* EFI stub boots only with setup_efi_pc() commented
@ 2013-08-02 23:06 Clea F. Rees
[not found] ` <CA+n4xYYJhoXp6KSeROda9ErrXZ0hFEBM-x-0Obvk3fmHwABfKw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
0 siblings, 1 reply; 7+ messages in thread
From: Clea F. Rees @ 2013-08-02 23:06 UTC (permalink / raw)
To: linux-efi-u79uwXL29TY76Z2rM5mHXA
Cc: matthew.garrett-05XSO3Yj/JvQT0dZR+AlfA, Matt Fleming
Matt Fleming asked me to email describing a problem I am experiencing
with the EFI stub loader. Matt has been trying to diagnose the cause
of consistent failures of the stub loader in recent kernels on my
machine.
Background: All kernels I tried in 3.7.* booted fine using the stub
loader (via rEFInd). No kernels tried from 3.8.* through 3.9.* and now
into 3.10.* boot using the stub loader. None of the workarounds which
others have used work for me. The problem is reproducible using a
direct NVRAM entry rather than rEFInd. All kernels boot fine with
GRUB.
The problem is reproducible with a kernel compiled from git. A series
of experiments (following Matt's instructions) have determined that
commenting out the call to setup_efi_pc() in eboot.c makes the kernels
boot.
I would be happy to provide any further information which might be
helpful or to run further tests.
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: EFI stub boots only with setup_efi_pc() commented
[not found] ` <CA+n4xYYJhoXp6KSeROda9ErrXZ0hFEBM-x-0Obvk3fmHwABfKw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
@ 2013-09-05 8:29 ` Matt Fleming
[not found] ` <20130905082924.GI28598-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>
0 siblings, 1 reply; 7+ messages in thread
From: Matt Fleming @ 2013-09-05 8:29 UTC (permalink / raw)
To: Clea F. Rees
Cc: linux-efi-u79uwXL29TY76Z2rM5mHXA,
matthew.garrett-05XSO3Yj/JvQT0dZR+AlfA, Matt Fleming
On Sat, 03 Aug, at 12:06:23AM, Clea F. Rees wrote:
> Matt Fleming asked me to email describing a problem I am experiencing
> with the EFI stub loader. Matt has been trying to diagnose the cause
> of consistent failures of the stub loader in recent kernels on my
> machine.
>
> Background: All kernels I tried in 3.7.* booted fine using the stub
> loader (via rEFInd). No kernels tried from 3.8.* through 3.9.* and now
> into 3.10.* boot using the stub loader. None of the workarounds which
> others have used work for me. The problem is reproducible using a
> direct NVRAM entry rather than rEFInd. All kernels boot fine with
> GRUB.
>
> The problem is reproducible with a kernel compiled from git. A series
> of experiments (following Matt's instructions) have determined that
> commenting out the call to setup_efi_pc() in eboot.c makes the kernels
> boot.
>
> I would be happy to provide any further information which might be
> helpful or to run further tests.
Sorry for the delay in replying to this.
Does your machine hang during setup_efi_pci() or later on in boot?
--
Matt Fleming, Intel Open Source Technology Center
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: EFI stub boots only with setup_efi_pc() commented
[not found] ` <20130905082924.GI28598-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>
@ 2013-09-09 11:45 ` Clea F. Rees
[not found] ` <CA+n4xYZq3wpY_j+w7BKJ1hV8s1=K-xqqYucZen8os7O_g71gSg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
0 siblings, 1 reply; 7+ messages in thread
From: Clea F. Rees @ 2013-09-09 11:45 UTC (permalink / raw)
To: Matt Fleming
Cc: linux-efi-u79uwXL29TY76Z2rM5mHXA,
matthew.garrett-05XSO3Yj/JvQT0dZR+AlfA, Matt Fleming
2013/9/5 Matt Fleming <matt-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>:
> On Sat, 03 Aug, at 12:06:23AM, Clea F. Rees wrote:
>> Matt Fleming asked me to email describing a problem I am experiencing
>> with the EFI stub loader. Matt has been trying to diagnose the cause
>> of consistent failures of the stub loader in recent kernels on my
>> machine.
>>
>> Background: All kernels I tried in 3.7.* booted fine using the stub
>> loader (via rEFInd). No kernels tried from 3.8.* through 3.9.* and now
>> into 3.10.* boot using the stub loader. None of the workarounds which
>> others have used work for me. The problem is reproducible using a
>> direct NVRAM entry rather than rEFInd. All kernels boot fine with
>> GRUB.
