March 2021 - U-Boot

[PATCH] arm: stm32mp1: Set soc_type, soc_pkg, soc_rev env variables
by Marek Vasut 11 Mar '21

11 Mar '21

Split up get_soc_name(), clean the decoding up a bit, and set up environment variables which contain the SoC type, package, revision. This is useful on SoMs, where multiple SoC options are populated. Signed-off-by: Marek Vasut <marex(a)denx.de> Cc: Patrick Delaunay <patrick.delaunay(a)st.com> Cc: Patrice Chotard <patrice.chotard(a)st.com> --- arch/arm/mach-stm32mp/cpu.c | 105 ++++++++++++++++++------------------ 1 file changed, 53 insertions(+), 52 deletions(-) diff --git a/arch/arm/mach-stm32mp/cpu.c b/arch/arm/mach-stm32mp/cpu.c index 3faa4ec18a4..887db9b5b0c 100644 --- a/arch/arm/mach-stm32mp/cpu.c +++ b/arch/arm/mach-stm32mp/cpu.c @@ -349,89 +349,78 @@ u32 get_cpu_package(void) return get_otp(BSEC_OTP_PKG, PKG_SHIFT, PKG_MASK); } -void get_soc_name(char name[SOC_NAME_SIZE]) +static const char * const soc_type[] = { + "????", + "151C", "151A", "151F", "151D", + "153C", "153A", "153F", "153D", + "157C", "157A", "157F", "157D" +}; + +static const char * const soc_pkg[] = { "??", "AD", "AC", "AB", "AA" }; +static const char * const soc_rev[] = { "?", "A", "B", "Z" }; + +static void get_cpu_string_offsets(unsigned int *type, unsigned int *pkg, + unsigned int *rev) { - char *cpu_s, *cpu_r, *pkg; + u32 cpu_type = get_cpu_type(); + u32 ct = cpu_type & ~(BIT(7) | BIT(0)); + u32 cm = ((cpu_type & BIT(7)) >> 6) | (cpu_type & BIT(0)); + u32 cp = get_cpu_package(); - /* MPUs Part Numbers */ - switch (get_cpu_type()) { - case CPU_STM32MP157Fxx: - cpu_s = "157F"; - break; - case CPU_STM32MP157Dxx: - cpu_s = "157D"; - break; - case CPU_STM32MP157Cxx: - cpu_s = "157C"; - break; - case CPU_STM32MP157Axx: - cpu_s = "157A"; - break; - case CPU_STM32MP153Fxx: - cpu_s = "153F"; - break; - case CPU_STM32MP153Dxx: - cpu_s = "153D"; + /* Bits 0 and 7 are the ACDF, 00:C 01:A 10:F 11:D */ + switch (ct) { + case CPU_STM32MP151Cxx: + *type = cm + 1; break; case CPU_STM32MP153Cxx: - cpu_s = "153C"; - break; - case CPU_STM32MP153Axx: - cpu_s = "153A"; - break; - case CPU_STM32MP151Fxx: - cpu_s = "151F"; - break; - case CPU_STM32MP151Dxx: - cpu_s = "151D"; + *type = cm + 5; break; - case CPU_STM32MP151Cxx: - cpu_s = "151C"; - break; - case CPU_STM32MP151Axx: - cpu_s = "151A"; + case CPU_STM32MP157Cxx: + *type = cm + 9; break; default: - cpu_s = "????"; + *type = 0; break; } /* Package */ - switch (get_cpu_package()) { + switch (cp) { case PKG_AA_LBGA448: - pkg = "AA"; - break; case PKG_AB_LBGA354: - pkg = "AB"; - break; case PKG_AC_TFBGA361: - pkg = "AC"; - break; case PKG_AD_TFBGA257: - pkg = "AD"; + *pkg = cp; break; default: - pkg = "??"; + pkg = 0; break; } - /* REVISION */ + /* Revision */ switch (get_cpu_rev()) { case CPU_REVA: - cpu_r = "A"; + *rev = 1; break; case CPU_REVB: - cpu_r = "B"; + *rev = 2; break; case CPU_REVZ: - cpu_r = "Z"; + *rev = 3; break; default: - cpu_r = "?"; + *rev = 0; break; } +} + +void get_soc_name(char name[SOC_NAME_SIZE]) +{ + unsigned int type, pkg, rev; - snprintf(name, SOC_NAME_SIZE, "STM32MP%s%s Rev.%s", cpu_s, pkg, cpu_r); + get_cpu_string_offsets(&type, &pkg, &rev); + + snprintf(name, SOC_NAME_SIZE, "STM32MP%s%s Rev.%s", + soc_type[type], soc_pkg[pkg], soc_rev[rev]); } #if defined(CONFIG_DISPLAY_CPUINFO) @@ -620,11 +609,23 @@ static int setup_serial_number(void) return 0; } +static void setup_soc_type_pkg_rev(void) +{ + unsigned int type, pkg, rev; + + get_cpu_string_offsets(&type, &pkg, &rev); + + env_set("soc_type", soc_type[type]); + env_set("soc_pkg", soc_pkg[pkg]); + env_set("soc_rev", soc_rev[rev]); +} + int arch_misc_init(void) { setup_boot_mode(); setup_mac_address(); setup_serial_number(); + setup_soc_type_pkg_rev(); return 0; } -- 2.30.1

