[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH] hw/arm/smmuv3: Another range invalidation fix
From: |
Auger Eric |
Subject: |
Re: [PATCH] hw/arm/smmuv3: Another range invalidation fix |
Date: |
Mon, 10 May 2021 13:44:13 +0200 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.8.0 |
Hi Peter,
On 5/10/21 1:31 PM, Peter Maydell wrote:
> On Wed, 21 Apr 2021 at 18:29, Eric Auger <eric.auger@redhat.com> wrote:
>>
>> 6d9cd115b9 ("hw/arm/smmuv3: Enforce invalidation on a power of two range")
>> failed to completely fix misalignment issues with range
>> invalidation. For instance invalidations patterns like "invalidate 32
>> 4kB pages starting from 0xff395000 are not correctly handled" due
>> to the fact the previous fix only made sure the number of invalidated
>> pages were a power of 2 but did not properly handle the start
>> address was not aligned with the range. This can be noticed when
>> boothing a fedora 33 with protected virtio-blk-pci.
>>
>> Signed-off-by: Eric Auger <eric.auger@redhat.com>
>> Fixes: 6d9cd115b9 ("hw/arm/smmuv3: Enforce invalidation on a power of two
>> range")
>>
>> ---
>>
>> This bug was found with SMMU RIL avocado-qemu acceptance tests
>> ---
>> hw/arm/smmuv3.c | 49 +++++++++++++++++++++++++------------------------
>> 1 file changed, 25 insertions(+), 24 deletions(-)
>>
>> diff --git a/hw/arm/smmuv3.c b/hw/arm/smmuv3.c
>> index 8705612535..16f285a566 100644
>> --- a/hw/arm/smmuv3.c
>> +++ b/hw/arm/smmuv3.c
>> @@ -856,43 +856,44 @@ static void smmuv3_inv_notifiers_iova(SMMUState *s,
>> int asid, dma_addr_t iova,
>>
>> static void smmuv3_s1_range_inval(SMMUState *s, Cmd *cmd)
>> {
>> - uint8_t scale = 0, num = 0, ttl = 0;
>> - dma_addr_t addr = CMD_ADDR(cmd);
>> + dma_addr_t end, addr = CMD_ADDR(cmd);
>> uint8_t type = CMD_TYPE(cmd);
>> uint16_t vmid = CMD_VMID(cmd);
>> + uint8_t scale = CMD_SCALE(cmd);
>> + uint8_t num = CMD_NUM(cmd);
>> + uint8_t ttl = CMD_TTL(cmd);
>> bool leaf = CMD_LEAF(cmd);
>> uint8_t tg = CMD_TG(cmd);
>> - uint64_t first_page = 0, last_page;
>> - uint64_t num_pages = 1;
>> + uint64_t num_pages;
>> + uint8_t granule;
>> int asid = -1;
>>
>> - if (tg) {
>> - scale = CMD_SCALE(cmd);
>> - num = CMD_NUM(cmd);
>> - ttl = CMD_TTL(cmd);
>> - num_pages = (num + 1) * BIT_ULL(scale);
>> - }
>> -
>> if (type == SMMU_CMD_TLBI_NH_VA) {
>> asid = CMD_ASID(cmd);
>> }
>>
>> + if (!tg) {
>> + trace_smmuv3_s1_range_inval(vmid, asid, addr, tg, 1, ttl, leaf);
>> + smmuv3_inv_notifiers_iova(s, asid, addr, tg, 1);
>> + smmu_iotlb_inv_iova(s, asid, addr, tg, 1, ttl);
>> + }
>
> Is this intended to fall through ?
hum no it isn't. I will fix that.
Thanks
Eric
>
>> +
>> + /* RIL in use */
>> +
>> + num_pages = (num + 1) * BIT_ULL(scale);
>> + granule = tg * 2 + 10;
>> +
>> /* Split invalidations into ^2 range invalidations */
>> - last_page = num_pages - 1;
>> - while (num_pages) {
>> - uint8_t granule = tg * 2 + 10;
>> - uint64_t mask, count;
>> + end = addr + (num_pages << granule) - 1;
>>
>> - mask = dma_aligned_pow2_mask(first_page, last_page, 64 - granule);
>> - count = mask + 1;
>> + while (addr != end + 1) {
>> + uint64_t mask = dma_aligned_pow2_mask(addr, end, 64);
>>
>> - trace_smmuv3_s1_range_inval(vmid, asid, addr, tg, count, ttl, leaf);
>> - smmuv3_inv_notifiers_iova(s, asid, addr, tg, count);
>> - smmu_iotlb_inv_iova(s, asid, addr, tg, count, ttl);
>> -
>> - num_pages -= count;
>> - first_page += count;
>> - addr += count * BIT_ULL(granule);
>> + num_pages = (mask + 1) >> granule;
>> + trace_smmuv3_s1_range_inval(vmid, asid, addr, tg, num_pages, ttl,
>> leaf);
>> + smmuv3_inv_notifiers_iova(s, asid, addr, tg, num_pages);
>> + smmu_iotlb_inv_iova(s, asid, addr, tg, num_pages, ttl);
>> + addr += mask + 1;
>> }
>> }
>
> thanks
> -- PMM
>