[PATCH 1/2] dm: core: Add late driver remove option
Add another flag to the DM core which could be assigned to drivers and which makes those drivers call their remove callbacks last, just before booting OS and after all the other drivers finished with their remove callbacks. This is necessary for things like clock drivers, where the other drivers might depend on the clock driver in their remove callbacks. Prime example is the mmc subsystem, which can reconfigure a card from HS mode to slower modes in the remove callback and for that it needs to reconfigure the controller clock.
Signed-off-by: Marek Vasut marek.vasut+renesas@gmail.com Cc: Simon Glass sjg@chromium.org Cc: Tom Rini trini@konsulko.com --- arch/arm/lib/bootm.c | 1 + board/Marvell/octeontx2/board.c | 4 ++-- drivers/core/device-remove.c | 11 ++++++++--- drivers/core/root.c | 2 ++ drivers/core/uclass.c | 32 +++++++++++++++++++++++++------- include/dm/device.h | 4 ++++ include/dm/uclass-internal.h | 3 ++- test/dm/core.c | 21 ++++++++++++--------- test/dm/test-main.c | 30 +++++++++++++++++------------- 9 files changed, 73 insertions(+), 35 deletions(-)
diff --git a/arch/arm/lib/bootm.c b/arch/arm/lib/bootm.c index 1206e306db..f9091a3d41 100644 --- a/arch/arm/lib/bootm.c +++ b/arch/arm/lib/bootm.c @@ -120,6 +120,7 @@ static void announce_and_cleanup(int fake) * of DMA operation or releasing device internal buffers. */ dm_remove_devices_flags(DM_REMOVE_ACTIVE_ALL); + dm_remove_devices_flags(DM_REMOVE_ACTIVE_ALL | DM_REMOVE_LATE);
cleanup_before_linux(); } diff --git a/board/Marvell/octeontx2/board.c b/board/Marvell/octeontx2/board.c index 50e903d9aa..d1e2372a37 100644 --- a/board/Marvell/octeontx2/board.c +++ b/board/Marvell/octeontx2/board.c @@ -65,7 +65,7 @@ void board_quiesce_devices(void) /* Removes all RVU PF devices */ ret = uclass_get(UCLASS_ETH, &uc_dev); if (uc_dev) - ret = uclass_destroy(uc_dev); + ret = uclass_destroy(uc_dev, DM_REMOVE_NORMAL); if (ret) printf("couldn't remove rvu pf devices\n");
@@ -77,7 +77,7 @@ void board_quiesce_devices(void) /* Removes all CGX and RVU AF devices */ ret = uclass_get(UCLASS_MISC, &uc_dev); if (uc_dev) - ret = uclass_destroy(uc_dev); + ret = uclass_destroy(uc_dev, DM_REMOVE_NORMAL); if (ret) printf("couldn't remove misc (cgx/rvu_af) devices\n");
diff --git a/drivers/core/device-remove.c b/drivers/core/device-remove.c index efdb0f2905..07b241b6bb 100644 --- a/drivers/core/device-remove.c +++ b/drivers/core/device-remove.c @@ -172,14 +172,19 @@ int device_remove(struct udevice *dev, uint flags) drv = dev->driver; assert(drv);
- ret = uclass_pre_remove_device(dev); - if (ret) - return ret; + if (!(flags & DM_REMOVE_LATE)) { + ret = uclass_pre_remove_device(dev); + if (ret) + return ret; + }
ret = device_chld_remove(dev, NULL, flags); if (ret) goto err;
+ if (!(flags & DM_REMOVE_LATE) && (drv->flags & DM_FLAG_REMOVE_LATE)) + return 0; + /* * Remove the device if called with the "normal" remove flag set, * or if the remove flag matches any of the drivers remove flags diff --git a/drivers/core/root.c b/drivers/core/root.c index 0726be6b79..21f3054125 100644 --- a/drivers/core/root.c +++ b/drivers/core/root.c @@ -168,6 +168,7 @@ int dm_init(bool of_live) int dm_uninit(void) { device_remove(dm_root(), DM_REMOVE_NORMAL); + device_remove(dm_root(), DM_REMOVE_LATE); device_unbind(dm_root()); gd->dm_root = NULL;
