From: Jason-JH Lin <jason-jh.lin@mediatek.com>
To: Rob Herring <robh@kernel.org>,
Krzysztof Kozlowski <krzk+dt@kernel.org>,
Conor Dooley <conor+dt@kernel.org>,
Jassi Brar <jassisinghbrar@gmail.com>,
Chun-Kuang Hu <chunkuang.hu@kernel.org>,
AngeloGioacchino Del Regno
<angelogioacchino.delregno@collabora.com>,
Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Matthias Brugger <matthias.bgg@gmail.com>,
Nicolas Dufresne <nicolas@ndufresne.ca>,
Jason-JH Lin <jason-jh.lin@mediatek.com>,
Nancy Lin <nancy.lin@mediatek.com>,
Singo Chang <singo.chang@mediatek.com>,
Paul-PL Chen <paul-pl.chen@mediatek.com>,
Moudy Ho <moudy.ho@mediatek.com>,
Xiandong Wang <xiandong.wang@mediatek.com>,
Sirius Wang <sirius.wang@mediatek.com>,
Fei Shao <fshao@chromium.org>, Chen-yu Tsai <wenst@chromium.org>,
<Project_Global_Chrome_Upstream_Group@mediatek.com>,
<devicetree@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<dri-devel@lists.freedesktop.org>,
<linux-mediatek@lists.infradead.org>,
<linux-arm-kernel@lists.infradead.org>,
<linux-media@vger.kernel.org>
Subject: [PATCH v7 07/20] mailbox: mtk-cmdq: Add mminfra_offset configuration for DRAM transaction
Date: Wed, 27 Aug 2025 19:37:39 +0800 [thread overview]
Message-ID: <20250827114006.3310175-8-jason-jh.lin@mediatek.com> (raw)
In-Reply-To: <20250827114006.3310175-1-jason-jh.lin@mediatek.com>
The GCE in MT8196 is placed in MMINFRA and requires all addresses
in GCE instructions for DRAM transactions to be IOVA.
Due to MMIO, if the GCE needs to access a hardware register at
0x1000_0000, but the SMMU is also mapping a DRAM block at 0x1000_0000,
the MMINFRA will not know whether to write to the hardware register or
the DRAM.
To solve this, MMINFRA treats addresses greater than 2G as data paths
and those less than 2G as config paths because the DRAM start address
is currently at 2G (0x8000_0000). On the data path, MMINFRA remaps
DRAM addresses by subtracting 2G, allowing SMMU to map DRAM addresses
less than 2G.
For example, if the DRAM start address 0x8000_0000 is mapped to
IOVA=0x0, when GCE accesses IOVA=0x0, it must add a 2G offset to
the address in the GCE instruction. MMINFRA will then see it as a
data path (IOVA >= 2G) and subtract 2G, allowing GCE to access IOVA=0x0.
Since the MMINFRA remap subtracting 2G is done in hardware and cannot
be configured by software, the address of DRAM in GCE instruction must
always add 2G to ensure proper access. After that, the shift functions
do more than just shift addresses, so the APIs were renamed to
cmdq_convert_gce_addr() and cmdq_revert_gce_addr().
This 2G adjustment is referred to as mminfra_offset in the CMDQ driver.
CMDQ helper can get the mminfra_offset from the cmdq_mbox_priv of
cmdq_pkt and add the mminfra_offset to the DRAM address in GCE
instructions.
