linux-stable.git/drivers/dma, branch v5.17.7 Linux kernel stable tree dmaengine: imx-sdma: fix init of uart scripts 2022-04-27T12:41:12+00:00 Kevin Groeneveld kgroeneveld@lenbrook.com 2022-04-10T22:31:18+00:00 7b203ecf14d02e3816e88b21a8a0b521e924ea21 commit a3ae97f4c87d9570e7e9a3e3324c443757f6e29a upstream. Commit b98ce2f4e32b ("dmaengine: imx-sdma: add uart rom script") broke uart rx on imx5 when using sdma firmware from older Freescale 2.6.35 kernel. In this case reading addr->uartXX_2_mcu_addr was going out of bounds of the firmware memory and corrupting the uart script addresses. Simply adding a bounds check before accessing addr->uartXX_2_mcu_addr does not work as the uartXX_2_mcu_addr members are now beyond the size of the older firmware and the uart addresses would never be populated in that case. There are other ways to fix this but overall the logic seems clearer to me to revert the uartXX_2_mcu_ram_addr structure entries back to uartXX_2_mcu_addr, change the newer entries to uartXX_2_mcu_rom_addr and update the logic accordingly. I have tested this patch on: 1. An i.MX53 system with sdma firmware from Freescale 2.6.35 kernel. Without this patch uart rx is broken in this scenario, with the patch uart rx is restored. 2. An i.MX6D system with no external sdma firmware. uart is okay with or without this patch. 3. An i.MX8MM system using current sdma-imx7d.bin firmware from linux-firmware. uart is okay with or without this patch and I confirmed the rom version of the uart script is being used which was the intention and reason for commit b98ce2f4e32b ("dmaengine: imx-sdma: add uart rom script") in the first place. Fixes: b98ce2f4e32b ("dmaengine: imx-sdma: add uart rom script") Cc: stable@vger.kernel.org Signed-off-by: Kevin Groeneveld <kgroeneveld@lenbrook.com> Reviewed-by: Lucas Stach <l.stach@pengutronix.de> Reviewed-by: Fabio Estevam <festevam@gmail.com> Acked-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Link: https://lore.kernel.org/r/20220410223118.15086-1-kgroeneveld@lenbrook.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit a3ae97f4c87d9570e7e9a3e3324c443757f6e29a upstream.

Commit b98ce2f4e32b ("dmaengine: imx-sdma: add uart rom script") broke
uart rx on imx5 when using sdma firmware from older Freescale 2.6.35
kernel. In this case reading addr->uartXX_2_mcu_addr was going out of
bounds of the firmware memory and corrupting the uart script addresses.

Simply adding a bounds check before accessing addr->uartXX_2_mcu_addr
does not work as the uartXX_2_mcu_addr members are now beyond the size
of the older firmware and the uart addresses would never be populated
in that case. There are other ways to fix this but overall the logic
seems clearer to me to revert the uartXX_2_mcu_ram_addr structure
entries back to uartXX_2_mcu_addr, change the newer entries to
uartXX_2_mcu_rom_addr and update the logic accordingly.

I have tested this patch on:
1. An i.MX53 system with sdma firmware from Freescale 2.6.35 kernel.
   Without this patch uart rx is broken in this scenario, with the
   patch uart rx is restored.
2. An i.MX6D system with no external sdma firmware. uart is okay with
   or without this patch.
3. An i.MX8MM system using current sdma-imx7d.bin firmware from
   linux-firmware. uart is okay with or without this patch and I
   confirmed the rom version of the uart script is being used which was
   the intention and reason for commit b98ce2f4e32b ("dmaengine:
   imx-sdma: add uart rom script") in the first place.

Fixes: b98ce2f4e32b ("dmaengine: imx-sdma: add uart rom script")
Cc: stable@vger.kernel.org
Signed-off-by: Kevin Groeneveld <kgroeneveld@lenbrook.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Acked-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/r/20220410223118.15086-1-kgroeneveld@lenbrook.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
dma: at_xdmac: fix a missing check on list iterator 2022-04-27T12:41:12+00:00 Xiaomeng Tong xiam0nd.tong@gmail.com 2022-03-27T06:11:54+00:00 a762a05b557c24b92c09b8e2d04d7925ea3a33d8 commit 206680c4e46b62fd8909385e0874a36952595b85 upstream. The bug is here: __func__, desc, &desc->tx_dma_desc.phys, ret, cookie, residue); The list iterator 'desc' will point to a bogus position containing HEAD if the list is empty or no element is found. To avoid dev_dbg() prints a invalid address, use a new variable 'iter' as the list iterator, while use the origin variable 'desc' as a dedicated pointer to point to the found element. Cc: stable@vger.kernel.org Fixes: 82e2424635f4c ("dmaengine: xdmac: fix print warning on dma_addr_t variable") Signed-off-by: Xiaomeng Tong <xiam0nd.tong@gmail.com> Link: https://lore.kernel.org/r/20220327061154.4867-1-xiam0nd.tong@gmail.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> 
commit 206680c4e46b62fd8909385e0874a36952595b85 upstream.