@@ -393,6 +394,7 @@ struct acpi_ops root_acpi_ops = { U_BOOT_DRIVER(root_driver) = { .name = "root_driver", .id = UCLASS_ROOT, + .flags = DM_FLAG_REMOVE_LATE, ACPI_OPS_PTR(&root_acpi_ops) };
diff --git a/drivers/core/uclass.c b/drivers/core/uclass.c index c3f1b73cd6..ac474d3ff8 100644 --- a/drivers/core/uclass.c +++ b/drivers/core/uclass.c @@ -104,10 +104,28 @@ fail_mem: return ret; }
-int uclass_destroy(struct uclass *uc) +int uclass_find_device_by_drv_flag(struct uclass *uc, const unsigned int flag, + const bool neg, struct udevice **devp) +{ + struct udevice *dev; + + *devp = NULL; + uclass_foreach_dev(dev, uc) { + if ((neg && (dev->driver->flags & flag)) || + (!neg && !(dev->driver->flags & flag))) { + *devp = dev; + return 0; + } + } + + return -ENODEV; +} + +int uclass_destroy(struct uclass *uc, unsigned int flag) { struct uclass_driver *uc_drv; struct udevice *dev; + bool late = flag & DM_REMOVE_LATE; int ret;
/* @@ -116,15 +134,15 @@ int uclass_destroy(struct uclass *uc) * unbound (by the recursion in the call to device_unbind() below). * We can loop until the list is empty. */ - while (!list_empty(&uc->dev_head)) { - dev = list_first_entry(&uc->dev_head, struct udevice, - uclass_node); - ret = device_remove(dev, DM_REMOVE_NORMAL | DM_REMOVE_NO_PD); - if (ret) + while (!uclass_find_device_by_drv_flag(uc, DM_FLAG_REMOVE_LATE, late, &dev)) { + ret = device_remove(dev, flag | DM_REMOVE_NO_PD); + if (ret) { return log_msg_ret("remove", ret); + } ret = device_unbind(dev); - if (ret) + if (ret) { return log_msg_ret("unbind", ret); + } }
uc_drv = uc->uc_drv; diff --git a/include/dm/device.h b/include/dm/device.h index ac3b6c1b8a..b3714ea851 100644 --- a/include/dm/device.h +++ b/include/dm/device.h @@ -73,6 +73,8 @@ struct driver_info; */ #define DM_FLAG_REMOVE_WITH_PD_ON (1 << 13)
+#define DM_FLAG_REMOVE_LATE (1 << 14) + /* * One or multiple of these flags are passed to device_remove() so that * a selective device removal as specified by the remove-stage and the @@ -95,6 +97,8 @@ enum {
/* Don't power down any attached power domains */ DM_REMOVE_NO_PD = 1 << 1, + + DM_REMOVE_LATE = 1 << 2, };
/** diff --git a/include/dm/uclass-internal.h b/include/dm/uclass-internal.h index 6e3f15c2b0..b5926b0f5c 100644 --- a/include/dm/uclass-internal.h +++ b/include/dm/uclass-internal.h @@ -247,8 +247,9 @@ struct uclass *uclass_find(enum uclass_id key); * Destroy a uclass and all its devices * * @uc: uclass to destroy + * @flag: driver flags (DM_REMOVE_NORMAL or DM_REMOVE_LATE) * @return 0 on success, -ve on error */ -int uclass_destroy(struct uclass *uc); +int uclass_destroy(struct uclass *uc, unsigned int flag);
