle
Thanks, sounds and looks good.
Acked-by: Keith Busch
> Background on QEMU boot indices
> ---
>
> Normally, the "bootindex" property is configured for bootable devices
> with:
>
> DEVICE_instance_init()
> device_add_
> any OS trying to use the Namespace List will fail the probe.
Thanks for the fix! Minor comment below, no need to fix.
Acked-by: Keith Busch
On Tue, Nov 17, 2015 at 09:33:11AM -0800, Busch, Keith wrote:
> I accidently deleted my comment. Here's what it said:
>
> +list = g_malloc(data_len);
> +for (i = 0; i < n->num_namespaces; i++) {
> +if (i <= min_nsid) {
> +continue;
> +}
> +list[i] = i;
>
o me. I had some difficulty finding the right
patches in the midst of Christoph's Linux patch bombs. :)
Acked-by: Keith Busch
On Mon, Jan 11, 2016 at 05:35:50PM +0100, Markus Armbruster wrote:
> Tom Yan writes:
> > Public bug reported:
> >
> > I am not exactly sure if this is a bug, but I don't see why the option
> > "serial" should be required for -device nvme like the option "drive".
> > Truth is it seem to accept rand
Initial commit for emulated nvme pci device. Implements the minimum from
the specification to work with existing drivers.
Cc: Keith Busch
Signed-off-by: Keith Busch
Just going for the simplest patch possible for this device. If this is
okay, the remainining features implementing the full
On Mon, 27 May 2013, Kevin Wolf wrote:
Am 10.04.2013 um 22:16 hat Keith Busch geschrieben:
Initial commit for emulated nvme pci storage device. Implements the
minimum from the specification to work with existing drivers.
Cc: Keith Busch
Signed-off-by: Keith Busch
A lot of people are very
on the organizations website at:
http://www.nvmexpress.org/
This commit implements the minimum from the specification to work with
existing drivers.
Cc: Keith Busch
Signed-off-by: Keith Busch
---
v4 -> v5:
Call bdrv_flush on controller shutdown.
Mark device as 'unmigratable'.
Hi,
On Tue, 2 Jun 2015, Laszlo Ersek wrote:
removed the nonzero initialization of Cc.Iosqes (submission queue size?)
and Cc.Iocqes (completion queue size?) in function
NvmeEnableController(). And the removal of these field initializations
seems to cause the early sanity check in QEMU's nvme_star
On Fri, 19 Jun 2015, Stefan Hajnoczi wrote:
From: Lu Lina
Signed-off-by: Lu Lina
Thanks for the fix.
Acked-by: Keith Busch
@@ -154,6 +154,7 @@ static uint16_t nvme_dma_read_prp(NvmeCtrl *n, uint8_t
*ptr, uint32_t len,
qemu_sglist_destroy(&qsg);
re
On Fri, Apr 29, 2016 at 10:10:39AM +0100, Stefan Hajnoczi wrote:
> On Thu, Apr 28, 2016 at 05:44:21PM -, Tom Yan wrote:
>
> CCing Keith Busch , maintainer of QEMU NVMe.
> Maybe he has an idea.
Thanks for the report. Sounds like a Windows specific issue as I have no
problem with m
> > C:\Windows\system32>sg_vpd -p sn PD1
> > Unit serial number VPD page:
> > Unit serial number: ___."
I checked your serial number against the SNT refernce on nvmexpress.org and
it's definitely the wrong translation, so that has to be a guest OS driver bug
(Linux has the
migratable
guest-visible state we always return 1 here.
Thanks, this fix was long overdue and already incorporated in my tree. I
really need to get my act together for a legit pull request with all the
other 1.0, 1.1 and 1.2 features.
Acked-by: Keith Busch
Signed-off-by: Christoph Hellwig
diff
UBH member from the above struct with this
patch.
Otherwise looks good.