Signed-off-by: Jason-JH Lin <jason-jh.lin@mediatek.com>
---
drivers/mailbox/mtk-cmdq-mailbox.c | 22 ++++++++++++----------
include/linux/mailbox/mtk-cmdq-mailbox.h | 1 +
2 files changed, 13 insertions(+), 10 deletions(-)
diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
index a9e8895d14df..373effbcfa40 100644
--- a/drivers/mailbox/mtk-cmdq-mailbox.c
+++ b/drivers/mailbox/mtk-cmdq-mailbox.c
@@ -94,20 +94,21 @@ struct cmdq {
struct gce_plat {
u32 thread_nr;
u8 shift;
+ dma_addr_t mminfra_offset;
bool control_by_sw;
bool sw_ddr_en;
bool gce_vm;
u32 gce_num;
};
-static inline u32 cmdq_reg_shift_addr(dma_addr_t addr, const struct gce_plat *pdata)
+static inline u32 cmdq_convert_gce_addr(dma_addr_t addr, const struct gce_plat *pdata)
{
- return (addr >> pdata->shift);
+ return ((addr + pdata->mminfra_offset) >> pdata->shift);
}
-static inline dma_addr_t cmdq_reg_revert_addr(u32 addr, const struct gce_plat *pdata)
+static inline dma_addr_t cmdq_revert_gce_addr(u32 addr, const struct gce_plat *pdata)
{
- return ((dma_addr_t)addr << pdata->shift);
+ return (((dma_addr_t)addr << pdata->shift) - pdata->mminfra_offset);
}
void cmdq_get_mbox_priv(struct mbox_chan *chan, struct cmdq_mbox_priv *priv)
@@ -115,6 +116,7 @@ void cmdq_get_mbox_priv(struct mbox_chan *chan, struct cmdq_mbox_priv *priv)
struct cmdq *cmdq = container_of(chan->mbox, struct cmdq, mbox);
priv->shift_pa = cmdq->pdata->shift;
+ priv->mminfra_offset = cmdq->pdata->mminfra_offset;
}
EXPORT_SYMBOL(cmdq_get_mbox_priv);
@@ -254,7 +256,7 @@ static void cmdq_task_insert_into_thread(struct cmdq_task *task)
struct cmdq_task *prev_task = list_last_entry(
&thread->task_busy_list, typeof(*task), list_entry);
u64 *prev_task_base = prev_task->pkt->va_base;
- u32 shift_addr = cmdq_reg_shift_addr(task->pa_base, task->cmdq->pdata);
+ u32 shift_addr = cmdq_convert_gce_addr(task->pa_base, task->cmdq->pdata);
/* let previous task jump to this task */
dma_sync_single_for_cpu(dev, prev_task->pa_base,
@@ -326,7 +328,7 @@ static void cmdq_thread_irq_handler(struct cmdq *cmdq,
return;
shift_addr = readl(thread->base + CMDQ_THR_CURR_ADDR);
- curr_pa = cmdq_reg_revert_addr(shift_addr, cmdq->pdata);
+ curr_pa = cmdq_revert_gce_addr(shift_addr, cmdq->pdata);
list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
list_entry) {
@@ -477,9 +479,9 @@ static int cmdq_mbox_send_data(struct mbox_chan *chan, void *data)
*/
WARN_ON(cmdq_thread_reset(cmdq, thread) < 0);
- shift_addr = cmdq_reg_shift_addr(task->pa_base, cmdq->pdata);
+ shift_addr = cmdq_convert_gce_addr(task->pa_base, cmdq->pdata);
writel(shift_addr, thread->base + CMDQ_THR_CURR_ADDR);
- shift_addr = cmdq_reg_shift_addr(task->pa_base + pkt->cmd_buf_size, cmdq->pdata);
+ shift_addr = cmdq_convert_gce_addr(task->pa_base + pkt->cmd_buf_size, cmdq->pdata);
writel(shift_addr, thread->base + CMDQ_THR_END_ADDR);
writel(thread->priority, thread->base + CMDQ_THR_PRIORITY);
@@ -488,9 +490,9 @@ static int cmdq_mbox_send_data(struct mbox_chan *chan, void *data)
} else {
WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
shift_addr = readl(thread->base + CMDQ_THR_CURR_ADDR);
- curr_pa = cmdq_reg_revert_addr(shift_addr, cmdq->pdata);
+ curr_pa = cmdq_revert_gce_addr(shift_addr, cmdq->pdata);
shift_addr = readl(thread->base + CMDQ_THR_END_ADDR);
- end_pa = cmdq_reg_revert_addr(shift_addr, cmdq->pdata);
+ end_pa = cmdq_revert_gce_addr(shift_addr, cmdq->pdata);
/* check boundary */
if (curr_pa == end_pa - CMDQ_INST_SIZE ||
curr_pa == end_pa) {
diff --git a/include/linux/mailbox/mtk-cmdq-mailbox.h b/include/linux/mailbox/mtk-cmdq-mailbox.h
index 73b70be4a8a7..07c1bfbdb8c4 100644
--- a/include/linux/mailbox/mtk-cmdq-mailbox.h
+++ b/include/linux/mailbox/mtk-cmdq-mailbox.h