>>
>> The problem is reproducible with a kernel compiled from git. A series
>> of experiments (following Matt's instructions) have determined that
>> commenting out the call to setup_efi_pc() in eboot.c makes the kernels
>> boot.
>>
>> I would be happy to provide any further information which might be
>> helpful or to run further tests.
>
> Sorry for the delay in replying to this.
>
> Does your machine hang during setup_efi_pci() or later on in boot?
I'm not certain. How would I tell this? I think the answer is probably
during because the tests Matt had me do involved inserting reboot
instructions into the code and testing where the hang occurred. But
I'm not sure whether that showed absolutely that the hang occurs
actually during setup_efi_pci().
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: EFI stub boots only with setup_efi_pc() commented
[not found] ` <CA+n4xYZq3wpY_j+w7BKJ1hV8s1=K-xqqYucZen8os7O_g71gSg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
@ 2013-09-16 11:58 ` Matt Fleming
[not found] ` <20130916115855.GC2697-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>
0 siblings, 1 reply; 7+ messages in thread
From: Matt Fleming @ 2013-09-16 11:58 UTC (permalink / raw)
To: Clea F. Rees
Cc: linux-efi-u79uwXL29TY76Z2rM5mHXA,
matthew.garrett-05XSO3Yj/JvQT0dZR+AlfA, Matt Fleming
On Mon, 09 Sep, at 12:45:00PM, Clea F. Rees wrote:
> 2013/9/5 Matt Fleming <matt-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>:
> >
> > Sorry for the delay in replying to this.
> >
> > Does your machine hang during setup_efi_pci() or later on in boot?
>
> I'm not certain. How would I tell this? I think the answer is probably
> during because the tests Matt had me do involved inserting reboot
> instructions into the code and testing where the hang occurred. But
> I'm not sure whether that showed absolutely that the hang occurs
> actually during setup_efi_pci().
Could you try the attached patch and let me know what you observe?
---
diff --git a/arch/x86/boot/compressed/eboot.c b/arch/x86/boot/compressed/eboot.c
index b7388a4..f4a966d 100644
--- a/arch/x86/boot/compressed/eboot.c
+++ b/arch/x86/boot/compressed/eboot.c
@@ -261,30 +261,37 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
int i;
struct setup_data *data;
+ efi_printk("1\n");
data = (struct setup_data *)(unsigned long)params->hdr.setup_data;
while (data && data->next)
data = (struct setup_data *)(unsigned long)data->next;
+ efi_printk("2\n");
status = efi_call_phys5(sys_table->boottime->locate_handle,
EFI_LOCATE_BY_PROTOCOL, &pci_proto,
NULL, &size, pci_handle);
+ efi_printk("3\n");
if (status == EFI_BUFFER_TOO_SMALL) {
status = efi_call_phys3(sys_table->boottime->allocate_pool,
EFI_LOADER_DATA, size, &pci_handle);
+ efi_printk("4\n");
if (status != EFI_SUCCESS)
return status;
+ efi_printk("5\n");
status = efi_call_phys5(sys_table->boottime->locate_handle,
EFI_LOCATE_BY_PROTOCOL, &pci_proto,
NULL, &size, pci_handle);
+ efi_printk("6\n");
}
if (status != EFI_SUCCESS)
goto free_handle;
+ efi_printk("7\n");
nr_pci = size / sizeof(void *);
for (i = 0; i < nr_pci; i++) {
void *h = pci_handle[i];
@@ -297,13 +304,16 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
if (status != EFI_SUCCESS)
continue;
+ efi_printk("8\n");
if (!pci)
continue;
+ efi_printk("9\n");
#ifdef CONFIG_X86_64
status = efi_call_phys4(pci->attributes, pci,
EfiPciIoAttributeOperationGet, 0,
&attributes);
+ efi_printk("10\n");
#else
status = efi_call_phys5(pci->attributes, pci,
EfiPciIoAttributeOperationGet, 0, 0,
@@ -312,43 +322,54 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
if (status != EFI_SUCCESS)
continue;
+ efi_printk("11\n");
if (!pci->romimage || !pci->romsize)
continue;
+ efi_printk("12\n");
size = pci->romsize + sizeof(*rom);
status = efi_call_phys3(sys_table->boottime->allocate_pool,
EFI_LOADER_DATA, size, &rom);
+ efi_printk("13\n");
if (status != EFI_SUCCESS)
continue;