Reviewed-by: Keith Busch
On Sun, Dec 04, 2022 at 11:06:13AM -0500, Stefan Hajnoczi wrote:
> On Thu, 1 Dec 2022 at 11:50, Klaus Jensen wrote:
> >
> > From: Klaus Jensen
> >
> > Hi,
> >
> > The following changes since commit c4ffd91aba1c3d878e99a3e7ba8aad4826728ece:
> >
> > Update VERSION for v7.2.0-rc3 (2022-11-29 18:15
On Tue, Jul 26, 2022 at 04:55:54PM +0800, Jinhao Fan wrote:
>
> Hi Klaus and Keith,
>
> I just added support for interrupt masking. How can I test interrupt
> masking?
Are you asking about MSI masking? I don't think any drivers are using the
feature, so you'd need to modify one to test it. I can
On Tue, Jul 26, 2022 at 11:32:57PM +0800, Jinhao Fan wrote:
> at 10:45 PM, Keith Busch wrote:
>
> > On Tue, Jul 26, 2022 at 04:55:54PM +0800, Jinhao Fan wrote:
> >> Hi Klaus and Keith,
> >>
> >> I just added support for interrupt masking. How can I test i
On Thu, Jul 28, 2022 at 10:25:19AM +0200, Klaus Jensen wrote:
> From: Klaus Jensen
>
> A set of fixes/changes to the ioeventfd support.
Series looks good.
Reviewed-by: Keith Busch
On Wed, Aug 03, 2022 at 09:46:05AM +0800, Jinhao Fan wrote:
> at 4:54 PM, Klaus Jensen wrote:
>
> > I am unsure if the compiler will transform that division into the shift
> > if it can infer that the divisor is a power of two (it most likely
> > will be able to).
> >
> > But I see no reason to
On Mon, Aug 08, 2022 at 10:23:03AM +0800, Jinhao Fan wrote:
> at 12:35 PM, Jinhao Fan wrote:
>
> > static void nvme_irq_assert(NvmeCtrl *n, NvmeCQueue *cq)
> > {
> > if (cq->irq_enabled) {
> > if (msix_enabled(&(n->parent_obj))) {
> > +/* Initialize CQ irqfd */
> > +
On Wed, Aug 10, 2022 at 12:48:53AM +0800, 樊金昊 wrote:
>> The driver will create the cq with an allocated vector, but it's not
>> activated
>> until after the driver wires it up to a handler. I think that's what you're
>> observing with the incomplete MSIx table entry on creation.
>
> Also, I'm won
create_sq() was getting its
limit from when processing the host command, and sure enough it's
directly from the register field.
Reviewed-by: Keith Busch
The implementation used blocks units rather than the expected bytes.
Fixes: c03e7ef12a9 ("nvme: Implement Write Zeroes")
Reported-by: Ming Lei
Signed-off-by: Keith Busch
---
hw/block/nvme.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/hw/block/nvme.c
d 1.
>
> Signed-off-by: Klaus Jensen
> Signed-off-by: Klaus Jensen
I like this feature a lot, thanks for doing it.
Reviewed-by: Keith Busch
> @@ -1256,18 +1272,24 @@ static uint16_t nvme_smart_info(NvmeCtrl *n, NvmeCmd
> *cmd, uint8_t rae,
> uint64_t units_read = 0, un
'x-use-intel-id' parameter to
> the nvme device to force use of the Intel vendor and device id. This is
> off by default but add a compat property to set this for machines 4.2
> and older.
>
> Signed-off-by: Klaus Jensen
Yay, thank you for following through on getting this identifier assigned.
Reviewed-by: Keith Busch
On Tue, Feb 04, 2020 at 10:52:08AM +0100, Klaus Jensen wrote:
> Signed-off-by: Klaus Jensen
> ---
> hw/block/nvme-ns.c | 2 +-
> hw/block/nvme-ns.h | 4 +++-
> hw/block/nvme.c| 1 +
> 3 files changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/hw/block/nvme-ns.c b/hw/block/nvme-ns.c
>
fault but is added as a machine compat
>property to be true for machine types <= 4.2.
>
> - SGL mapping code has been refactored.
