ACK: [SRU][X/aws][PATCH 1/1] Revert "iommu/vt-d: Gracefully handle DMAR units with no supported address widths"
Guilherme Piccoli
gpiccoli at canonical.com
Tue Feb 23 17:11:51 UTC 2021
On Tue, Feb 23, 2021 at 1:36 PM Andrea Righi <andrea.righi at canonical.com> wrote:
>
> BugLink: https://bugs.launchpad.net/bugs/1913739
>
> This reverts commit 561ef22a7d396289a1c5a1d18057ccdf9bb59826.
>
> It seems that this patch is addressing the problem correctly only with
> kernels >4.4. For now it is probably safer to simply drop this patch and
> revert this change in 4.4 to avoid introducing potential regressions.
>
> Signed-off-by: Andrea Righi <andrea.righi at canonical.com>
> ---
> drivers/iommu/dmar.c | 45 ++++++++++++++------------------------------
> 1 file changed, 14 insertions(+), 31 deletions(-)
>
> diff --git a/drivers/iommu/dmar.c b/drivers/iommu/dmar.c
> index 78174f2f9fafb..00169c9eb3eed 100644
> --- a/drivers/iommu/dmar.c
> +++ b/drivers/iommu/dmar.c
> @@ -1012,8 +1012,8 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd)
> {
> struct intel_iommu *iommu;
> u32 ver, sts;
> - int agaw = -1;
> - int msagaw = -1;
> + int agaw = 0;
> + int msagaw = 0;
> int err;
>
> if (!drhd->reg_base_addr) {
> @@ -1038,28 +1038,17 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd)
> }
>
> err = -EINVAL;
> - if (cap_sagaw(iommu->cap) == 0) {
> - pr_info("%s: No supported address widths. Not attempting DMA translation.\n",
> - iommu->name);
> - drhd->ignored = 1;
> + agaw = iommu_calculate_agaw(iommu);
> + if (agaw < 0) {
> + pr_err("Cannot get a valid agaw for iommu (seq_id = %d)\n",
> + iommu->seq_id);
> + goto err_unmap;
> }
> -
> - if (!drhd->ignored) {
> - agaw = iommu_calculate_agaw(iommu);
> - if (agaw < 0) {
> - pr_err("Cannot get a valid agaw for iommu (seq_id = %d)\n",
> - iommu->seq_id);
> - drhd->ignored = 1;
> - }
> - }
> - if (!drhd->ignored) {
> - msagaw = iommu_calculate_max_sagaw(iommu);
> - if (msagaw < 0) {
> - pr_err("Cannot get a valid max agaw for iommu (seq_id = %d)\n",
> - iommu->seq_id);
> - drhd->ignored = 1;
> - agaw = -1;
> - }
> + msagaw = iommu_calculate_max_sagaw(iommu);
> + if (msagaw < 0) {
> + pr_err("Cannot get a valid max agaw for iommu (seq_id = %d)\n",
> + iommu->seq_id);
> + goto err_unmap;
> }
> iommu->agaw = agaw;
> iommu->msagaw = msagaw;
> @@ -1088,12 +1077,7 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd)
>
> drhd->iommu = iommu;
>
> - /*
> - * This is only for hotplug; at boot time intel_iommu_enabled won't
> - * be set yet. When intel_iommu_init() runs, it registers the units
> - * present at boot time, then sets intel_iommu_enabled.
> - */
> - if (intel_iommu_enabled && !drhd->ignored)
> + if (intel_iommu_enabled)
> iommu->iommu_dev = iommu_device_create(NULL, iommu,
> intel_iommu_groups,
> "%s", iommu->name);
> @@ -1111,8 +1095,7 @@ error:
>
> static void free_iommu(struct intel_iommu *iommu)
> {
> - if (intel_iommu_enabled && iommu->iommu_dev)
> - iommu_device_destroy(iommu->iommu_dev);
> + iommu_device_destroy(iommu->iommu_dev);
>
> if (iommu->irq) {
> if (iommu->pr_irq) {
> --
> 2.30.0
>
Thanks Andrea, a clean patch, with reasonable justification.
Acked-by: Guilherme G. Piccoli <gpiccoli at canonical.com>
More information about the kernel-team
mailing list