3 2

[BUG] double free in dfu_free_entities()
by Heinrich Schuchardt 11 Mar '21

11 Mar '21

Hello Lukasz, Jose and I have observed segmentation violations when dfu_free_entities() is called. In our scenario we have: dfu_alt_info= sf 0:0=u-boot-bin raw 0x100000 0x50000;u-boot-env raw 0x150000 0x200000 dfu_free_entities() is called with these entries in dfu: u-boot-env: dfu->data.sf.dev 0000000015baf420 m25p16 dfu->data.sf.dev->dev 0000000015888410 spi.bin@0 u-boot-bin: dfu->data.sf.dev 0000000015baf420 m25p16 dfu->data.sf.dev->dev 0000000015888410 spi.bin@0 The same device is released twice which eventually leads to the segmentation fault in device_chld_remove(). Program received signal SIGSEGV, Segmentation fault. device_chld_remove ( dev=dev@entry=0x5555557f6750 <av_+1936>, drv=drv@entry=0x0, flags=flags@entry=1) at drivers/core/device-remove.c:55 55 list_for_each_entry_safe(pos, n, &dev->child_head, sibling_node) { We must ensure that dfu_free_entities() removes each device only once. Another bug in dfu_free_entities() is that only the first list member is freed. free() should be called for all list members. Best regards Heinrich

2 2

[PATCH] mtd: rawnand: fsl_ifc: fix FSL NAND driver to read all ONFI parameter pages
by Maxim Kochetkov 11 Mar '21

11 Mar '21

Linux commit a75bbe71a27 ("mtd: rawnand: fsl_ifc: fix FSL NAND driver to read all ONFI parameter pages") Per ONFI specification (Rev. 4.0), if the CRC of the first parameter page read is not valid, the host should read redundant parameter page copies. Fix FSL NAND driver to read the two redundant copies which are mandatory in the specification. Signed-off-by: Jane Wan <Jane.Wan(a)nokia.com> Signed-off-by: Boris Brezillon <boris.brezillon(a)bootlin.com> Signed-off-by: Maxim Kochetkov <fido_max(a)inbox.ru> --- drivers/mtd/nand/raw/fsl_ifc_nand.c | 17 ++++++++++------- 1 file changed, 10 insertions(+), 7 deletions(-) diff --git a/drivers/mtd/nand/raw/fsl_ifc_nand.c b/drivers/mtd/nand/raw/fsl_ifc_nand.c index cf20238782..e5ff937872 100644 --- a/drivers/mtd/nand/raw/fsl_ifc_nand.c +++ b/drivers/mtd/nand/raw/fsl_ifc_nand.c @@ -411,9 +411,16 @@ static void fsl_ifc_cmdfunc(struct mtd_info *mtd, unsigned int command, /* READID must read all possible bytes while CEB is active */ case NAND_CMD_READID: case NAND_CMD_PARAM: { + /* + * For READID, read 8 bytes that are currently used. + * For PARAM, read all 3 copies of 256-bytes pages. + */ + int len = 8; int timing = IFC_FIR_OP_RB; - if (command == NAND_CMD_PARAM) + if (command == NAND_CMD_PARAM) { timing = IFC_FIR_OP_RBCD; + len = 256 * 3; + } ifc_out32(&ifc->ifc_nand.nand_fir0, (IFC_FIR_OP_CW0 << IFC_NAND_FIR0_OP0_SHIFT) | @@ -423,12 +430,8 @@ static void fsl_ifc_cmdfunc(struct mtd_info *mtd, unsigned int command, command << IFC_NAND_FCR0_CMD0_SHIFT); ifc_out32(&ifc->ifc_nand.row3, column); - /* - * although currently it's 8 bytes for READID, we always read - * the maximum 256 bytes(for PARAM) - */ - ifc_out32(&ifc->ifc_nand.nand_fbcr, 256); - ctrl->read_bytes = 256; + ifc_out32(&ifc->ifc_nand.nand_fbcr, len); + ctrl->read_bytes = len; set_addr(mtd, 0, 0, 0); fsl_ifc_run_command(mtd); -- 2.30.2