Looking pretty good to me. For the series beyond the individually
reviewed patches:
Acked-by: Keith Busch
If you need to send a v5, you may add my tag
On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
>
> The pin-based interrupt logic in hw/nvme seems sound enough to me, so I
> am wondering if there is something going on with the kernel driver (but
> I certainly do not rule out that hw/nvme is at fault here, since
> pin-based interru
On Thu, Jan 12, 2023 at 06:45:55PM +0100, Klaus Jensen wrote:
> On Jan 12 09:34, Keith Busch wrote:
> > On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
> > >
> > > The pin-based interrupt logic in hw/nvme seems sound enough to me, so I
> > > am w
On Thu, Jan 12, 2023 at 06:45:55PM +0100, Klaus Jensen wrote:
> On Jan 12 09:34, Keith Busch wrote:
> > On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
> > >
> > > The pin-based interrupt logic in hw/nvme seems sound enough to me, so I
> > > am w
On Fri, Jan 13, 2023 at 12:32:29PM +, Peter Maydell wrote:
> On Fri, 13 Jan 2023 at 08:55, Klaus Jensen wrote:
> >
> > +CC qemu pci maintainers
> >
> > Michael, Marcel,
> >
> > Do you have any comments on this thread? As you can see one solution is
> > to simply deassert prior to asserting, th
On Mon, Jan 16, 2023 at 10:14:07PM +0100, Klaus Jensen wrote:
> I noticed that the Linux driver does not use the INTMS/INTMC registers
> to mask interrupts on the controller while processing CQEs. While not
> required by the spec, it is *recommended* in setups not using MSI-X to
> reduce the risk o
On Thu, Jan 12, 2023 at 02:10:51PM +0100, Klaus Jensen wrote:
> Hi all (linux-nvme, qemu-devel, maintainers),
>
> On QEMU riscv64, which does not use MSI/MSI-X and thus relies on
> pin-based interrupts, I'm seeing occasional completion timeouts, i.e.
>
> nvme nvme0: I/O 333 QID 1 timeout, compl
On Wed, Jan 18, 2023 at 03:04:06PM +, Peter Maydell wrote:
> On Tue, 17 Jan 2023 at 19:21, Guenter Roeck wrote:
> > Anyway - any idea what to do to help figuring out what is happening ?
> > Add tracing support to pci interrupt handling, maybe ?
>
> For intermittent bugs, I like recording the
Klaus,
This isn't going to help your issue, but there are at least two legacy
irq bugs in the nvme qemu implementation.
1. The admin queue breaks if start with legacy and later initialize
msix.
2. The legacy vector is shared among all queues, but it's being
deasserted when the first queue's door
On Wed, Jan 18, 2023 at 09:33:05AM -0700, Keith Busch wrote:
> On Wed, Jan 18, 2023 at 03:04:06PM +, Peter Maydell wrote:
> > On Tue, 17 Jan 2023 at 19:21, Guenter Roeck wrote:
> > > Anyway - any idea what to do to help figuring out what is happening ?
> > >
On Thu, Jan 19, 2023 at 10:41:42AM +1000, Alistair Francis wrote:
> On Thu, Jan 19, 2023 at 9:07 AM Keith Busch wrote:
> > ---
> > diff --git a/hw/intc/sifive_plic.c b/hw/intc/sifive_plic.c
> > index c2dfacf028..f8f7af08dc 100644
> > --- a/hw/intc/sifive_plic.c
>
On Thu, Jan 19, 2023 at 01:10:57PM +1000, Alistair Francis wrote:
> On Thu, Jan 19, 2023 at 12:44 PM Keith Busch wrote:
> >
> > Further up, it says the "interrupt gateway" is responsible for
> > forwarding new interrupt requests while the level remains asserte
ond patch changes the default for 'shared' such that namespaces
> are shared by default and will thus by default be attached to hotplugged
> controllers. This adds a compat property for older machine versions and
> updates the documentation to reflect this.
Series looks good.
