From: Auger Eric <eric.auger@redhat.com>
To: Simon Veith <sveith@amazon.de>,
qemu-devel@nongnu.org, qemu-arm@nongnu.org
Subject: Re: [PATCH 3/5] hw/arm/smmuv3: Align stream table base address to table size
Date: Thu, 5 Dec 2019 11:39:55 +0100 [thread overview]
Message-ID: <e6edab1a-25c3-3561-ea61-e2182a4391ad@redhat.com> (raw)
In-Reply-To: <1575467748-28898-4-git-send-email-sveith@amazon.de>
Hi Simon,
On 12/4/19 2:55 PM, Simon Veith wrote:
> Per the specification, and as observed in hardware, the SMMUv3 aligns
> the SMMU_STRTAB_BASE address to the size of the table by masking out the
> respective least significant bits in the ADDR field.
>
> Apply this masking logic to our smmu_find_ste() lookup function per the
> specification.
>
> ref. ARM IHI 0070C, section 6.3.23.
>
> Signed-off-by: Simon Veith <sveith@amazon.de>
> Cc: Eric Auger <eric.auger@redhat.com>
> Cc: qemu-devel@nongnu.org
> Cc: qemu-arm@nongnu.org
> ---
> hw/arm/smmuv3.c | 29 +++++++++++++++++++++++++----
> 1 file changed, 25 insertions(+), 4 deletions(-)
>
> diff --git a/hw/arm/smmuv3.c b/hw/arm/smmuv3.c
> index aad4639..2d6c275 100644
> --- a/hw/arm/smmuv3.c
> +++ b/hw/arm/smmuv3.c
> @@ -376,8 +376,9 @@ bad_ste:
> static int smmu_find_ste(SMMUv3State *s, uint32_t sid, STE *ste,
> SMMUEventInfo *event)
> {
> - dma_addr_t addr;
> + dma_addr_t addr, strtab_base;
> uint32_t log2size;
> + int strtab_size_shift;
> int ret;
>
> trace_smmuv3_find_ste(sid, s->features, s->sid_split);
> @@ -391,10 +392,23 @@ static int smmu_find_ste(SMMUv3State *s, uint32_t sid, STE *ste,
> }
> if (s->features & SMMU_FEATURE_2LVL_STE) {
> int l1_ste_offset, l2_ste_offset, max_l2_ste, span;
> - dma_addr_t strtab_base, l1ptr, l2ptr;
> + dma_addr_t l1ptr, l2ptr;
> STEDesc l1std;
>
> - strtab_base = s->strtab_base & SMMU_BASE_ADDR_MASK;
> + /*
> + * Align strtab base address to table size. For this purpose, assume it
> + * is not bounded by SMMU_IDR1_SIDSIZE.
> + */
> + strtab_size_shift = log2size - s->sid_split - 1 + 3;
Maybe just as the spec use MAX(5, (LOG2SIZE-SPLIT-1+3))?
> + if (strtab_size_shift < DMA_ADDR_BITS) {
> + if (strtab_size_shift < 5) {
> + strtab_size_shift = 5;
> + }
> + strtab_base = s->strtab_base & SMMU_BASE_ADDR_MASK &
> + ~((1ULL << strtab_size_shift) - 1);
nit: you may use ~MAKE_64BIT_MASK(0, strtab_size_shift)
> + } else {
see below
> + strtab_base = 0;
> + }
> l1_ste_offset = sid >> s->sid_split;
> l2_ste_offset = sid & ((1 << s->sid_split) - 1);
> l1ptr = (dma_addr_t)(strtab_base + l1_ste_offset * sizeof(l1std));
> @@ -433,7 +447,14 @@ static int smmu_find_ste(SMMUv3State *s, uint32_t sid, STE *ste,
> }
> addr = l2ptr + l2_ste_offset * sizeof(*ste);
> } else {
> - addr = (s->strtab_base & SMMU_BASE_ADDR_MASK) + sid * sizeof(*ste);
> + strtab_size_shift = log2size + 5;
> + if (strtab_size_shift < DMA_ADDR_BITS) {
> + strtab_base = s->strtab_base & SMMU_BASE_ADDR_MASK &
> + ~((1ULL << strtab_size_shift) - 1);
> + } else {
Can it happen? I understand LOG2SIZE <= SMMU_S_IDR_5.SIDSIZE
and SIDSIZE is max 32. Same above?
> + strtab_base = 0;
> + }
> + addr = strtab_base + sid * sizeof(*ste);
> }
>
> if (smmu_get_ste(s, addr, ste, event)) {
>
Thank you for those series fixes.
Best Regards
Eric
next prev parent reply other threads:[~2019-12-05 10:42 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-12-04 13:55 [PATCH 0/5] hw/arm/smmuv3: Correct stream ID and event address handling Simon Veith
2019-12-04 13:55 ` [PATCH 1/5] hw/arm/smmuv3: Apply address mask to linear strtab base address Simon Veith
2019-12-05 8:42 ` Auger Eric
2019-12-05 22:04 ` Simon Veith
2019-12-09 9:14 ` Auger Eric
2019-12-04 13:55 ` [PATCH 2/5] hw/arm/smmuv3: Check stream IDs against actual table LOG2SIZE Simon Veith
2019-12-05 8:41 ` Auger Eric
2019-12-04 13:55 ` [PATCH 3/5] hw/arm/smmuv3: Align stream table base address to table size Simon Veith
2019-12-05 10:39 ` Auger Eric [this message]
2019-12-04 13:55 ` [PATCH 4/5] hw/arm/smmuv3: Use correct bit positions in EVT_SET_ADDR2 macro Simon Veith
2019-12-05 8:37 ` Auger Eric
2019-12-04 13:55 ` [PATCH 5/5] hw/arm/smmuv3: Report F_STE_FETCH fault address in correct word position Simon Veith
2019-12-05 8:39 ` Auger Eric
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=e6edab1a-25c3-3561-ea61-e2182a4391ad@redhat.com \
--to=eric.auger@redhat.com \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=sveith@amazon.de \
/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).