+ efi_printk("14\n");
rom->data.type = SETUP_PCI;
rom->data.len = size - sizeof(struct setup_data);
rom->data.next = 0;
rom->pcilen = pci->romsize;
+ efi_printk("15\n");
status = efi_call_phys5(pci->pci.read, pci,
EfiPciIoWidthUint16, PCI_VENDOR_ID,
1, &(rom->vendor));
+ efi_printk("16\n");
if (status != EFI_SUCCESS)
goto free_struct;
+ efi_printk("17\n");
status = efi_call_phys5(pci->pci.read, pci,
EfiPciIoWidthUint16, PCI_DEVICE_ID,
1, &(rom->devid));
+ efi_printk("18\n");
if (status != EFI_SUCCESS)
goto free_struct;
+ efi_printk("19\n");
status = efi_call_phys5(pci->get_location, pci,
&(rom->segment), &(rom->bus),
&(rom->device), &(rom->function));
+ efi_printk("20\n");
if (status != EFI_SUCCESS)
goto free_struct;
+ efi_printk("21\n");
memcpy(rom->romdata, pci->romimage, pci->romsize);
if (data)
@@ -358,13 +379,18 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
data = (struct setup_data *)rom;
+ efi_printk("22\n");
continue;
free_struct:
+ efi_printk("23\n");
efi_call_phys1(sys_table->boottime->free_pool, rom);
+ efi_printk("24\n");
}
free_handle:
+ efi_printk("25\n");
efi_call_phys1(sys_table->boottime->free_pool, pci_handle);
+ efi_printk("26\n");
return status;
}
--
Matt Fleming, Intel Open Source Technology Center
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: EFI stub boots only with setup_efi_pc() commented
[not found] ` <20130916115855.GC2697-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>
@ 2013-09-30 22:57 ` Clea F. Rees
[not found] ` <CA+n4xYa0qnsKHDsG5T8+PGKJOwXAEseEGcx4+8fw9scSMttkGg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
0 siblings, 1 reply; 7+ messages in thread
From: Clea F. Rees @ 2013-09-30 22:57 UTC (permalink / raw)
To: Matt Fleming
Cc: linux-efi-u79uwXL29TY76Z2rM5mHXA,
matthew.garrett-05XSO3Yj/JvQT0dZR+AlfA, Matt Fleming
2013/9/16 Matt Fleming <matt-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>:
> On Mon, 09 Sep, at 12:45:00PM, Clea F. Rees wrote:
>> 2013/9/5 Matt Fleming <matt-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>:
>> >
>> > Sorry for the delay in replying to this.
>> >
>> > Does your machine hang during setup_efi_pci() or later on in boot?
>>
>> I'm not certain. How would I tell this? I think the answer is probably
>> during because the tests Matt had me do involved inserting reboot
>> instructions into the code and testing where the hang occurred. But
>> I'm not sure whether that showed absolutely that the hang occurs
>> actually during setup_efi_pci().
>
> Could you try the attached patch and let me know what you observe?
>
> ---
>
> diff --git a/arch/x86/boot/compressed/eboot.c b/arch/x86/boot/compressed/eboot.c
> index b7388a4..f4a966d 100644
> --- a/arch/x86/boot/compressed/eboot.c
> +++ b/arch/x86/boot/compressed/eboot.c
> @@ -261,30 +261,37 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
> int i;
> struct setup_data *data;
>
> + efi_printk("1\n");
> data = (struct setup_data *)(unsigned long)params->hdr.setup_data;
>
> while (data && data->next)
> data = (struct setup_data *)(unsigned long)data->next;
>
> + efi_printk("2\n");
> status = efi_call_phys5(sys_table->boottime->locate_handle,
> EFI_LOCATE_BY_PROTOCOL, &pci_proto,
> NULL, &size, pci_handle);
>
> + efi_printk("3\n");
> if (status == EFI_BUFFER_TOO_SMALL) {
> status = efi_call_phys3(sys_table->boottime->allocate_pool,
> EFI_LOADER_DATA, size, &pci_handle);
>
> + efi_printk("4\n");
> if (status != EFI_SUCCESS)
> return status;
>
> + efi_printk("5\n");
> status = efi_call_phys5(sys_table->boottime->locate_handle,
> EFI_LOCATE_BY_PROTOCOL, &pci_proto,
> NULL, &size, pci_handle);
> + efi_printk("6\n");
> }
>
> if (status != EFI_SUCCESS)
> goto free_handle;
>
> + efi_printk("7\n");
> nr_pci = size / sizeof(void *);
> for (i = 0; i < nr_pci; i++) {
> void *h = pci_handle[i];
> @@ -297,13 +304,16 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
> if (status != EFI_SUCCESS)
> continue;
>