Reviewed-by: Keith Busch
On Tue, Nov 16, 2021 at 04:34:39PM +0100, Łukasz Gieryk wrote:
> if (!pci_is_vf(&n->parent_obj) && n->params.sriov_max_vfs) {
> -pcie_sriov_pf_disable_vfs(&n->parent_obj);
> +if (rst != NVME_RESET_CONTROLLER) {
> +pcie_sriov_pf_disable_vfs(&n->parent_obj);
Shouldn'
On Thu, Jun 16, 2022 at 08:34:07PM +0800, Jinhao Fan wrote:
> }
> sq->timer = timer_new_ns(QEMU_CLOCK_VIRTUAL, nvme_process_sq, sq);
>
> +if (n->dbbuf_enabled) {
> +sq->db_addr = n->dbbuf_dbs + (sqid << 3);
> +sq->ei_addr = n->dbbuf_eis + (sqid << 3);
> +}
> +
>
On Wed, Jun 29, 2022 at 05:04:25PM +0800, Jinhao Fan wrote:
> Ping~
>
> > @@ -4271,6 +4343,11 @@ static void nvme_init_sq(NvmeSQueue *sq, NvmeCtrl
> > *n, uint64_t dma_addr,
> > if (n->dbbuf_enabled) {
> > sq->db_addr = n->dbbuf_dbs + (sqid << 3);
> > sq->ei_addr = n->dbbuf_ei
On Thu, Jun 30, 2022 at 11:22:31AM +0800, Jinhao Fan wrote:
> +static int nvme_init_sq_ioeventfd(NvmeSQueue *sq)
> +{
> +NvmeCtrl *n = sq->ctrl;
> +uint16_t offset = sq->sqid << 3;
> +int ret;
> +
> +ret = event_notifier_init(&sq->notifier, 0);
> +if (ret < 0) {
> +retur
On Tue, Jul 05, 2022 at 07:11:36PM +0200, Klaus Jensen wrote:
> On Jul 5 22:24, Jinhao Fan wrote:
> > @@ -1374,7 +1374,14 @@ static void nvme_enqueue_req_completion(NvmeCQueue
> > *cq, NvmeRequest *req)
> >
> > QTAILQ_REMOVE(&req->sq->out_req_list, req, entry);
> > QTAILQ_INSERT_TAIL(
inux kernel[1].
>
> [1]:
> https://lore.kernel.org/linux-nvme/20220201190128.3075065-1-kbu...@kernel.org/
Thanks Klaus, this looks good to me.
Reviewed-by: Keith Busch
updated if
that assumption ever changes.
Reviewed-by: Keith Busch
t; +uint64_tslba;
> +uint32_trsvd12;
> +uint8_t zsa;
> +uint8_t zsflags[3];
This should be just a single uint8_t for zsflags, followed by
'uint8_t rsvd[2]'.
Otherwise, looks good.
Reviewed-by: Keith Busch
On Thu, Nov 25, 2021 at 08:37:34AM +0100, Klaus Jensen wrote:
> From: Klaus Jensen
>
> Add enumeration for OZCS values.
Looks good.
Reviewed-by: Keith Busch
On Thu, Nov 25, 2021 at 08:37:33AM +0100, Klaus Jensen wrote:
> @@ -295,7 +295,7 @@ static void nvme_assign_zone_state(NvmeNamespace *ns,
> NvmeZone *zone,
> case NVME_ZONE_STATE_READ_ONLY:
> break;
> default:
> -zone->d.za = 0;
> +NVME_ZA_CLEAR_ALL(zone->d.za);
rwa resources, i.e. number of zones that can have a zrwa),
> "zoned.zrwas" (zrwa size in LBAs), "zoned.zrwafg" (granularity in LBAs
> for flushes).
>
> Signed-off-by: Klaus Jensen
Looks good, and will just need a minor update if you choose to take the
feedback from patch 2 onboard.
Reviewed-by: Keith Busch
On Tue, May 17, 2022 at 01:04:56PM +0200, Klaus Jensen wrote:
> >
> > Should we consider this series ready to merge?
> >
>
> Hi Lukasz and Lukasz :)
>
> Yes, I'm queing this up.