#endif diff --git a/test/dm/core.c b/test/dm/core.c index 6f380a574c..43327c29d3 100644 --- a/test/dm/core.c +++ b/test/dm/core.c @@ -82,16 +82,19 @@ void dm_leak_check_start(struct unit_test_state *uts) int dm_leak_check_end(struct unit_test_state *uts) { struct mallinfo end; - int id, diff; + int i, id, diff;
/* Don't delete the root class, since we started with that */ - for (id = UCLASS_ROOT + 1; id < UCLASS_COUNT; id++) { - struct uclass *uc; - - uc = uclass_find(id); - if (!uc) - continue; - ut_assertok(uclass_destroy(uc)); + for (i = 0; i < 2; i++) { + for (id = UCLASS_ROOT + 1; id < UCLASS_COUNT; id++) { + struct uclass *uc; + + uc = uclass_find(id); + if (!uc) + continue; + ut_assertok(uclass_destroy(uc, + i ? DM_REMOVE_LATE : DM_REMOVE_NORMAL)); + } }
end = mallinfo(); @@ -514,7 +517,7 @@ static int dm_test_uclass(struct unit_test_state *uts) ut_asserteq(0, dm_testdrv_op_count[DM_TEST_OP_DESTROY]); ut_assert(uc->priv);
- ut_assertok(uclass_destroy(uc)); + ut_assertok(uclass_destroy(uc, DM_REMOVE_LATE)); ut_asserteq(1, dm_testdrv_op_count[DM_TEST_OP_INIT]); ut_asserteq(1, dm_testdrv_op_count[DM_TEST_OP_DESTROY]);
diff --git a/test/dm/test-main.c b/test/dm/test-main.c index 38b7b1481a..f1c1011d8e 100644 --- a/test/dm/test-main.c +++ b/test/dm/test-main.c @@ -60,19 +60,23 @@ static int do_autoprobe(struct unit_test_state *uts)
static int dm_test_destroy(struct unit_test_state *uts) { - int id; - - for (id = 0; id < UCLASS_COUNT; id++) { - struct uclass *uc; - - /* - * If the uclass doesn't exist we don't want to create it. So - * check that here before we call uclass_find_device(). - */ - uc = uclass_find(id); - if (!uc) - continue; - ut_assertok(uclass_destroy(uc)); + int i, id; + + for (i = 0; i < 2; i++) { + for (id = 0; id < UCLASS_COUNT; id++) { + struct uclass *uc; + + /* + * If the uclass doesn't exist we don't want to + * create it. So check that here before we call + * uclass_find_device(). + */ + uc = uclass_find(id); + if (!uc) + continue; + ut_assertok(uclass_destroy(uc, + i ? DM_REMOVE_LATE : DM_REMOVE_NORMAL)); + } }
return 0;
Once all the other drivers got torn down in preparation for the OS to start, tear down the clock controller last. The clock controller must be torn down last as some of the clock which get turned off might have still been needed during the teardown stage of the other drivers.
Signed-off-by: Marek Vasut marek.vasut+renesas@gmail.com Cc: Nobuhiro Iwamatsu iwamatsu@nigauri.org Cc: Simon Glass sjg@chromium.org Cc: Tom Rini trini@konsulko.com --- drivers/clk/renesas/r8a7795-cpg-mssr.c | 1 + drivers/clk/renesas/r8a7796-cpg-mssr.c | 1 + drivers/clk/renesas/r8a77965-cpg-mssr.c | 1 + drivers/clk/renesas/r8a77970-cpg-mssr.c | 1 + drivers/clk/renesas/r8a77980-cpg-mssr.c | 1 + drivers/clk/renesas/r8a77990-cpg-mssr.c | 1 + drivers/clk/renesas/r8a77995-cpg-mssr.c | 1 + 7 files changed, 7 insertions(+)
