Re: 回复: Re: [PATCH v1 1/2] drm/qxl: replace ioremap by ioremap_cache on arm64

[Christian König <christian.koenig@xxxxxxx>]

    Hi Cong,
    
    well than Dave must decide what to do here.
    
    When QXL emulates a device it should also not use memory accesses
    which won't work on a physical device.
    
    BTW: Your patch is really buggy, it misses the cases in ttm_module.c
    
    Regards,
    Christian.

Am 23.03.22 um 09:48 schrieb liucong2@xxxxxxxxxx:

Hi Christian,

according to 'Arm Architecture Reference Manual Armv8,for Armv8-A

architecture profile' pdf E2.6.5

E2.6.5 Generation of Alignment faults by load/store multiple accesses to

 Device memory

When FEAT_LSMAOC is implemented and the value of the applicable nTLSMD

field is 0, any memory access by an AArch32 Load Multiple or Store 

Multiple instruction to an address that the stage 1 translation 

assigns as Device-nGRE, Device-nGnRE, or Device-nGnRnE generates 

an Alignment fault.

so it seems not just some ARM boards doesn't allow unaligned access to MMIO 

space, all pci memory mapped as Device-nGnRE in arm64 cannot support

unaligned access. and qxl is a device simulated by qemu, it should be able to access 

to MMIO space in a more flexible way(PROT_NORMAL) than the real physical 

graphics card.

----

Cong.

       

主　题：Re: [PATCH v1 1/2] drm/qxl: replace ioremap by ioremap_cache on arm64            
日　期：2022-03-23 15:15            
发件人：Christian König            
收件人：Cong Liuairlied@redhat.comkraxel@redhat.comairlied@linux.iedaniel@ffwll.chray.huang@amd.comvirtualization@lists.linux-foundation.orgspice-devel@lists.freedesktop.orgdri-devel@xxxxxxxxxxxxxxxxxxxxx                    

       

Am 22.03.22 um 10:34 schrieb Cong Liu:
> qxl use ioremap to map ram_header and rom, in the arm64 implementation,
> the device is mapped as DEVICE_nGnRE, it can not support unaligned
> access.

Well that some ARM boards doesn't allow unaligned access to MMIO space
is a well known bug of those ARM boards.

So as far as I know this is a hardware bug you are trying to workaround
here and I'm not 100% sure that this is correct.

Christian.

>
>    6.620515] pc : setup_hw_slot+0x24/0x60 [qxl]
> [    6.620961] lr : setup_slot+0x34/0xf0 [qxl]
> [    6.621376] sp : ffff800012b73760
> [    6.621701] x29: ffff800012b73760 x28: 0000000000000001 x27: 0000000010000000
> [    6.622400] x26: 0000000000000001 x25: 0000000004000000 x24: ffffcf376848c000
> [    6.623099] x23: ffff0000c4087400 x22: ffffcf3718e17140 x21: 0000000000000000
> [    6.623823] x20: ffff0000c4086000 x19: ffff0000c40870b0 x18: 0000000000000014
> [    6.624519] x17: 000000004d3605ab x16: 00000000bb3b6129 x15: 000000006e771809
> [    6.625214] x14: 0000000000000001 x13: 007473696c5f7974 x12: 696e696666615f65
> [    6.625909] x11: 00000000d543656a x10: 0000000000000000 x9 : ffffcf3718e085a4
> [    6.626616] x8 : 00000000006c7871 x7 : 000000000000000a x6 : 0000000000000017
> [    6.627343] x5 : 0000000000001400 x4 : ffff800011f63400 x3 : 0000000014000000
> [    6.628047] x2 : 0000000000000000 x1 : ffff0000c40870b0 x0 : ffff0000c4086000
> [    6.628751] Call trace:
> [    6.628994]  setup_hw_slot+0x24/0x60 [qxl]
> [    6.629404]  setup_slot+0x34/0xf0 [qxl]
> [    6.629790]  qxl_device_init+0x6f0/0x7f0 [qxl]
> [    6.630235]  qxl_pci_probe+0xdc/0x1d0 [qxl]
> [    6.630654]  local_pci_probe+0x48/0xb8
> [    6.631027]  pci_device_probe+0x194/0x208
> [    6.631464]  really_probe+0xd0/0x458
> [    6.631818]  __driver_probe_device+0x124/0x1c0
> [    6.632256]  driver_probe_device+0x48/0x130
> [    6.632669]  __driver_attach+0xc4/0x1a8
> [    6.633049]  bus_for_each_dev+0x78/0xd0
> [    6.633437]  driver_attach+0x2c/0x38
> [    6.633789]  bus_add_driver+0x154/0x248
> [    6.634168]  driver_register+0x6c/0x128
> [    6.635205]  __pci_register_driver+0x4c/0x58
> [    6.635628]  qxl_init+0x48/0x1000 [qxl]
> [    6.636013]  do_one_initcall+0x50/0x240
> [    6.636390]  do_init_module+0x60/0x238
> [    6.636768]  load_module+0x2458/0x2900
> [    6.637136]  __do_sys_finit_module+0xbc/0x128
> [    6.637561]  __arm64_sys_finit_module+0x28/0x38
> [    6.638004]  invoke_syscall+0x74/0xf0
> [    6.638366]  el0_svc_common.constprop.0+0x58/0x1a8
> [    6.638836]  do_el0_svc+0x2c/0x90
> [    6.639216]  el0_svc+0x40/0x190
> [    6.639526]  el0t_64_sync_handler+0xb0/0xb8
> [    6.639934]  el0t_64_sync+0x1a4/0x1a8
> [    6.640294] Code: 910003fd f9484804 f9400c23 8b050084 (f809c083)
> [    6.640889] ---[ end trace 95615d89b7c87f95 ]---
>
> Signed-off-by: Cong Liu
> ---
>   drivers/gpu/drm/qxl/qxl_kms.c | 10 ++++++++++
>   1 file changed, 10 insertions(+)
>
> diff --git a/drivers/gpu/drm/qxl/qxl_kms.c b/drivers/gpu/drm/qxl/qxl_kms.c
> index 4dc5ad13f12c..0e61ac04d8ad 100644
> --- a/drivers/gpu/drm/qxl/qxl_kms.c
> +++ b/drivers/gpu/drm/qxl/qxl_kms.c
> @@ -165,7 +165,11 @@ int qxl_device_init(struct qxl_device *qdev,
>   (int)qdev->surfaceram_size / 1024,
>   (sb == 4) ? "64bit" : "32bit");
>  
> +#ifdef CONFIG_ARM64
> + qdev->rom = ioremap_cache(qdev->rom_base, qdev->rom_size);
> +#else
>   qdev->rom = ioremap(qdev->rom_base, qdev->rom_size);
> +#endif
>   if (!qdev->rom) {
>   pr_err("Unable to ioremap ROM\n");
>   r = -ENOMEM;
> @@ -183,9 +187,15 @@ int qxl_device_init(struct qxl_device *qdev,
>   goto rom_unmap;
>   }
>  
> +#ifdef CONFIG_ARM64
> + qdev->ram_header = ioremap_cache(qdev->vram_base +
> +   qdev->rom->ram_header_offset,
> +   sizeof(*qdev->ram_header));
> +#else
>   qdev->ram_header = ioremap(qdev->vram_base +
>     qdev->rom->ram_header_offset,
>     sizeof(*qdev->ram_header));
> +#endif
>   if (!qdev->ram_header) {
>   DRM_ERROR("Unable to ioremap RAM header\n");
>   r = -ENOMEM;