The bug is here:
	__func__, desc, &desc->tx_dma_desc.phys, ret, cookie, residue);

The list iterator 'desc' will point to a bogus position containing
HEAD if the list is empty or no element is found. To avoid dev_dbg()
prints a invalid address, use a new variable 'iter' as the list
iterator, while use the origin variable 'desc' as a dedicated
pointer to point to the found element.

Cc: stable@vger.kernel.org
Fixes: 82e2424635f4c ("dmaengine: xdmac: fix print warning on dma_addr_t variable")
Signed-off-by: Xiaomeng Tong <xiam0nd.tong@gmail.com>
Link: https://lore.kernel.org/r/20220327061154.4867-1-xiam0nd.tong@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
dmaengine: idxd: skip clearing device context when device is read-only 2022-04-27T12:41:04+00:00 Dave Jiang dave.jiang@intel.com 2022-04-11T22:06:34+00:00 bad2e51e49f1d5442f4863d39f308fb36104f490 [ Upstream commit 1cd8e751d96c43ece3f6842ac2244a37d9332c3a ] If the device shows up as read-only configuration, skip the clearing of the state as the context must be preserved for device re-enable after being disabled. Fixes: 0dcfe41e9a4c ("dmanegine: idxd: cleanup all device related bits after disabling device") Reported-by: Tony Zhu <tony.zhu@intel.com> Tested-by: Tony Zhu <tony.zhu@intel.com> Signed-off-by: Dave Jiang <dave.jiang@intel.com> Link: https://lore.kernel.org/r/164971479479.2200566.13980022473526292759.stgit@djiang5-desk3.ch.intel.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 1cd8e751d96c43ece3f6842ac2244a37d9332c3a ]

If the device shows up as read-only configuration, skip the clearing of the
state as the context must be preserved for device re-enable after being
disabled.

Fixes: 0dcfe41e9a4c ("dmanegine: idxd: cleanup all device related bits after disabling device")
Reported-by: Tony Zhu <tony.zhu@intel.com>
Tested-by: Tony Zhu <tony.zhu@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/164971479479.2200566.13980022473526292759.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
dmaengine: idxd: add RO check for wq max_transfer_size write 2022-04-27T12:41:04+00:00 Dave Jiang dave.jiang@intel.com 2022-04-11T22:08:01+00:00 54fd7e00419628d13b0cc3d41d8484d7029a3afb [ Upstream commit 505a2d1032ae656b0a8c736be110255503941cde ] Block wq_max_transfer_size_store() when the device is configured as read-only and not configurable. Fixes: d7aad5550eca ("dmaengine: idxd: add support for configurable max wq xfer size") Reported-by: Bernice Zhang <bernice.zhang@intel.com> Tested-by: Bernice Zhang <bernice.zhang@intel.com> Signed-off-by: Dave Jiang <dave.jiang@intel.com> Link: https://lore.kernel.org/r/164971488154.2200913.10706665404118545941.stgit@djiang5-desk3.ch.intel.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 505a2d1032ae656b0a8c736be110255503941cde ]

Block wq_max_transfer_size_store() when the device is configured as
read-only and not configurable.

Fixes: d7aad5550eca ("dmaengine: idxd: add support for configurable max wq xfer size")
Reported-by: Bernice Zhang <bernice.zhang@intel.com>
Tested-by: Bernice Zhang <bernice.zhang@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/164971488154.2200913.10706665404118545941.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
dmaengine: idxd: add RO check for wq max_batch_size write 2022-04-27T12:41:04+00:00 Dave Jiang dave.jiang@intel.com 2022-04-11T22:08:55+00:00 2dab40fbb9a73450b01b6e01c1e7e2649fee454d [ Upstream commit 66903461ffed0b66fc3e0200082d4e09365aacdc ] Block wq_max_batch_size_store() when the device is configured as read-only and not configurable. Fixes: e7184b159dd3 ("dmaengine: idxd: add support for configurable max wq batch size") Reported-by: Bernice Zhang <bernice.zhang@intel.com> Tested-by: Bernice Zhang <bernice.zhang@intel.com> Signed-off-by: Dave Jiang <dave.jiang@intel.com> Link: https://lore.kernel.org/r/164971493551.2201159.1942042593642155209.stgit@djiang5-desk3.ch.intel.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 66903461ffed0b66fc3e0200082d4e09365aacdc ]

Block wq_max_batch_size_store() when the device is configured as read-only
and not configurable.