diff --git a/drivers/clk/renesas/r8a7795-cpg-mssr.c b/drivers/clk/renesas/r8a7795-cpg-mssr.c index feb41b120d..12b181dcf8 100644 --- a/drivers/clk/renesas/r8a7795-cpg-mssr.c +++ b/drivers/clk/renesas/r8a7795-cpg-mssr.c @@ -377,4 +377,5 @@ U_BOOT_DRIVER(clk_r8a7795) = { .ops = &gen3_clk_ops, .probe = gen3_clk_probe, .remove = gen3_clk_remove, + .flags = DM_FLAG_OS_PREPARE | DM_FLAG_REMOVE_LATE, }; diff --git a/drivers/clk/renesas/r8a7796-cpg-mssr.c b/drivers/clk/renesas/r8a7796-cpg-mssr.c index ef7c554ac6..5ecd906b3a 100644 --- a/drivers/clk/renesas/r8a7796-cpg-mssr.c +++ b/drivers/clk/renesas/r8a7796-cpg-mssr.c @@ -355,4 +355,5 @@ U_BOOT_DRIVER(clk_r8a7796) = { .ops = &gen3_clk_ops, .probe = gen3_clk_probe, .remove = gen3_clk_remove, + .flags = DM_FLAG_OS_PREPARE | DM_FLAG_REMOVE_LATE, }; diff --git a/drivers/clk/renesas/r8a77965-cpg-mssr.c b/drivers/clk/renesas/r8a77965-cpg-mssr.c index f8d954242d..08c51ff183 100644 --- a/drivers/clk/renesas/r8a77965-cpg-mssr.c +++ b/drivers/clk/renesas/r8a77965-cpg-mssr.c @@ -353,4 +353,5 @@ U_BOOT_DRIVER(clk_r8a77965) = { .ops = &gen3_clk_ops, .probe = gen3_clk_probe, .remove = gen3_clk_remove, + .flags = DM_FLAG_OS_PREPARE | DM_FLAG_REMOVE_LATE, }; diff --git a/drivers/clk/renesas/r8a77970-cpg-mssr.c b/drivers/clk/renesas/r8a77970-cpg-mssr.c index c1e9b91b0d..613575900d 100644 --- a/drivers/clk/renesas/r8a77970-cpg-mssr.c +++ b/drivers/clk/renesas/r8a77970-cpg-mssr.c @@ -230,4 +230,5 @@ U_BOOT_DRIVER(clk_r8a77970) = { .ops = &gen3_clk_ops, .probe = gen3_clk_probe, .remove = gen3_clk_remove, + .flags = DM_FLAG_OS_PREPARE | DM_FLAG_REMOVE_LATE, }; diff --git a/drivers/clk/renesas/r8a77980-cpg-mssr.c b/drivers/clk/renesas/r8a77980-cpg-mssr.c index e291774791..d41dd348cd 100644 --- a/drivers/clk/renesas/r8a77980-cpg-mssr.c +++ b/drivers/clk/renesas/r8a77980-cpg-mssr.c @@ -253,4 +253,5 @@ U_BOOT_DRIVER(clk_r8a77980) = { .ops = &gen3_clk_ops, .probe = gen3_clk_probe, .remove = gen3_clk_remove, + .flags = DM_FLAG_OS_PREPARE | DM_FLAG_REMOVE_LATE, }; diff --git a/drivers/clk/renesas/r8a77990-cpg-mssr.c b/drivers/clk/renesas/r8a77990-cpg-mssr.c index a990251ac5..b3ea96a077 100644 --- a/drivers/clk/renesas/r8a77990-cpg-mssr.c +++ b/drivers/clk/renesas/r8a77990-cpg-mssr.c @@ -312,4 +312,5 @@ U_BOOT_DRIVER(clk_r8a77990) = { .ops = &gen3_clk_ops, .probe = gen3_clk_probe, .remove = gen3_clk_remove, + .flags = DM_FLAG_OS_PREPARE | DM_FLAG_REMOVE_LATE, }; diff --git a/drivers/clk/renesas/r8a77995-cpg-mssr.c b/drivers/clk/renesas/r8a77995-cpg-mssr.c index c49a5ffae8..b66b0b46df 100644 --- a/drivers/clk/renesas/r8a77995-cpg-mssr.c +++ b/drivers/clk/renesas/r8a77995-cpg-mssr.c @@ -251,4 +251,5 @@ U_BOOT_DRIVER(clk_r8a77995) = { .ops = &gen3_clk_ops, .probe = gen3_clk_probe, .remove = gen3_clk_remove, + .flags = DM_FLAG_OS_PREPARE | DM_FLAG_REMOVE_LATE, };
Hi Marek,
On Sun, 11 Oct 2020 at 10:03, Marek Vasut marek.vasut@gmail.com wrote:
Add another flag to the DM core which could be assigned to drivers and which makes those drivers call their remove callbacks last, just before booting OS and after all the other drivers finished with their remove callbacks. This is necessary for things like clock drivers, where the other drivers might depend on the clock driver in their remove callbacks. Prime example is the mmc subsystem, which can reconfigure a card from HS mode to slower modes in the remove callback and for that it needs to reconfigure the controller clock.