1 0

[PATCH] checkpatch: Add warnings for using strn(cat|cpy)
by Sean Anderson 11 Mar '21

11 Mar '21

strn(cat|cpy) has a bad habit of not nul-terminating the destination, resulting in constructions like strncpy(foo, bar, sizeof(foo) - 1); foo[sizeof(foo) - 1] = '\0'; However, it is very easy to forget about this behavior and accidentally leave a string unterminated. This has shown up in some recent coverity scans [1, 2] (including code recently touched by yours truly). Fortunately, the guys at OpenBSD came up with strl(cat|cpy), which always nul-terminate strings. These functions are already in U-Boot, so we should encourage new code to use them instead of strn(cat|cpy). [1] https://lists.denx.de/pipermail/u-boot/2021-March/442888.html [2] https://lists.denx.de/pipermail/u-boot/2021-January/438073.html Signed-off-by: Sean Anderson <seanga2(a)gmail.com> --- scripts/checkpatch.pl | 6 ++++++ tools/patman/test_checkpatch.py | 14 +++++++++++++- 2 files changed, 19 insertions(+), 1 deletion(-) diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl index 755f4802a4..91365a5529 100755 --- a/scripts/checkpatch.pl +++ b/scripts/checkpatch.pl @@ -5892,6 +5892,12 @@ sub process { } } +# prefer strl(cpy|cat) over strn(cpy|cat) + if ($line =~ /\bstrn(cpy|cat)\s*$/) { + WARN("STRL", + "strl$1 is preferred over strn$1 because it always produces a nul-terminated string\n" . $herecurr); + } + # prefer usleep_range over udelay if ($line =~ /\budelay\s*\(\s*(\d+)\s*$/) { my $delay = $1; diff --git a/tools/patman/test_checkpatch.py b/tools/patman/test_checkpatch.py index a4fec1d4c1..56af5265cc 100644 --- a/tools/patman/test_checkpatch.py +++ b/tools/patman/test_checkpatch.py @@ -353,7 +353,7 @@ index 0000000..2234c87 Args: pm: PatchMaker object to use - msg" Expected message (e.g. 'LIVETREE') + msg: Expected message (e.g. 'LIVETREE') pmtype: Type of problem ('error', 'warning') """ result = pm.run_checkpatch() @@ -439,6 +439,18 @@ index 0000000..2234c87 self.check_struct('per_device_auto', '_priv', 'DEVICE_PRIV_AUTO') self.check_struct('per_device_plat_auto', '_plat', 'DEVICE_PLAT_AUTO') + def check_strl(self, func): + """Check one of the checks for strn(cpy|cat)""" + pm = PatchMaker() + pm.add_line('common/main.c', "strn%s(foo, bar, sizeof(foo));" % func) + self.checkSingleMessage(pm, "STRL", + "strl%s is preferred over strn%s because it always produces a nul-terminated string\n" + % (func, func)) + + def testStrl(self): + """Check for uses of strn(cat|cpy)""" + self.check_strl("cat"); + self.check_strl("cpy"); if __name__ == "__main__": unittest.main() -- 2.30.1

2 2

[PATCH v2] riscv: Remove unused define in maix header
by Sean Anderson 11 Mar '21

11 Mar '21

This define was left over from a previous revision, and was never used. Signed-off-by: Sean Anderson <seanga2(a)gmail.com> Reviewed-by: Heinrich Schuchardt <xypron.glpk(a)gmx.de> Reviewed-by: Bin Meng <bin.meng(a)windriver.com> --- Changes in v2: - Rebase onto u-boot/master include/configs/sipeed-maix.h | 3 --- 1 file changed, 3 deletions(-) diff --git a/include/configs/sipeed-maix.h b/include/configs/sipeed-maix.h index 08acb25075..4c1ff98ec6 100644 --- a/include/configs/sipeed-maix.h +++ b/include/configs/sipeed-maix.h @@ -18,9 +18,6 @@ /* Don't relocate into AI ram since it isn't set up yet */ #define CONFIG_SYS_SDRAM_SIZE (SZ_4M + SZ_2M) -/* For early init */ -#define K210_SYSCTL_BASE 0x50440000 - #ifndef CONFIG_EXTRA_ENV_SETTINGS #define CONFIG_EXTRA_ENV_SETTINGS \ "loadaddr=0x80060000\0" \ -- 2.30.1