@@ -72,6 +72,7 @@ struct cmdq_cb_data {
struct cmdq_mbox_priv {
u8 shift_pa;
+ dma_addr_t mminfra_offset;
};
struct cmdq_pkt {
--
2.43.0
next prev parent reply other threads:[~2025-08-27 15:49 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-08-27 11:37 [PATCH v7 00/20] Add GCE support for MT8196 Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 01/20] arm64: dts: mediatek: Add GCE header " Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 02/20] mailbox: mtk-cmdq: Refine DMA address handling for the command buffer Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 03/20] mailbox: mtk-cmdq: Add cmdq private data to cmdq_pkt for generating instruction Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 04/20] soc: mediatek: mtk-cmdq: Add cmdq_get_mbox_priv() in cmdq_pkt_create() Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 05/20] soc: mediatek: mtk-cmdq: Add cmdq_pkt_jump_rel_temp() for removing shift_pa Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 06/20] mailbox: mtk-cmdq: Add GCE hardware virtualization configuration Jason-JH Lin
2025-08-27 11:37 ` Jason-JH Lin [this message]
2025-08-27 11:37 ` [PATCH v7 08/20] mailbox: mtk-cmdq: Add driver data to support for MT8196 Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 09/20] soc: mediatek: mtk-cmdq: Add pa_base parsing for hardware without subsys ID support Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 10/20] soc: mediatek: mtk-cmdq: Add new APIs to replace cmdq_pkt_write() and cmdq_pkt_write_mask() Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 11/20] soc: mediatek: mtk-cmdq: Add mminfra_offset adjustment for DRAM addresses Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 12/20] soc: mediatek: Add programming flow for unsupported subsys ID hardware Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 13/20] drm/mediatek: " Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 14/20] media: platform: mtk-mdp3: " Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 15/20] media: platform: mtk-mdp3: Change cmdq_pkt_jump_rel() to cmdq_pkt_jump_rel_temp() Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 16/20] soc: mediatek: mtk-cmdq: Remove shift_pa parameter from cmdq_pkt_jump() Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 17/20] media: platform: mtk-mdp3: Use cmdq_pkt_jump_rel() without shift_pa Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 18/20] soc: mediatek: mtk-cmdq: Remove cmdq_pkt_jump() and cmdq_pkt_jump_rel_temp() Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 19/20] soc: mediatek: mtk-cmdq: Remove cmdq_pkt_write() and cmdq_pkt_write_mask() Jason-JH Lin
2025-08-27 11:37 ` [PATCH v7 20/20] mailbox: mtk-cmdq: Remove unsued cmdq_get_shift_pa() Jason-JH Lin
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=20250827114006.3310175-8-jason-jh.lin@mediatek.com \
--to=jason-jh.lin@mediatek.com \
--cc=Project_Global_Chrome_Upstream_Group@mediatek.com \
--cc=angelogioacchino.delregno@collabora.com \
--cc=chunkuang.hu@kernel.org \
--cc=conor+dt@kernel.org \
--cc=devicetree@vger.kernel.org \
--cc=dri-devel@lists.freedesktop.org \
--cc=fshao@chromium.org \
--cc=jassisinghbrar@gmail.com \
--cc=krzk+dt@kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-media@vger.kernel.org \
--cc=linux-mediatek@lists.infradead.org \
--cc=matthias.bgg@gmail.com \
--cc=mchehab@kernel.org \
--cc=moudy.ho@mediatek.com \
--cc=nancy.lin@mediatek.com \
--cc=nicolas@ndufresne.ca \
--cc=paul-pl.chen@mediatek.com \
--cc=robh@kernel.org \
--cc=singo.chang@mediatek.com \
--cc=sirius.wang@mediatek.com \
--cc=wenst@chromium.org \
--cc=xiandong.wang@mediatek.com \
/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).