FWIW, this looks good to me. I was hoping to give it a test run, but I don't
think I'll get to that for another week
On Mon, Feb 14, 2022 at 01:30:29PM +0100, Klaus Jensen wrote:
> @@ -384,6 +389,12 @@ static int nvme_ns_check_constraints(NvmeNamespace *ns,
> Error **errp)
> return -1;
> }
>
> +if (ns->params.pif != NVME_PI_GUARD_16 &&
> +ns->params.pif != NVME_PI_GUARD_64) {
> +
Linux kernel[1].
>
> [1]:
> https://lore.kernel.org/linux-nvme/20220201190128.3075065-1-kbu...@kernel.org/
Other than comment on 6/6, series looks good to me.
Reviewed-by: Keith Busch
On Tue, Feb 22, 2022 at 09:48:06AM +, Stefan Hajnoczi wrote:
> On Mon, 21 Feb 2022 at 12:00, Klaus Jensen wrote:
> >
> > Yes, I'll go ahead as mentor for this.
> >
> > @Keith, if you want to join in, let us know :)
Thank you for setting this up, I would be happy assist with the cause!
> > Su
On Fri, Aug 26, 2022 at 05:45:21PM +0200, Klaus Jensen wrote:
> On Aug 26 09:34, Keith Busch wrote:
> > On Fri, Aug 26, 2022 at 11:12:04PM +0800, Jinhao Fan wrote:
> > > Use KVM's irqfd to send interrupts when possible. This approach is
> > > thread safe. More
On Fri, Aug 26, 2022 at 11:12:04PM +0800, Jinhao Fan wrote:
> Use KVM's irqfd to send interrupts when possible. This approach is
> thread safe. Moreover, it does not have the inter-thread communication
> overhead of plain event notifiers since handler callback are called
> in the same system call a
On Thu, Feb 16, 2023 at 05:48:05PM +0100, Jesper Devantier wrote:
> +enum NvmeDirective {
> +NVME_DIRECTIVE_SUPPORTED = 0x0,
> +NVME_DIRECTIVE_ENABLED = 0x1,
> +};
What's this?
On Thu, Feb 16, 2023 at 06:35:27PM +0100, Klaus Jensen wrote:
> On Thu, Feb 16, 2023, at 18:23, Keith Busch wrote:
> > On Thu, Feb 16, 2023 at 05:48:05PM +0100, Jesper Devantier wrote:
> >> +enum NvmeDirective {
> >> +NVME_DIRECTIVE_SUPPORTED = 0x0,
> >>
This mostly looks fine. I need to clarify some spec decisions, though.
By default, FDP feature is disabled: "The default value of this Feature shall
be 0h." You can't change the value as long as namespaces exist within the
group, so FDP requires NS Mgmt be supported if you're going to enable it. T
On Thu, Feb 16, 2023 at 05:48:06PM +0100, Jesper Devantier wrote:
> +static bool nvme_ns_init_fdp(NvmeNamespace *ns, Error **errp)
> +{
> +NvmeEnduranceGroup *endgrp = ns->endgrp;
> +NvmeRuHandle *ruh;
> +uint8_t lbafi = NVME_ID_NS_FLBAS_INDEX(ns->id_ns.flbas);
> +unsigned int *ruhi
nlb);
> +
> +while (nlb) {
> +if (nlb < ru->ruamw) {
> +ru->ruamw -= nlb;
> +break;
> +}
> +
> +nlb -= ru->ruamw;
> +//trace_pci_nvme_fdp_ruh_change(ruh->rgid, ruh->ruhid);
Please use the trace points if you find them useful, otherwise just delete
them instead of committing commented out code.
