Linux CXL
 help / color / mirror / Atom feed
From: Vinayak Holikatti <vinayak.kh@samsung.com>
To: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: qemu-devel@nongnu.org, gost.dev@samsung.com,
	linux-cxl@vger.kernel.org, nifan.cxl@gmail.com,
	dave@stgolabs.net, vishak.g@samsung.com, krish.reddy@samsung.com,
	a.manzanares@samsung.com, alok.rathore@samsung.com
Subject: Re: [PATCH v2 1/3] hw/cxl/cxl-mailbox-utils: Add support for Media operations discovery commands (8.2.9.9.5.3)
Date: Tue, 18 Feb 2025 11:56:16 +0530	[thread overview]
Message-ID: <20250218062616.6k6p4k6e2p4huwou@test-PowerEdge-R740xd> (raw)
In-Reply-To: <20250214140851.000073fe@huawei.com>

[-- Attachment #1: Type: text/plain, Size: 7342 bytes --]

On 14/02/25 02:08PM, Jonathan Cameron wrote:
>On Thu, 13 Feb 2025 14:45:56 +0530
>Vinayak Holikatti <vinayak.kh@samsung.com> wrote:
>
>> CXL spec 3.1 section 8.2.9.9.5.3 describes media operations commands.
>
>Given the CXL consortium only makes the latest spec available,
>generally we try to reference that.
>It's move to 8.2.10.9.5.3 in r3.2
>
>Otherwise mostly minor style comments inline.
>
>Thanks,
>
>Jonathan
>
Thank You for feed back will update accordingly in V3
>
>
>> CXL devices supports media operations discovery command.
>>
>> Signed-off-by: Vinayak Holikatti <vinayak.kh@samsung.com>
>> ---
>>  hw/cxl/cxl-mailbox-utils.c | 136 +++++++++++++++++++++++++++++++++++++
>>  1 file changed, 136 insertions(+)
>>
>> diff --git a/hw/cxl/cxl-mailbox-utils.c b/hw/cxl/cxl-mailbox-utils.c
>> index 9c7ea5bc35..fa38ecf507 100644
>> --- a/hw/cxl/cxl-mailbox-utils.c
>> +++ b/hw/cxl/cxl-mailbox-utils.c
>> @@ -89,6 +89,7 @@ enum {
>>      SANITIZE    = 0x44,
>>          #define OVERWRITE     0x0
>>          #define SECURE_ERASE  0x1
>> +        #define MEDIA_OPERATIONS 0x2
>>      PERSISTENT_MEM = 0x45,
>>          #define GET_SECURITY_STATE     0x0
>>      MEDIA_AND_POISON = 0x43,
>> @@ -1721,6 +1722,137 @@ static CXLRetCode cmd_sanitize_overwrite(const struct cxl_cmd *cmd,
>>      return CXL_MBOX_BG_STARTED;
>>  }
>>
>> +#define CXL_CACHELINE_SIZE 64
>
>Already defined in include/hw/cxl/cxl.h
ok 
>
>> +enum {
>> +    MEDIA_OP_CLASS_GENERAL  = 0x0,
>> +        #define MEDIA_OP_GEN_SUBC_DISCOVERY 0x0
>> +    MEDIA_OP_CLASS_SANITIZE = 0x1,
>> +        #define MEDIA_OP_SAN_SUBC_SANITIZE 0x0
>> +        #define MEDIA_OP_SAN_SUBC_ZERO 0x1
>> +    MEDIA_OP_CLASS_MAX
>> +};
>> +
>> +struct media_op_supported_list_entry {
>> +    uint8_t media_op_class;
>> +    uint8_t media_op_subclass;
>> +};
>> +
>> +struct media_op_discovery_out_pl {
>> +    uint64_t dpa_range_granularity;
>> +    uint16_t total_supported_operations;
>> +    uint16_t num_of_supported_operations;
>> +    struct media_op_supported_list_entry entry[];
>> +} QEMU_PACKED;
>> +
>> +static const struct media_op_supported_list_entry media_op_matrix[] = {
>> +    {MEDIA_OP_CLASS_GENERAL, MEDIA_OP_GEN_SUBC_DISCOVERY},
>> +    {MEDIA_OP_CLASS_SANITIZE, MEDIA_OP_SAN_SUBC_SANITIZE},
>> +    {MEDIA_OP_CLASS_SANITIZE, MEDIA_OP_SAN_SUBC_ZERO}
>Add trailing comma as we may well get more of these in future.
>In general use a trailing comma whenever there isn't a definite reason
>we will never get them.
ok
>
>Also I'd prefer space after { and before } to match local style.
>    { MEDIA_OP_CLASS_SANITIZE, MEDIA_OP_SAN_SUBC_ZERO },
>
ok
>> +};
>> +
>> +static CXLRetCode media_operations_discovery(uint8_t *payload_in,
>> +                                                size_t len_in,
>> +                                                uint8_t *payload_out,
>> +                                                size_t *len_out)
>Align to opening bracket (just after it)
ok
>> +{
>> +    struct {
>> +        uint8_t media_operation_class;
>> +        uint8_t media_operation_subclass;
>> +        uint8_t rsvd[2];
>> +        uint32_t dpa_range_count;
>> +        struct {
>> +            uint16_t start_index;
>> +            uint16_t num_supported_ops;
>
ok
>I'd just call this num or num_ops
>
ok 
>
>> +        } discovery_osa;