Fixes: e7184b159dd3 ("dmaengine: idxd: add support for configurable max wq batch size")
Reported-by: Bernice Zhang <bernice.zhang@intel.com>
Tested-by: Bernice Zhang <bernice.zhang@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/164971493551.2201159.1942042593642155209.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
dmaengine: idxd: fix retry value to be constant for duration of function call 2022-04-27T12:41:03+00:00 Dave Jiang dave.jiang@intel.com 2022-04-18T21:33:21+00:00 18dd60b7a45da4c6e0d40d084a154fb0f1aedf2e [ Upstream commit bc3452cdfc468a65965d0ac397c940acb787ea4d ] When retries is compared to wq->enqcmds_retries each loop of idxd_enqcmds(), wq->enqcmds_retries can potentially changed by user. Assign the value of retries to wq->enqcmds_retries during initialization so it is the original value set when entering the function. Fixes: 7930d8553575 ("dmaengine: idxd: add knob for enqcmds retries") Suggested-by: Dave Hansen <dave.hansen@intel.com> Signed-off-by: Dave Jiang <dave.jiang@intel.com> Link: https://lore.kernel.org/r/165031760154.3658664.1983547716619266558.stgit@djiang5-desk3.ch.intel.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit bc3452cdfc468a65965d0ac397c940acb787ea4d ]

When retries is compared to wq->enqcmds_retries each loop of idxd_enqcmds(),
wq->enqcmds_retries can potentially changed by user. Assign the value
of retries to wq->enqcmds_retries during initialization so it is the
original value set when entering the function.

Fixes: 7930d8553575 ("dmaengine: idxd: add knob for enqcmds retries")
Suggested-by: Dave Hansen <dave.hansen@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165031760154.3658664.1983547716619266558.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
dmaengine: idxd: match type for retries var in idxd_enqcmds() 2022-04-27T12:41:03+00:00 Dave Jiang dave.jiang@intel.com 2022-04-18T21:31:10+00:00 12b5ce51a12563e6b7ff23d0f1fc5f58185a6190 [ Upstream commit 5d9d16e5aa0cf023e600bf716239fd9caa2d4148 ] wq->enqcmds_retries is defined as unsigned int. However, retries on the stack is defined as int. Change retries to unsigned int to compare the same type. Fixes: 7930d8553575 ("dmaengine: idxd: add knob for enqcmds retries") Suggested-by: Thiago Macieira <thiago.macieira@intel.com> Signed-off-by: Dave Jiang <dave.jiang@intel.com> Link: https://lore.kernel.org/r/165031747059.3658198.6035308204505664375.stgit@djiang5-desk3.ch.intel.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 5d9d16e5aa0cf023e600bf716239fd9caa2d4148 ]

wq->enqcmds_retries is defined as unsigned int. However, retries on the
stack is defined as int. Change retries to unsigned int to compare the same
type.

Fixes: 7930d8553575 ("dmaengine: idxd: add knob for enqcmds retries")
Suggested-by: Thiago Macieira <thiago.macieira@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165031747059.3658198.6035308204505664375.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
dmaengine: dw-edma: Fix unaligned 64bit access 2022-04-27T12:40:57+00:00 Herve Codina herve.codina@bootlin.com 2022-02-25T12:02:52+00:00 6c7f5be40a4e11a07bc4d8e75d110fbc360590f3 [ Upstream commit 8fc5133d6d4da65cad6b73152fc714ad3d7f91c1 ] On some arch (ie aarch64 iMX8MM) unaligned PCIe accesses are not allowed and lead to a kernel Oops. [ 1911.668835] Unable to handle kernel paging request at virtual address ffff80001bc00a8c [ 1911.668841] Mem abort info: [ 1911.668844] ESR = 0x96000061 [ 1911.668847] EC = 0x25: DABT (current EL), IL = 32 bits [ 1911.668850] SET = 0, FnV = 0 [ 1911.668852] EA = 0, S1PTW = 0 [ 1911.668853] Data abort info: [ 1911.668855] ISV = 0, ISS = 0x00000061 [ 1911.668857] CM = 0, WnR = 1 [ 1911.668861] swapper pgtable: 4k pages, 48-bit VAs, pgdp=0000000040ff4000 [ 1911.668864] [ffff80001bc00a8c] pgd=00000000bffff003, pud=00000000bfffe003, pmd=0068000018400705 [ 1911.668872] Internal error: Oops: 96000061 [#1] PREEMPT SMP ... The llp register present in the channel group registers is not aligned on 64bit. Fix unaligned 64bit access using two 32bit accesses Fixes: 04e0a39fc10f ("dmaengine: dw-edma: Add writeq() and readq() for 64 bits architectures") Signed-off-by: Herve Codina <herve.codina@bootlin.com> Link: https://lore.kernel.org/r/20220225120252.309404-1-herve.codina@bootlin.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 8fc5133d6d4da65cad6b73152fc714ad3d7f91c1 ]

