[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 07/43] hw/block/nvme: fix pci doorbell size calculation
From: |
Kevin Wolf |
Subject: |
[PULL 07/43] hw/block/nvme: fix pci doorbell size calculation |
Date: |
Wed, 17 Jun 2020 16:48:33 +0200 |
From: Klaus Jensen <k.jensen@samsung.com>
The size of the BAR is 0x1000 (main registers) + 8 bytes for each
queue. Currently, the size of the BAR is calculated like so:
n->reg_size = pow2ceil(0x1004 + 2 * (n->num_queues + 1) * 4);
Since the 'num_queues' parameter already accounts for the admin queue,
this should in any case not need to be incremented by one. Also, the
size should be initialized to (0x1000).
n->reg_size = pow2ceil(0x1000 + 2 * n->num_queues * 4);
This, with the default value of num_queues (64), we will set aside room
for 1 admin queue and 63 I/O queues (4 bytes per doorbell, 2 doorbells
per queue).
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Message-Id: <20200609190333.59390-2-its@irrelevant.dk>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
hw/block/nvme.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/hw/block/nvme.c b/hw/block/nvme.c
index a21eeca2fb..c1476e8b2a 100644
--- a/hw/block/nvme.c
+++ b/hw/block/nvme.c
@@ -53,6 +53,9 @@
#include "trace.h"
#include "nvme.h"
+#define NVME_REG_SIZE 0x1000
+#define NVME_DB_SIZE 4
+
#define NVME_GUEST_ERR(trace, fmt, ...) \
do { \
(trace_##trace)(__VA_ARGS__); \
@@ -1401,7 +1404,9 @@ static void nvme_realize(PCIDevice *pci_dev, Error **errp)
pcie_endpoint_cap_init(pci_dev, 0x80);
n->num_namespaces = 1;
- n->reg_size = pow2ceil(0x1004 + 2 * (n->num_queues + 1) * 4);
+
+ /* num_queues is really number of pairs, so each has two doorbells */
+ n->reg_size = pow2ceil(NVME_REG_SIZE + 2 * n->num_queues * NVME_DB_SIZE);
n->ns_size = bs_size / (uint64_t)n->num_namespaces;
n->namespaces = g_new0(NvmeNamespace, n->num_namespaces);
--
2.25.4
- [PULL 00/43] Block layer patches, Kevin Wolf, 2020/06/17
- [PULL 01/43] hw/ide: Make IDEDMAOps handlers take a const IDEDMA pointer, Kevin Wolf, 2020/06/17
- [PULL 03/43] virtio-blk: Refactor the code that processes queued requests, Kevin Wolf, 2020/06/17
- [PULL 04/43] virtio-blk: On restart, process queued requests in the proper context, Kevin Wolf, 2020/06/17
- [PULL 06/43] qcow2: Tweak comments on qcow2_get_persistent_dirty_bitmap_size, Kevin Wolf, 2020/06/17
- [PULL 05/43] block: Refactor subdirectory recursion during make, Kevin Wolf, 2020/06/17
- [PULL 02/43] icount: make dma reads deterministic, Kevin Wolf, 2020/06/17
- [PULL 10/43] hw/block/nvme: move device parameters to separate struct, Kevin Wolf, 2020/06/17
- [PULL 07/43] hw/block/nvme: fix pci doorbell size calculation,
Kevin Wolf <=
- [PULL 09/43] hw/block/nvme: remove superfluous breaks, Kevin Wolf, 2020/06/17
- [PULL 12/43] hw/block/nvme: refactor nvme_addr_read, Kevin Wolf, 2020/06/17
- [PULL 14/43] hw/block/nvme: add max_ioqpairs device parameter, Kevin Wolf, 2020/06/17
- [PULL 15/43] hw/block/nvme: remove redundant cmbloc/cmbsz members, Kevin Wolf, 2020/06/17
- [PULL 16/43] hw/block/nvme: factor out property/constraint checks, Kevin Wolf, 2020/06/17
- [PULL 13/43] hw/block/nvme: fix pin-based interrupt behavior, Kevin Wolf, 2020/06/17
- [PULL 11/43] hw/block/nvme: use constants in identify, Kevin Wolf, 2020/06/17
- [PULL 18/43] hw/block/nvme: factor out block backend setup, Kevin Wolf, 2020/06/17
- [PULL 17/43] hw/block/nvme: factor out device state setup, Kevin Wolf, 2020/06/17
- [PULL 19/43] hw/block/nvme: add namespace helpers, Kevin Wolf, 2020/06/17