1 0

[PATCH 0/4 v6] Add ESRT and test ESRT creation
by Jose Marinho 10 Mar '21

10 Mar '21

The following 4 commits add the ESRT and provide tests for the functionality. The first commit adds the ESRT as defined in the UEFI 2.8 specification. An empty ESRT is created during the execution of the efi_init_obj_list(). The ESRT is updated when: 1) a FMP protocol is installed in the system: this will add the corresponding entries to the ESRT. 2) a capsule is installed via UpdateCapsule: this should update entries already present in the ESRT. This implementation of the ESRT creation takes input from FMP only. It is assumed that the FMP will maintain the following values across reboot: - LastAttemptVersion. - LastAttemptStatus. The second commit introduces a helper cmd function to print the ESRT in the u-boot shell. The third and forth commits contain two test for the ESRT creation in the sandbox platform. - Test 1 executes from the u-boot shell with "bootefi selftest". - Test 2 executes in the pytest environment. Patch v6: - split the v5 3/3 commit into v6 3/4 (efi selftest) and v6 4/4 (addition to test_efi_capsule_fw3). - Address v5 comments. Patch v5: - Address v4 comments. - split the v4 2/2 commit into the v5 2/3 (debug print ESRT) and 3/3 (ESRT tests). Patch v4: - update stale [Patch 1/2 v3] commit message. Patch v3: - Address v2 comments. Patch v2: - The ESRT is now regenerated from scratch at every FMP EVT_NOTIFY_SIGNAL and whenever a capsule is updated. - Extended TestEfiCapsuleFirmwareFit::test_efi_capsule_fw3 to verify that the ESRT is correctly populated after an UpdateCapsule. - Addressed v1 comments. Patch v1: - reworked the ESRT creation code, allowing table to resize as FMPs are installed. - registered a callback for the FMP protocol install. - Created a unit test running on the sandbox platform. rfc: initial patch submission CC: Heinrich Schuchardt <xypron.glpk(a)gmx.de> CC: Sughosh Ganu <sughosh.ganu(a)linaro.org> CC: AKASHI Takahiro <takahiro.akashi(a)linaro.org> CC: Ilias Apalodimas <ilias.apalodimas(a)linaro.org> CC: Andre Przywara <andre.przywara(a)arm.com> CC: Alexander Graf <agraf(a)csgraf.de> CC: nd(a)arm.com Jose Marinho (4): efi: Add ESRT to the EFI system table cmd: efi: ESRT table debug print efi: ESRT creation tests Test the ESRT creation within an existing pytest. cmd/efidebug.c | 70 +++ include/efi_api.h | 33 ++ include/efi_loader.h | 24 + lib/efi_loader/Kconfig | 7 + lib/efi_loader/Makefile | 1 + lib/efi_loader/efi_boottime.c | 7 +- lib/efi_loader/efi_capsule.c | 8 + lib/efi_loader/efi_esrt.c | 518 ++++++++++++++++++ lib/efi_loader/efi_setup.c | 6 + lib/efi_selftest/Makefile | 2 + lib/efi_selftest/efi_selftest_esrt.c | 264 +++++++++ .../test_efi_capsule/test_capsule_firmware.py | 8 + 12 files changed, 944 insertions(+), 4 deletions(-) create mode 100644 lib/efi_loader/efi_esrt.c create mode 100644 lib/efi_selftest/efi_selftest_esrt.c -- 2.17.1

2 7

[PATCH v5 0/2] gpio: Update and simplify the uclass API
by Simon Glass 10 Mar '21

