[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH for-6.0 v2 4/8] hw/block/nvme: fix controller namespaces arra
From: |
Philippe Mathieu-Daudé |
Subject: |
Re: [PATCH for-6.0 v2 4/8] hw/block/nvme: fix controller namespaces array indexing |
Date: |
Tue, 6 Apr 2021 09:01:44 +0200 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.8.1 |
On 4/5/21 7:54 PM, Klaus Jensen wrote:
> From: Klaus Jensen <k.jensen@samsung.com>
>
> The controller namespaces array being 0-indexed requires 'nsid - 1'
> everywhere. Something that is easy to miss. Align the controller
> namespaces array with the subsystem namespaces array such that both are
> 1-indexed.
TBH I don't understand the justification. Assuming you hit a
bug and try to protect yourself, maybe now you should also
check for
assert(n->namespaces[0] == NULL);
somewhere. In nvme_ns() maybe?
>
> Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
> Reviewed-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
> ---
> hw/block/nvme.h | 8 ++++----
> hw/block/nvme.c | 2 +-
> 2 files changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/hw/block/nvme.h b/hw/block/nvme.h
> index 9edc86d79e98..c610ab30dc5c 100644
> --- a/hw/block/nvme.h
> +++ b/hw/block/nvme.h
> @@ -217,7 +217,7 @@ typedef struct NvmeCtrl {
> * Attached namespaces to this controller. If subsys is not given, all
> * namespaces in this list will always be attached.
> */
> - NvmeNamespace *namespaces[NVME_MAX_NAMESPACES];
> + NvmeNamespace *namespaces[NVME_MAX_NAMESPACES + 1];
> NvmeSQueue **sq;
> NvmeCQueue **cq;
> NvmeSQueue admin_sq;
> @@ -232,7 +232,7 @@ static inline NvmeNamespace *nvme_ns(NvmeCtrl *n,
> uint32_t nsid)
> return NULL;
> }
>
> - return n->namespaces[nsid - 1];
> + return n->namespaces[nsid];
> }
>
> static inline bool nvme_ns_is_attached(NvmeCtrl *n, NvmeNamespace *ns)
> @@ -253,7 +253,7 @@ static inline void nvme_ns_attach(NvmeCtrl *n,
> NvmeNamespace *ns)
> uint32_t nsid = nvme_nsid(ns);
> assert(nsid && nsid <= NVME_MAX_NAMESPACES);
>
> - n->namespaces[nsid - 1] = ns;
> + n->namespaces[nsid] = ns;
> }
>
> static inline void nvme_ns_detach(NvmeCtrl *n, NvmeNamespace *ns)
> @@ -261,7 +261,7 @@ static inline void nvme_ns_detach(NvmeCtrl *n,
> NvmeNamespace *ns)
> uint32_t nsid = nvme_nsid(ns);
> assert(nsid && nsid <= NVME_MAX_NAMESPACES);
>
> - n->namespaces[nsid - 1] = NULL;
> + n->namespaces[nsid] = NULL;
> }
>
> static inline NvmeCQueue *nvme_cq(NvmeRequest *req)
> diff --git a/hw/block/nvme.c b/hw/block/nvme.c
> index c54ec3c9523c..6d31d5b17a0b 100644
> --- a/hw/block/nvme.c
> +++ b/hw/block/nvme.c
> @@ -5915,7 +5915,7 @@ int nvme_register_namespace(NvmeCtrl *n, NvmeNamespace
> *ns, Error **errp)
> return -1;
> }
> } else {
> - if (n->namespaces[nsid - 1]) {
> + if (n->namespaces[nsid]) {
> error_setg(errp, "namespace id '%d' is already in use", nsid);
> return -1;
> }
>
[PATCH for-6.0 v2 8/8] hw/block/nvme: add missing copyright headers, Klaus Jensen, 2021/04/05
[PATCH for-6.0 v2 7/8] hw/block/nvme: fix handling of private namespaces, Klaus Jensen, 2021/04/05