qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Philippe Mathieu-Daudé" <philmd@redhat.com>
To: David Gibson <david@gibson.dropbear.id.au>
Cc: Fam Zheng <fam@euphon.net>,
	qemu-ppc@nongnu.org, qemu-devel@nongnu.org,
	Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [PATCH 5/5] hw/scsi/spapr_vscsi: Do not mix SRP IU size with DMA buffer size
Date: Thu, 5 Mar 2020 07:47:53 +0100	[thread overview]
Message-ID: <020a0760-bca5-346e-810c-28d438206458@redhat.com> (raw)
In-Reply-To: <20200305004337.GL593957@umbus.fritz.box>

On 3/5/20 1:43 AM, David Gibson wrote:
> On Wed, Mar 04, 2020 at 04:33:11PM +0100, Philippe Mathieu-Daudé wrote:
>> The 'union srp_iu' is meant as a pointer to any SRP Information
>> Unit type, it is not related to the size of a VIO DMA buffer.
>>
>> Use a plain buffer for the VIO DMA read/write calls.
>> We can remove the reserved buffer from the 'union srp_iu'.
>>
>> This issue was noticed when replacing the zero-length arrays
>> from hw/scsi/srp.h with flexible array member,
>> 'clang -fsanitize=undefined' reported:
>>
>>    hw/scsi/spapr_vscsi.c:69:29: error: field 'iu' with variable sized type 'union viosrp_iu' not at the end of a struct or class is a GNU extension [-Werror,-Wgnu-variable-sized-type-not-at-end]
>>         union viosrp_iu         iu;
>>                                 ^
>>
>> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>> ---
>>   hw/scsi/viosrp.h      |  2 +-
>>   hw/scsi/spapr_vscsi.c | 10 +++++-----
>>   2 files changed, 6 insertions(+), 6 deletions(-)
>>
>> diff --git a/hw/scsi/viosrp.h b/hw/scsi/viosrp.h
>> index 25676c2383..aba3203028 100644
>> --- a/hw/scsi/viosrp.h
>> +++ b/hw/scsi/viosrp.h
>> @@ -49,8 +49,8 @@ union srp_iu {
>>       struct srp_tsk_mgmt tsk_mgmt;
>>       struct srp_cmd cmd;
>>       struct srp_rsp rsp;
>> -    uint8_t reserved[SRP_MAX_IU_LEN];
>>   };
>> +_Static_assert(sizeof(union srp_iu) <= SRP_MAX_IU_LEN, "srp_iu size incorrect");
> 
> Hrm.  Given that srp_iu will be a variably sized type, is this
> assertion actually testing anything meaningful?

I wanted to assert that if another SRP IU is added, the DMA buffer will 
be big enough to hold it. I'll simply remove the check.

> Otherwise, LGTM.

Thanks for reviewing the series,

Phil.

> 
>>   
>>   enum viosrp_crq_formats {
>>       VIOSRP_SRP_FORMAT = 0x01,
>> diff --git a/hw/scsi/spapr_vscsi.c b/hw/scsi/spapr_vscsi.c
>> index f1a0bbdc31..f9be68e44e 100644
>> --- a/hw/scsi/spapr_vscsi.c
>> +++ b/hw/scsi/spapr_vscsi.c
>> @@ -66,7 +66,7 @@ typedef union vscsi_crq {
>>   
>>   typedef struct vscsi_req {
>>       vscsi_crq               crq;
>> -    union viosrp_iu         iu;
>> +    uint8_t                 viosrp_iu_buf[SRP_MAX_IU_LEN];
>>   
>>       /* SCSI request tracking */
>>       SCSIRequest             *sreq;
>> @@ -99,7 +99,7 @@ typedef struct {
>>   
>>   static union viosrp_iu *req_iu(vscsi_req *req)
>>   {
>> -    return (union viosrp_iu *)req->iu.srp.reserved;
>> +    return (union viosrp_iu *)req->viosrp_iu_buf;
>>   }
>>   
>>   
>> @@ -184,7 +184,7 @@ static int vscsi_send_iu(VSCSIState *s, vscsi_req *req,
>>   
>>       /* First copy the SRP */
>>       rc = spapr_vio_dma_write(&s->vdev, req->crq.s.IU_data_ptr,
>> -                             &req->iu, length);
>> +                             &req->viosrp_iu_buf, length);
>>       if (rc) {
>>           fprintf(stderr, "vscsi_send_iu: DMA write failure !\n");
>>       }
>> @@ -603,7 +603,7 @@ static const VMStateDescription vmstate_spapr_vscsi_req = {
>>       .minimum_version_id = 1,
>>       .fields = (VMStateField[]) {
>>           VMSTATE_BUFFER(crq.raw, vscsi_req),
>> -        VMSTATE_BUFFER(iu.srp.reserved, vscsi_req),
>> +        VMSTATE_BUFFER(viosrp_iu_buf, vscsi_req),
>>           VMSTATE_UINT32(qtag, vscsi_req),
>>           VMSTATE_BOOL(active, vscsi_req),
>>           VMSTATE_UINT32(data_len, vscsi_req),
>> @@ -1104,7 +1104,7 @@ static void vscsi_got_payload(VSCSIState *s, vscsi_crq *crq)
>>       }
>>   
>>       /* XXX Handle failure differently ? */
>> -    if (spapr_vio_dma_read(&s->vdev, crq->s.IU_data_ptr, &req->iu,
>> +    if (spapr_vio_dma_read(&s->vdev, crq->s.IU_data_ptr, &req->viosrp_iu_buf,
>>                              crq->s.IU_length)) {
>>           fprintf(stderr, "vscsi_got_payload: DMA read failure !\n");
>>           vscsi_put_req(req);
> 



      reply	other threads:[~2020-03-05  6:48 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-04 15:33 [PATCH 0/5] hw/scsi/spapr_vscsi: Fix time bomb zero-length array use Philippe Mathieu-Daudé
2020-03-04 15:33 ` [PATCH 1/5] hw/scsi/viosrp: Add missing 'hw/scsi/srp.h' include Philippe Mathieu-Daudé
2020-03-05  0:39   ` David Gibson
2020-03-04 15:33 ` [PATCH 2/5] hw/scsi/spapr_vscsi: Use SRP_MAX_IU_LEN instead of sizeof flexible array Philippe Mathieu-Daudé
2020-03-05  0:40   ` David Gibson
2020-03-04 15:33 ` [PATCH 3/5] hw/scsi/spapr_vscsi: Simplify a bit Philippe Mathieu-Daudé
2020-03-05  0:40   ` David Gibson
2020-03-05 10:21   ` Greg Kurz
2020-03-04 15:33 ` [PATCH 4/5] hw/scsi/spapr_vscsi: Introduce req_ui() helper Philippe Mathieu-Daudé
2020-03-05  0:41   ` David Gibson
2020-03-05  0:42     ` David Gibson
2020-03-04 15:33 ` [PATCH 5/5] hw/scsi/spapr_vscsi: Do not mix SRP IU size with DMA buffer size Philippe Mathieu-Daudé
2020-03-05  0:43   ` David Gibson
2020-03-05  6:47     ` Philippe Mathieu-Daudé [this message]

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=020a0760-bca5-346e-810c-28d438206458@redhat.com \
    --to=philmd@redhat.com \
    --cc=david@gibson.dropbear.id.au \
    --cc=fam@euphon.net \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-ppc@nongnu.org \
    /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;
as well as URLs for NNTP newsgroup(s).