10 Mar '21

At present the GPIO uclass mirrors what was in U-Boot before driver model. It works well in most cases but is becoming cumbersome with things like pull-up/down and drive strength. In those cases it is easier for the driver to deal with all the flags at one, rather than piece by piece. In fact the current API does not officially have a method for adjusting anything other than the direction flags. While set_dir_flags() and get_dir_flags() do in fact support changing other flags, this is not documented. Secondly, set_dir_flags actually ORs the current flags with the new ones so it is not possible to clear flags. It seems better to use a clr/set interface (bit clear followed by OR) to provide more flexibility. Finally, direction_input() and direction_output() are really just the same thing as set_dir_flags(), with a different name. We currently use the latter if available, failing back to the former. But it makes sense to deprecate the old methods. This series makes the above changes. Existing drivers are mostly left alone, since they should continue to operate as is. The sandbox driver is updated to add the required new tests and the x86 driver is switched over to the new API. The STM32 driver should be checked to make sure the open source/open drain features still work as intended. Changes in v5: - Drop patches previously applied Simon Glass (2): gpio: i2c-gpio: Drop use of dm_gpio_set_dir() gpio: Drop dm_gpio_set_dir() drivers/gpio/gpio-uclass.c | 11 ----------- drivers/i2c/i2c-gpio.c | 19 +++++++++---------- include/asm-generic/gpio.h | 11 ----------- 3 files changed, 9 insertions(+), 32 deletions(-) -- 2.30.1.766.gb4fecdf3b7-goog

4 8

[PATCH] tee: optee: Change printing during optee_probe
by Ilias Apalodimas 10 Mar '21

10 Mar '21

Right now the error messages when optee has a version mismatch or shared memory is not configured are done with a debug(). That's not very convenient since you have to enable debugging to figure out what's going on, although this is an actual error. So let's switch the debug() -> log_err() and report those explicitly. Signed-off-by: Ilias Apalodimas <ilias.apalodimas(a)linaro.org> --- drivers/tee/optee/core.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/tee/optee/core.c b/drivers/tee/optee/core.c index b898c32edc0b..1b0a7c0c9668 100644 --- a/drivers/tee/optee/core.c +++ b/drivers/tee/optee/core.c @@ -624,14 +624,14 @@ static int optee_probe(struct udevice *dev) u32 sec_caps; if (!is_optee_api(pdata->invoke_fn)) { - debug("%s: OP-TEE api uid mismatch\n", __func__); + log_err("%s: OP-TEE api uid mismatch\n", __func__); return -ENOENT; } print_os_revision(dev, pdata->invoke_fn); if (!api_revision_is_compatible(pdata->invoke_fn)) { - debug("%s: OP-TEE api revision mismatch\n", __func__); + log_err("%s: OP-TEE api revision mismatch\n", __func__); return -ENOENT; } @@ -642,7 +642,7 @@ static int optee_probe(struct udevice *dev) */ if (!exchange_capabilities(pdata->invoke_fn, &sec_caps) || !(sec_caps & OPTEE_SMC_SEC_CAP_DYNAMIC_SHM)) { - debug("%s: OP-TEE capabilities mismatch\n", __func__); + log_err("%s: OP-TEE capabilities mismatch\n", __func__); return -ENOENT; } -- 2.30.1

2 2

[IMPORTANT] gitlab relocation / rename
by Wolfgang Denk 10 Mar '21

10 Mar '21

Hi all, I'm sorry that I have to inform you about a disruption of the gitlab services. This has become necessary for a number of reasons: - repeated cases of poor performance - difficulties with CI setup - security issues We have to move all public projects off gitlab.denx.de to a new gitlab instance; the new host name is then source.denx.de . We are fully aware that this will cause inconveniences, and I can only assure you that we try hard to keep these as limited as possible. # Details Around noon on Saturday, 2021-02-27, we will freeze all public repositories on gitlab.denx.de. They will still be readable/ cloneable but any attempt to push new commits after this point will fail. We will then copy them over to the now host, source.denx.de . All custodian accounts will also be transferred, the same login credentials from gitlab.denx.de will work on source.denx.de . The URLs for the repositories will stay the same, except of course for the host name. So, for example, the U-Boot mainline repository will then be hosted at https://source.denx.de/u-boot/u-boot.git We will install redirects to forward HTTP accesses from the old repository URLs to the new host. This should make the transition mostly transparent, but does not cover everything. The following needs to be taken care of manually: - Update the SSH URI for pushing to the repositories. Just replace gitlab.denx.de with source.denx.de . We will transfer all the SSH-keys to the new host so nothing else should be needed. - Make sure you're logging in on source.denx.de . Access to gitlab.denx.de will no longer work! - CI runners connected to gitlab.denx.de need to be re-registered with source.denx.de . The new host should hopefully be operational sometime on Sunday, 2021-02-28. We will send a further announcement once everything is back up and running. Thanks for your understanding and patience! Best regards, Wolfgang Denk -- DENX Software Engineering GmbH, Managing Director: Wolfgang Denk HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd(a)denx.de At the source of every error which is blamed on the computer you will find at least two human errors, including the error of blaming it on the computer.