> + efi_printk("8\n");
> if (!pci)
> continue;
>
> + efi_printk("9\n");
> #ifdef CONFIG_X86_64
> status = efi_call_phys4(pci->attributes, pci,
> EfiPciIoAttributeOperationGet, 0,
> &attributes);
> + efi_printk("10\n");
> #else
> status = efi_call_phys5(pci->attributes, pci,
> EfiPciIoAttributeOperationGet, 0, 0,
> @@ -312,43 +322,54 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
> if (status != EFI_SUCCESS)
> continue;
>
> + efi_printk("11\n");
> if (!pci->romimage || !pci->romsize)
> continue;
>
> + efi_printk("12\n");
> size = pci->romsize + sizeof(*rom);
>
> status = efi_call_phys3(sys_table->boottime->allocate_pool,
> EFI_LOADER_DATA, size, &rom);
>
> + efi_printk("13\n");
> if (status != EFI_SUCCESS)
> continue;
>
> + efi_printk("14\n");
> rom->data.type = SETUP_PCI;
> rom->data.len = size - sizeof(struct setup_data);
> rom->data.next = 0;
> rom->pcilen = pci->romsize;
>
> + efi_printk("15\n");
> status = efi_call_phys5(pci->pci.read, pci,
> EfiPciIoWidthUint16, PCI_VENDOR_ID,
> 1, &(rom->vendor));
> + efi_printk("16\n");
>
> if (status != EFI_SUCCESS)
> goto free_struct;
>
> + efi_printk("17\n");
> status = efi_call_phys5(pci->pci.read, pci,
> EfiPciIoWidthUint16, PCI_DEVICE_ID,
> 1, &(rom->devid));
>
> + efi_printk("18\n");
> if (status != EFI_SUCCESS)
> goto free_struct;
>
> + efi_printk("19\n");
> status = efi_call_phys5(pci->get_location, pci,
> &(rom->segment), &(rom->bus),
> &(rom->device), &(rom->function));
>
> + efi_printk("20\n");
> if (status != EFI_SUCCESS)
> goto free_struct;
>
> + efi_printk("21\n");
> memcpy(rom->romdata, pci->romimage, pci->romsize);
>
> if (data)
> @@ -358,13 +379,18 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
>
> data = (struct setup_data *)rom;
>
> + efi_printk("22\n");
> continue;
> free_struct:
> + efi_printk("23\n");
> efi_call_phys1(sys_table->boottime->free_pool, rom);
> + efi_printk("24\n");
> }
>
> free_handle:
> + efi_printk("25\n");
> efi_call_phys1(sys_table->boottime->free_pool, pci_handle);
> + efi_printk("26\n");
> return status;
> }
>
Could you give me a revision number this patch should work with? I
tried it with both the revision I was using before and the current git
and in both cases, the patch failed:
patching file arch/x86/boot/compressed/eboot.c
Hunk #1 FAILED at 261.
Hunk #2 FAILED at 297.
Hunk #3 FAILED at 312.
Hunk #4 FAILED at 358.
4 out of 4 hunks FAILED -- saving rejects to file
arch/x86/boot/compressed/eboot.c.rej
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: EFI stub boots only with setup_efi_pc() commented
[not found] ` <CA+n4xYa0qnsKHDsG5T8+PGKJOwXAEseEGcx4+8fw9scSMttkGg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
@ 2013-10-02 23:22 ` Clea F. Rees
2013-10-04 0:48 ` Clea F. Rees
0 siblings, 1 reply; 7+ messages in thread
From: Clea F. Rees @ 2013-10-02 23:22 UTC (permalink / raw)
To: Matt Fleming
Cc: linux-efi-u79uwXL29TY76Z2rM5mHXA,
matthew.garrett-05XSO3Yj/JvQT0dZR+AlfA, Matt Fleming
Please ignore my previous message. I did not realise that gmail's new
interface was mangling the patch. Compiling now...
2013/9/30 Clea F. Rees <reescf-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>:
> 2013/9/16 Matt Fleming <matt-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>:
>> On Mon, 09 Sep, at 12:45:00PM, Clea F. Rees wrote:
>>> 2013/9/5 Matt Fleming <matt-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>:
>>> >
>>> > Sorry for the delay in replying to this.
>>> >
>>> > Does your machine hang during setup_efi_pci() or later on in boot?
>>>
>>> I'm not certain. How would I tell this? I think the answer is probably
>>> during because the tests Matt had me do involved inserting reboot
>>> instructions into the code and testing where the hang occurred. But
>>> I'm not sure whether that showed absolutely that the hang occurs
>>> actually during setup_efi_pci().
>>
>> Could you try the attached patch and let me know what you observe?