On some arch (ie aarch64 iMX8MM) unaligned PCIe accesses are
not allowed and lead to a kernel Oops.
  [ 1911.668835] Unable to handle kernel paging request at virtual address ffff80001bc00a8c
  [ 1911.668841] Mem abort info:
  [ 1911.668844]   ESR = 0x96000061
  [ 1911.668847]   EC = 0x25: DABT (current EL), IL = 32 bits
  [ 1911.668850]   SET = 0, FnV = 0
  [ 1911.668852]   EA = 0, S1PTW = 0
  [ 1911.668853] Data abort info:
  [ 1911.668855]   ISV = 0, ISS = 0x00000061
  [ 1911.668857]   CM = 0, WnR = 1
  [ 1911.668861] swapper pgtable: 4k pages, 48-bit VAs, pgdp=0000000040ff4000
  [ 1911.668864] [ffff80001bc00a8c] pgd=00000000bffff003, pud=00000000bfffe003, pmd=0068000018400705
  [ 1911.668872] Internal error: Oops: 96000061 [#1] PREEMPT SMP
  ...

The llp register present in the channel group registers is not
aligned on 64bit.

Fix unaligned 64bit access using two 32bit accesses

Fixes: 04e0a39fc10f ("dmaengine: dw-edma: Add writeq() and readq() for 64 bits architectures")
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Link: https://lore.kernel.org/r/20220225120252.309404-1-herve.codina@bootlin.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
dmaengine: mediatek:Fix PM usage reference leak of mtk_uart_apdma_alloc_chan_resources 2022-04-27T12:40:57+00:00 zhangqilong zhangqilong3@huawei.com 2022-03-19T02:21:42+00:00 6db489660841b174faf948c94ae660f83e9e3208 [ Upstream commit 545b2baac89b859180e51215468c05d85ea8465a ] pm_runtime_get_sync will increment pm usage counter even it failed. Forgetting to putting operation will result in reference leak here. We fix it: 1) Replacing it with pm_runtime_resume_and_get to keep usage counter balanced. 2) Add putting operation before returning error. Fixes:9135408c3ace4 ("dmaengine: mediatek: Add MediaTek UART APDMA support") Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com> Link: https://lore.kernel.org/r/20220319022142.142709-1-zhangqilong3@huawei.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 545b2baac89b859180e51215468c05d85ea8465a ]

pm_runtime_get_sync will increment pm usage counter even it failed.
Forgetting to putting operation will result in reference leak here.
We fix it:
1) Replacing it with pm_runtime_resume_and_get to keep usage counter
   balanced.
2) Add putting operation before returning error.

Fixes:9135408c3ace4 ("dmaengine: mediatek: Add MediaTek UART APDMA support")
Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com>
Link: https://lore.kernel.org/r/20220319022142.142709-1-zhangqilong3@huawei.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
dmaengine: imx-sdma: Fix error checking in sdma_event_remap 2022-04-27T12:40:56+00:00 Miaoqian Lin linmq006@gmail.com 2022-03-08T06:49:51+00:00 cd9464a4a0c7616bd5d7b4bdda4f487632915b88 [ Upstream commit 7104b9cb35a33ad803a1adbbfa50569b008faf15 ] of_parse_phandle() returns NULL on errors, rather than error pointers. Using NULL check on grp_np to fix this. Fixes: d078cd1b4185 ("dmaengine: imx-sdma: Add imx6sx platform support") Signed-off-by: Miaoqian Lin <linmq006@gmail.com> Link: https://lore.kernel.org/r/20220308064952.15743-1-linmq006@gmail.com Signed-off-by: Vinod Koul <vkoul@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> 
[ Upstream commit 7104b9cb35a33ad803a1adbbfa50569b008faf15 ]

of_parse_phandle() returns NULL on errors, rather than error
pointers. Using NULL check on grp_np to fix this.

Fixes: d078cd1b4185 ("dmaengine: imx-sdma: Add imx6sx platform support")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Link: https://lore.kernel.org/r/20220308064952.15743-1-linmq006@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>