>> +    } QEMU_PACKED *media_op_in_disc_pl = (void *)payload_in;
>> +    int count = 0;
>> +
>> +    if (len_in < sizeof(*media_op_in_disc_pl)) {
>> +        return CXL_MBOX_INVALID_PAYLOAD_LENGTH;
>> +    }
>> +
>> +    struct media_op_discovery_out_pl *media_out_pl =
>> +                                                  (void *)payload_out;
>> +    int num_ops = media_op_in_disc_pl->discovery_osa.num_supported_ops;
>> +    int start_index = media_op_in_disc_pl->discovery_osa.start_index;
>
>Generally we don't mix declarations and code. So move these local variable
>declarations up.
>
ok
>
>> +
>> +    if (start_index + num_ops > ARRAY_SIZE(media_op_matrix)) {
>> +        return CXL_MBOX_INVALID_INPUT;
>> +    }
>> +
>> +    media_out_pl->dpa_range_granularity = CXL_CACHELINE_SIZE;
>> +    media_out_pl->total_supported_operations =
>> +                                     ARRAY_SIZE(media_op_matrix);
>> +    if (num_ops > 0) {
>> +        for (int i = start_index; i < ARRAY_SIZE(media_op_matrix); i++) {
>
>Given you already checked for going out of range, can just do
>i < start_index + num_ops
>I think and avoid the need to break or keep a count.
>
ok
>Keep to local style and declare i outside the loop
>
ok
>
>> +            media_out_pl->entry[count].media_op_class =
>> +                    media_op_matrix[i].media_op_class;
>> +            media_out_pl->entry[count].media_op_subclass =
>> +                        media_op_matrix[i].media_op_subclass;
>> +            count++;
>> +            if (count == num_ops) {
>> +                break;
>> +            }
>> +        }
>> +    }
>> +
>> +    media_out_pl->num_of_supported_operations = count;
>> +    *len_out = sizeof(struct media_op_discovery_out_pl) +
>> +        (sizeof(struct media_op_supported_list_entry) * count);
>> +    return CXL_MBOX_SUCCESS;
>> +}
>> +
>> +static CXLRetCode cmd_media_operations(const struct cxl_cmd *cmd,
>> +                                         uint8_t *payload_in,
>
>Alignment should be to opening bracket.
>
ok
>> +                                         size_t len_in,
>> +                                         uint8_t *payload_out,
>> +                                         size_t *len_out,
>> +                                         CXLCCI *cci)
>> +{
>> +    struct {
>> +        uint8_t media_operation_class;
>> +        uint8_t media_operation_subclass;
>> +        uint8_t rsvd[2];
>> +        uint32_t dpa_range_count;
>> +    } QEMU_PACKED *media_op_in_common_pl = (void *)payload_in;
>> +
>> +    if (len_in < sizeof(*media_op_in_common_pl)) {
>> +        return CXL_MBOX_INVALID_PAYLOAD_LENGTH;
>> +    }
>> +
>> +    uint8_t media_op_cl = media_op_in_common_pl->media_operation_class;
>> +    uint8_t media_op_subclass =
>> +                      media_op_in_common_pl->media_operation_subclass;
>> +    uint32_t dpa_range_count = media_op_in_common_pl->dpa_range_count;
>
>As above, traditional c style with declarations before code.
>
ok
>> +
>> +    switch (media_op_cl) {
>> +    case MEDIA_OP_CLASS_GENERAL:
>> +        if (media_op_subclass != MEDIA_OP_GEN_SUBC_DISCOVERY) {
>> +            return CXL_MBOX_UNSUPPORTED;
>> +        }
>> +
>> +        /*
>> +         * As per spec CXL 3.1 8.2.9.9.5.3 dpa_range_count
>> +         * should be zero for discovery sub class command
>> +         */
>
>I would move this into media_operations_discovery.
>
ok
>> +        if (dpa_range_count) {
>> +            return CXL_MBOX_INVALID_INPUT;
>> +        }
>> +
>> +        return media_operations_discovery(payload_in, len_in, payload_out,
>> +                                             len_out);
>> +    case MEDIA_OP_CLASS_SANITIZE:
>
>Easier to introduce this case in next patch.  Until then can just let
>the default deal with it.
>
ok
>> +        switch (media_op_subclass) {
>> +        default:
>> +            return CXL_MBOX_UNSUPPORTED;
>> +        }
>> +    default:
>> +        return CXL_MBOX_UNSUPPORTED;
>> +    }
>> +
>> +    return CXL_MBOX_SUCCESS;
>> +}
>> +
>
>