Beyond that, looks good! For the series:
Reviewed-by: Keith Busch
On Fri, Sep 23, 2022 at 08:34:51AM -0700, Keith Busch wrote:
>
> An iov length needs to be aligned to the logical block size, which may
> be larger than the memory alignment. And since this is only used with
> file-posix backing storage, move the alignment function to there, where
&g
On Thu, Sep 29, 2022 at 07:16:29PM +0300, Maxim Levitsky wrote:
> On Thu, 2022-09-29 at 09:48 -0600, Keith Busch wrote:
> > I am aware, and I've submitted the fix to qemu here:
> >
> > https://lists.nongnu.org/archive/html/qemu-block/2022-09/msg00398.html
> >
&
I am aware, and I've submitted the fix to qemu here:
https://lists.nongnu.org/archive/html/qemu-block/2022-09/msg00398.html
On Thu, Sep 29, 2022 at 07:59:50PM +0200, Kevin Wolf wrote:
> Am 29.09.2022 um 18:09 hat Keith Busch geschrieben:
> > On Fri, Sep 23, 2022 at 08:34:51AM -0700, Keith Busch wrote:
> > >
> > > An iov length needs to be aligned to the logical block size, which may
>
From: Keith Busch
An iov length needs to be aligned to the logical block size, which may
be larger than the memory alignment.
Tested-by: Jens Axboe
Signed-off-by: Keith Busch
---
block/file-posix.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/block/file-posix.c b
From: Keith Busch
There is only user of bdrv_qiov_is_aligned(), so move the alignment
function to there and make it static.
Signed-off-by: Keith Busch
---
block/file-posix.c | 21 +
block/io.c | 21 -
include/block/block-io.h | 1
From: Keith Busch
Changes from v2:
Split the patch so that the function move is separate from the
functional change. This makes it immediately obvious what criteria is
changing. (Kevin Wolf)
Added received Tested-by tag in the changelog.
Keith Busch (2):
block: move
On Sun, Oct 02, 2022 at 11:59:42AM +0300, Maxim Levitsky wrote:
> On Thu, 2022-09-29 at 19:35 +0200, Paolo Bonzini wrote:
> > On 9/29/22 18:39, Christoph Hellwig wrote:
> > > On Thu, Sep 29, 2022 at 10:37:22AM -0600, Keith Busch wrote:
> > > > > I am aware, and I&
bottom halfs and instead of
> calling nvme_post_cqes() immediately (causing an interrupt per cqe), we
> defer the call.
Nice change, looks good! Timers never did seem to be the best fit for
this.
Reviewed-by: Keith Busch
On Thu, Oct 27, 2022 at 01:00:45PM -0500, Jonathan Derrick wrote:
> +Parameters:
> +
> +``auto-ns-path=``
> + If specified indicates a support for dynamic management of nvme namespaces
> + by means of nvme create-ns command. This path points
> + to the storage area for backend images must exist.
On Wed, Jun 08, 2022 at 10:55:30PM +0200, Klaus Jensen wrote:
>
> Keith, is this a bug in the kernel? If the code here would expect the
> doorbell buffer to be updated for the admin queue as well, would we
> break?
The admin queue has to be created before db buffer can be set up, so we can't
rely
On Sun, Jun 12, 2022 at 07:40:55PM +0800, Jinhao Fan wrote:
>
> > On Jun 10, 2022, at 1:27 AM, Klaus Jensen wrote:
> >
> > I'm ok with following the concensus here, but we all agree that this is
> > a blatant spec violation that ended up manifesting itself down the
> > stack, right?
> >
> > So.
On Tue, Jun 14, 2022 at 03:24:37PM +0800, Jinhao Fan wrote:
> > On Jun 14, 2022, at 5:15 AM, Keith Busch wrote:
> > @@ -6538,9 +6544,25 @@ static void nvme_process_db(NvmeCtrl *n, hwaddr
> > addr, int val)
> >
> > trace_pci_nvme_mmio_doorbell_sq(sq->sqid
On Thu, Jun 16, 2022 at 12:42:49PM +0200, Klaus Jensen wrote:
> On Jun 8 03:28, Niklas Cassel via wrote:
> > Hello there,
> >
> > considering that Linux v5.19-rc1 is out which includes support for
> > NVMe TP4084:
> > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drive
; entry to being greater than that value. Therefore, the number of MMIO's
> on the doorbell registers is greatly reduced.
Looks good to me, and passes my sanity tests.
Reviewed-by: Keith Busch
1 - 100 of 303 matches
Mail list logo