2 4

[PATCH v2] usb: ehci-hcd: Add IAA handshake for removing async QH
by Ye Li 10 Mar '21

10 Mar '21

According to EHCI spec, software needs to do handshake with HC for safely removing QH from async list. This handshake is implemented by setting IAAD (Interrupt on Async Advance Doorbell) bit in USB_USBCMD register and poll the IAA (Interrupt on Async Advance bit) in the USB_USBSTS to ensure the HC has released all on-chip state that may potentially reference one of the data structures just removed. Current codes only check active status of the last QTD, but this can't ensure the QH is released from HC. We can meet unrecoverable "EHCI timed out on TD" errors when running UEFI SCT tests on USB disk. The USB_ASYNCLISTADDR register is changed to a invalid address when the issue happens. It is fixed after adding the IAA handshake. Steps to reproduce the issue: 1. Build the UEFI SCT from https://github.com/tianocore/edk2-test 2. Build the EDK2 UEFI Shell from https://github.com/tianocore/edk2 3. Copy SCT files and Shell.efi to USB disk FAT partition 4. Load the Shell.efi from USB FAT, and run bootefi to execute it 5. After booting into Shell, enter the SCT directory and run "sct -a" to execute all tests. 6. Tests run about 1 hour and stop with many EHCI timeout errors like EHCI timed out on TD - token=0x801f8c80 Signed-off-by: Ye Li <ye.li(a)nxp.com> --- Changes in v2: - Remove unnecessary cast and parenthesis - Abort the transfer when IAA cycle timeout - Add steps to reproduce the issue drivers/usb/host/ehci-hcd.c | 27 +++++++++++++++++++++++++++ drivers/usb/host/ehci.h | 1 + 2 files changed, 28 insertions(+) diff --git a/drivers/usb/host/ehci-hcd.c b/drivers/usb/host/ehci-hcd.c index 8933f60..ba75c27 100644 --- a/drivers/usb/host/ehci-hcd.c +++ b/drivers/usb/host/ehci-hcd.c @@ -346,6 +346,28 @@ static int ehci_disable_async(struct ehci_ctrl *ctrl) return ret; } +static int ehci_iaa_cycle(struct ehci_ctrl *ctrl) +{ + u32 cmd, status; + int ret; + + /* Enable Interrupt on Async Advance Doorbell. */ + cmd = ehci_readl(&ctrl->hcor->or_usbcmd); + cmd |= CMD_IAAD; + ehci_writel(&ctrl->hcor->or_usbcmd, cmd); + + ret = handshake(&ctrl->hcor->or_usbsts, STS_IAA, STS_IAA, + 10 * 1000); /* 10ms timeout */ + if (ret < 0) + printf("EHCI fail timeout STS_IAA set\n"); + + status = ehci_readl(&ctrl->hcor->or_usbsts); + if (status & STS_IAA) + ehci_writel(&ctrl->hcor->or_usbsts, STS_IAA); + + return ret; +} + static int ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer, int length, struct devrequest *req) @@ -631,6 +653,11 @@ ehci_submit_async(struct usb_device *dev, unsigned long pipe, void *buffer, flush_dcache_range((unsigned long)&ctrl->qh_list, ALIGN_END_ADDR(struct QH, &ctrl->qh_list, 1)); + /* Set IAAD, poll IAA */ + ret = ehci_iaa_cycle(ctrl); + if (ret) + goto fail; + /* * Invalidate the memory area occupied by buffer * Don't try to fix the buffer alignment, if it isn't properly diff --git a/drivers/usb/host/ehci.h b/drivers/usb/host/ehci.h index 8e07554..e9e6f2a5 100644 --- a/drivers/usb/host/ehci.h +++ b/drivers/usb/host/ehci.h @@ -44,6 +44,7 @@ struct ehci_hcor { #define STS_ASS (1 << 15) #define STS_PSS (1 << 14) #define STS_HALT (1 << 12) +#define STS_IAA (1 << 5) uint32_t or_usbintr; #define INTR_UE (1 << 0) /* USB interrupt enable */ #define INTR_UEE (1 << 1) /* USB error interrupt enable */ -- 2.7.4

2 1