Signed-off-by: Marek Vasut marek.vasut+renesas@gmail.com Cc: Simon Glass sjg@chromium.org Cc: Tom Rini trini@konsulko.com
arch/arm/lib/bootm.c | 1 + board/Marvell/octeontx2/board.c | 4 ++-- drivers/core/device-remove.c | 11 ++++++++--- drivers/core/root.c | 2 ++ drivers/core/uclass.c | 32 +++++++++++++++++++++++++------- include/dm/device.h | 4 ++++ include/dm/uclass-internal.h | 3 ++- test/dm/core.c | 21 ++++++++++++--------- test/dm/test-main.c | 30 +++++++++++++++++------------- 9 files changed, 73 insertions(+), 35 deletions(-)
I remember seeing a similar patch before. Do you have a change log?
Regards, Simon
On 10/12/20 5:35 AM, Simon Glass wrote:
Hi Marek,
On Sun, 11 Oct 2020 at 10:03, Marek Vasut marek.vasut@gmail.com wrote:
Add another flag to the DM core which could be assigned to drivers and which makes those drivers call their remove callbacks last, just before booting OS and after all the other drivers finished with their remove callbacks. This is necessary for things like clock drivers, where the other drivers might depend on the clock driver in their remove callbacks. Prime example is the mmc subsystem, which can reconfigure a card from HS mode to slower modes in the remove callback and for that it needs to reconfigure the controller clock.
[...]
arch/arm/lib/bootm.c | 1 + board/Marvell/octeontx2/board.c | 4 ++-- drivers/core/device-remove.c | 11 ++++++++--- drivers/core/root.c | 2 ++ drivers/core/uclass.c | 32 +++++++++++++++++++++++++------- include/dm/device.h | 4 ++++ include/dm/uclass-internal.h | 3 ++- test/dm/core.c | 21 ++++++++++++--------- test/dm/test-main.c | 30 +++++++++++++++++------------- 9 files changed, 73 insertions(+), 35 deletions(-)
I remember seeing a similar patch before. Do you have a change log?
Nothing changed since except for the octeon fix, I still require this patch in to correctly stop SD/MMC controller and then the clock before booting Linux.
Hi Marek,
On Mon, 12 Oct 2020 at 02:31, Marek Vasut marek.vasut@gmail.com wrote:
On 10/12/20 5:35 AM, Simon Glass wrote:
Hi Marek,
On Sun, 11 Oct 2020 at 10:03, Marek Vasut marek.vasut@gmail.com wrote:
Add another flag to the DM core which could be assigned to drivers and which makes those drivers call their remove callbacks last, just before booting OS and after all the other drivers finished with their remove callbacks. This is necessary for things like clock drivers, where the other drivers might depend on the clock driver in their remove callbacks. Prime example is the mmc subsystem, which can reconfigure a card from HS mode to slower modes in the remove callback and for that it needs to reconfigure the controller clock.
[...]
arch/arm/lib/bootm.c | 1 + board/Marvell/octeontx2/board.c | 4 ++-- drivers/core/device-remove.c | 11 ++++++++--- drivers/core/root.c | 2 ++ drivers/core/uclass.c | 32 +++++++++++++++++++++++++------- include/dm/device.h | 4 ++++ include/dm/uclass-internal.h | 3 ++- test/dm/core.c | 21 ++++++++++++--------- test/dm/test-main.c | 30 +++++++++++++++++------------- 9 files changed, 73 insertions(+), 35 deletions(-)
I remember seeing a similar patch before. Do you have a change log?
Nothing changed since except for the octeon fix, I still require this patch in to correctly stop SD/MMC controller and then the clock before booting Linux.
Did you see my comments on the previous version of the patch? These need to be addressed.
Regards, Simon
On 10/15/20 5:05 PM, Simon Glass wrote:
[...]
Add another flag to the DM core which could be assigned to drivers and which makes those drivers call their remove callbacks last, just before booting OS and after all the other drivers finished with their remove callbacks. This is necessary for things like clock drivers, where the other drivers might depend on the clock driver in their remove callbacks. Prime example is the mmc subsystem, which can reconfigure a card from HS mode to slower modes in the remove callback and for that it needs to reconfigure the controller clock.