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]



  reply	other threads:[~2025-02-18  6:26 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CGME20250213091628epcas5p31ec87df7fc4ce2d47db40b693239d7ad@epcas5p3.samsung.com>
2025-02-13  9:15 ` [PATCH v2 0/3] CXL CCI Media Operations Vinayak Holikatti
2025-02-13  9:15   ` [PATCH v2 1/3] hw/cxl/cxl-mailbox-utils: Add support for Media operations discovery commands (8.2.9.9.5.3) Vinayak Holikatti
2025-02-14 14:08     ` Jonathan Cameron
2025-02-18  6:26       ` Vinayak Holikatti [this message]
2025-02-13  9:15   ` [PATCH v2 2/3] hw/cxl: factor out calculation of sanitize duration from cmd_santize_overwrite Vinayak Holikatti
2025-02-13  9:15   ` [PATCH v2 3/3] hw/cxl/cxl-mailbox-utils: Add support for Media operations Sanitize and Write Zeros commands (8.2.9.9.5.3) Vinayak Holikatti
2025-02-14 14:40     ` Jonathan Cameron
2025-02-18  6:34       ` Vinayak Holikatti

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20250218062616.6k6p4k6e2p4huwou@test-PowerEdge-R740xd \
    --to=vinayak.kh@samsung.com \
    --cc=Jonathan.Cameron@huawei.com \
    --cc=a.manzanares@samsung.com \
    --cc=alok.rathore@samsung.com \
    --cc=dave@stgolabs.net \
    --cc=gost.dev@samsung.com \
    --cc=krish.reddy@samsung.com \
    --cc=linux-cxl@vger.kernel.org \
    --cc=nifan.cxl@gmail.com \
    --cc=qemu-devel@nongnu.org \
    --cc=vishak.g@samsung.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox