hw/sd/sdhci: Strengthen multiple DMA transfers

[PATCH 0/4] hw/sd/sdhci: Strengthen multiple DMA transfers

Posted by Philippe Mathieu-Daudé 3 years, 7 months ago

Still trying to fix the bugs reported by Aleksander...

- Do not send 0 block count
- Reduce DMA to MMIO re-entrancy by yielding when pending IRQ

Based-on: <20200901140411.112150-1-f4bug@amsat.org>

Philippe Mathieu-Daudé (4):
  hw/sd/sdhci: Stop multiple transfers when block count is cleared
  hw/sd/sdhci: Resume pending DMA transfers on MMIO accesses
  hw/sd/sdhci: Let sdhci_update_irq() return if IRQ was delivered
  hw/sd/sdhci: Yield if interrupt delivered during multiple transfer

 hw/sd/sdhci.c | 35 +++++++++++++++++++++++++++++++----
 1 file changed, 31 insertions(+), 4 deletions(-)

-- 
2.26.2

Re: [PATCH 0/4] hw/sd/sdhci: Strengthen multiple DMA transfers

Posted by Philippe Mathieu-Daudé 3 years, 7 months ago

On 9/3/20 7:28 PM, Philippe Mathieu-Daudé wrote:
> Still trying to fix the bugs reported by Aleksander...
> 
> - Do not send 0 block count
> - Reduce DMA to MMIO re-entrancy by yielding when pending IRQ
> 
> Based-on: <20200901140411.112150-1-f4bug@amsat.org>
> 
> Philippe Mathieu-Daudé (4):
>   hw/sd/sdhci: Stop multiple transfers when block count is cleared
>   hw/sd/sdhci: Resume pending DMA transfers on MMIO accesses
>   hw/sd/sdhci: Let sdhci_update_irq() return if IRQ was delivered
>   hw/sd/sdhci: Yield if interrupt delivered during multiple transfer
> 
>  hw/sd/sdhci.c | 35 +++++++++++++++++++++++++++++++----
>  1 file changed, 31 insertions(+), 4 deletions(-)
> 

ping for review?

Re: [PATCH 0/4] hw/sd/sdhci: Strengthen multiple DMA transfers

Posted by Alexander Bulekov 3 years, 7 months ago

I fuzzed the SDHCI with this applied. There are still bugs in SDHCI, but
this fixes the ones triggered by my initial bug-reproducers, and doesn't
appear to create any new bugs.

In the interest of incrementally fixing the issues, for this series:

Tested-by: Alexander Bulekov <alxndr@bu.edu>

On 200903 1928, Philippe Mathieu-Daudé wrote:
> Still trying to fix the bugs reported by Aleksander...
> 
> - Do not send 0 block count
> - Reduce DMA to MMIO re-entrancy by yielding when pending IRQ
> 
> Based-on: <20200901140411.112150-1-f4bug@amsat.org>
> 
> Philippe Mathieu-Daudé (4):
>   hw/sd/sdhci: Stop multiple transfers when block count is cleared
>   hw/sd/sdhci: Resume pending DMA transfers on MMIO accesses
>   hw/sd/sdhci: Let sdhci_update_irq() return if IRQ was delivered
>   hw/sd/sdhci: Yield if interrupt delivered during multiple transfer
> 
>  hw/sd/sdhci.c | 35 +++++++++++++++++++++++++++++++----
>  1 file changed, 31 insertions(+), 4 deletions(-)
> 
> -- 
> 2.26.2
>

Re: [PATCH 0/4] hw/sd/sdhci: Strengthen multiple DMA transfers

Posted by Philippe Mathieu-Daudé 3 years, 7 months ago

On 9/3/20 7:28 PM, Philippe Mathieu-Daudé wrote:
> Still trying to fix the bugs reported by Aleksander...
> 
> - Do not send 0 block count
> - Reduce DMA to MMIO re-entrancy by yielding when pending IRQ
> 
> Based-on: <20200901140411.112150-1-f4bug@amsat.org>
> 
> Philippe Mathieu-Daudé (4):
>   hw/sd/sdhci: Stop multiple transfers when block count is cleared
>   hw/sd/sdhci: Resume pending DMA transfers on MMIO accesses
>   hw/sd/sdhci: Let sdhci_update_irq() return if IRQ was delivered
>   hw/sd/sdhci: Yield if interrupt delivered during multiple transfer

Thanks, series applied to my sd-next tree.