[...]
arch/arm/lib/bootm.c | 1 + board/Marvell/octeontx2/board.c | 4 ++-- drivers/core/device-remove.c | 11 ++++++++--- drivers/core/root.c | 2 ++ drivers/core/uclass.c | 32 +++++++++++++++++++++++++------- include/dm/device.h | 4 ++++ include/dm/uclass-internal.h | 3 ++- test/dm/core.c | 21 ++++++++++++--------- test/dm/test-main.c | 30 +++++++++++++++++------------- 9 files changed, 73 insertions(+), 35 deletions(-)
I remember seeing a similar patch before. Do you have a change log?
Nothing changed since except for the octeon fix, I still require this patch in to correctly stop SD/MMC controller and then the clock before booting Linux.
Did you see my comments on the previous version of the patch? These need to be addressed.
I believe all comments are addressed. What is the problem with this patch ?
On Wed, Oct 21, 2020 at 12:27:45AM +0200, Marek Vasut wrote:
On 10/15/20 5:05 PM, Simon Glass wrote:
[...]
Add another flag to the DM core which could be assigned to drivers and which makes those drivers call their remove callbacks last, just before booting OS and after all the other drivers finished with their remove callbacks. This is necessary for things like clock drivers, where the other drivers might depend on the clock driver in their remove callbacks. Prime example is the mmc subsystem, which can reconfigure a card from HS mode to slower modes in the remove callback and for that it needs to reconfigure the controller clock.
[...]
arch/arm/lib/bootm.c | 1 + board/Marvell/octeontx2/board.c | 4 ++-- drivers/core/device-remove.c | 11 ++++++++--- drivers/core/root.c | 2 ++ drivers/core/uclass.c | 32 +++++++++++++++++++++++++------- include/dm/device.h | 4 ++++ include/dm/uclass-internal.h | 3 ++- test/dm/core.c | 21 ++++++++++++--------- test/dm/test-main.c | 30 +++++++++++++++++------------- 9 files changed, 73 insertions(+), 35 deletions(-)
I remember seeing a similar patch before. Do you have a change log?
Nothing changed since except for the octeon fix, I still require this patch in to correctly stop SD/MMC controller and then the clock before booting Linux.
Did you see my comments on the previous version of the patch? These need to be addressed.
I believe all comments are addressed. What is the problem with this patch ?
Um. This patch doesn't say "v2" or "v3" or anything, and you said the only change you made since last posting was adding the octeon changes. I see a bit of a thread in https://patchwork.ozlabs.org/project/uboot/patch/20200802150640.114716-1-mar... where Simon and Stefan expanded on things and no further follow up from you. Did you address everything in that thread?
On 10/21/20 1:23 AM, Tom Rini wrote:
On Wed, Oct 21, 2020 at 12:27:45AM +0200, Marek Vasut wrote:
On 10/15/20 5:05 PM, Simon Glass wrote:
[...]
Add another flag to the DM core which could be assigned to drivers and which makes those drivers call their remove callbacks last, just before booting OS and after all the other drivers finished with their remove callbacks. This is necessary for things like clock drivers, where the other drivers might depend on the clock driver in their remove callbacks. Prime example is the mmc subsystem, which can reconfigure a card from HS mode to slower modes in the remove callback and for that it needs to reconfigure the controller clock.
[...]
arch/arm/lib/bootm.c | 1 + board/Marvell/octeontx2/board.c | 4 ++-- drivers/core/device-remove.c | 11 ++++++++--- drivers/core/root.c | 2 ++ drivers/core/uclass.c | 32 +++++++++++++++++++++++++------- include/dm/device.h | 4 ++++ include/dm/uclass-internal.h | 3 ++- test/dm/core.c | 21 ++++++++++++--------- test/dm/test-main.c | 30 +++++++++++++++++------------- 9 files changed, 73 insertions(+), 35 deletions(-)
I remember seeing a similar patch before. Do you have a change log?
Nothing changed since except for the octeon fix, I still require this patch in to correctly stop SD/MMC controller and then the clock before booting Linux.
Did you see my comments on the previous version of the patch? These need to be addressed.
I believe all comments are addressed. What is the problem with this patch ?
Um. This patch doesn't say "v2" or "v3" or anything, and you said the only change you made since last posting was adding the octeon changes. I see a bit of a thread in https://patchwork.ozlabs.org/project/uboot/patch/20200802150640.114716-1-mar... where Simon and Stefan expanded on things and no further follow up from you. Did you address everything in that thread?
I think so.
Hi Marek,
On Tue, 20 Oct 2020 at 17:37, Marek Vasut marex@denx.de wrote:
On 10/21/20 1:23 AM, Tom Rini wrote:
On Wed, Oct 21, 2020 at 12:27:45AM +0200, Marek Vasut wrote:
On 10/15/20 5:05 PM, Simon Glass wrote:
[...]
> Add another flag to the DM core which could be assigned to drivers and > which makes those drivers call their remove callbacks last, just before > booting OS and after all the other drivers finished with their remove > callbacks. This is necessary for things like clock drivers, where the > other drivers might depend on the clock driver in their remove callbacks. > Prime example is the mmc subsystem, which can reconfigure a card from HS > mode to slower modes in the remove callback and for that it needs to > reconfigure the controller clock.
[...]
> arch/arm/lib/bootm.c | 1 + > board/Marvell/octeontx2/board.c | 4 ++-- > drivers/core/device-remove.c | 11 ++++++++--- > drivers/core/root.c | 2 ++ > drivers/core/uclass.c | 32 +++++++++++++++++++++++++------- > include/dm/device.h | 4 ++++ > include/dm/uclass-internal.h | 3 ++- > test/dm/core.c | 21 ++++++++++++--------- > test/dm/test-main.c | 30 +++++++++++++++++------------- > 9 files changed, 73 insertions(+), 35 deletions(-)
I remember seeing a similar patch before. Do you have a change log?
Nothing changed since except for the octeon fix, I still require this patch in to correctly stop SD/MMC controller and then the clock before booting Linux.
Did you see my comments on the previous version of the patch? These need to be addressed.
I believe all comments are addressed. What is the problem with this patch ?
Um. This patch doesn't say "v2" or "v3" or anything, and you said the only change you made since last posting was adding the octeon changes. I see a bit of a thread in https://patchwork.ozlabs.org/project/uboot/patch/20200802150640.114716-1-mar... where Simon and Stefan expanded on things and no further follow up from you. Did you address everything in that thread?
I think so.
I can't see it. See here for my feedback:
https://patchwork.ozlabs.org/project/uboot/patch/20200802150640.114716-1-mar...
Also please do include a changelog with the next version.
Regards, Simon
On 10/12/20 10:31 AM, Marek Vasut wrote:
[...]
Add another flag to the DM core which could be assigned to drivers and which makes those drivers call their remove callbacks last, just before booting OS and after all the other drivers finished with their remove callbacks. This is necessary for things like clock drivers, where the other drivers might depend on the clock driver in their remove callbacks. Prime example is the mmc subsystem, which can reconfigure a card from HS mode to slower modes in the remove callback and for that it needs to reconfigure the controller clock.
[...]
arch/arm/lib/bootm.c | 1 + board/Marvell/octeontx2/board.c | 4 ++-- drivers/core/device-remove.c | 11 ++++++++--- drivers/core/root.c | 2 ++ drivers/core/uclass.c | 32 +++++++++++++++++++++++++------- include/dm/device.h | 4 ++++ include/dm/uclass-internal.h | 3 ++- test/dm/core.c | 21 ++++++++++++--------- test/dm/test-main.c | 30 +++++++++++++++++------------- 9 files changed, 73 insertions(+), 35 deletions(-)
I remember seeing a similar patch before. Do you have a change log?
Nothing changed since except for the octeon fix, I still require this patch in to correctly stop SD/MMC controller and then the clock before booting Linux.
This patch is still missing from upstream, and I see no feedback. I have multiple patches depending on this functionality too, and it already missed two merge windows.
participants (4)
-
Marek Vasut -
Marek Vasut -
Simon Glass -
Tom Rini