>>
>> ---
>>
>> diff --git a/arch/x86/boot/compressed/eboot.c b/arch/x86/boot/compressed/eboot.c
>> index b7388a4..f4a966d 100644
>> --- a/arch/x86/boot/compressed/eboot.c
>> +++ b/arch/x86/boot/compressed/eboot.c
>> @@ -261,30 +261,37 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
>> int i;
>> struct setup_data *data;
>>
>> + efi_printk("1\n");
>> data = (struct setup_data *)(unsigned long)params->hdr.setup_data;
>>
>> while (data && data->next)
>> data = (struct setup_data *)(unsigned long)data->next;
>>
>> + efi_printk("2\n");
>> status = efi_call_phys5(sys_table->boottime->locate_handle,
>> EFI_LOCATE_BY_PROTOCOL, &pci_proto,
>> NULL, &size, pci_handle);
>>
>> + efi_printk("3\n");
>> if (status == EFI_BUFFER_TOO_SMALL) {
>> status = efi_call_phys3(sys_table->boottime->allocate_pool,
>> EFI_LOADER_DATA, size, &pci_handle);
>>
>> + efi_printk("4\n");
>> if (status != EFI_SUCCESS)
>> return status;
>>
>> + efi_printk("5\n");
>> status = efi_call_phys5(sys_table->boottime->locate_handle,
>> EFI_LOCATE_BY_PROTOCOL, &pci_proto,
>> NULL, &size, pci_handle);
>> + efi_printk("6\n");
>> }
>>
>> if (status != EFI_SUCCESS)
>> goto free_handle;
>>
>> + efi_printk("7\n");
>> nr_pci = size / sizeof(void *);
>> for (i = 0; i < nr_pci; i++) {
>> void *h = pci_handle[i];
>> @@ -297,13 +304,16 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
>> if (status != EFI_SUCCESS)
>> continue;
>>
>> + efi_printk("8\n");
>> if (!pci)
>> continue;
>>
>> + efi_printk("9\n");
>> #ifdef CONFIG_X86_64
>> status = efi_call_phys4(pci->attributes, pci,
>> EfiPciIoAttributeOperationGet, 0,
>> &attributes);
>> + efi_printk("10\n");
>> #else
>> status = efi_call_phys5(pci->attributes, pci,
>> EfiPciIoAttributeOperationGet, 0, 0,
>> @@ -312,43 +322,54 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
>> if (status != EFI_SUCCESS)
>> continue;
>>
>> + efi_printk("11\n");
>> if (!pci->romimage || !pci->romsize)
>> continue;
>>
>> + efi_printk("12\n");
>> size = pci->romsize + sizeof(*rom);
>>
>> status = efi_call_phys3(sys_table->boottime->allocate_pool,
>> EFI_LOADER_DATA, size, &rom);
>>
>> + efi_printk("13\n");
>> if (status != EFI_SUCCESS)
>> continue;
>>
>> + efi_printk("14\n");
>> rom->data.type = SETUP_PCI;
>> rom->data.len = size - sizeof(struct setup_data);
>> rom->data.next = 0;
>> rom->pcilen = pci->romsize;
>>
>> + efi_printk("15\n");
>> status = efi_call_phys5(pci->pci.read, pci,
>> EfiPciIoWidthUint16, PCI_VENDOR_ID,
>> 1, &(rom->vendor));
>> + efi_printk("16\n");
>>
>> if (status != EFI_SUCCESS)
>> goto free_struct;
>>
>> + efi_printk("17\n");
>> status = efi_call_phys5(pci->pci.read, pci,
>> EfiPciIoWidthUint16, PCI_DEVICE_ID,
>> 1, &(rom->devid));
>>
>> + efi_printk("18\n");
>> if (status != EFI_SUCCESS)
>> goto free_struct;
>>
>> + efi_printk("19\n");
>> status = efi_call_phys5(pci->get_location, pci,
>> &(rom->segment), &(rom->bus),
>> &(rom->device), &(rom->function));
>>
>> + efi_printk("20\n");
>> if (status != EFI_SUCCESS)
>> goto free_struct;
>>
>> + efi_printk("21\n");
>> memcpy(rom->romdata, pci->romimage, pci->romsize);
>>
>> if (data)
>> @@ -358,13 +379,18 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
>>
>> data = (struct setup_data *)rom;
>>
>> + efi_printk("22\n");
>> continue;
>> free_struct:
>> + efi_printk("23\n");
>> efi_call_phys1(sys_table->boottime->free_pool, rom);
>> + efi_printk("24\n");
>> }
>>
>> free_handle:
>> + efi_printk("25\n");
>> efi_call_phys1(sys_table->boottime->free_pool, pci_handle);
>> + efi_printk("26\n");
>> return status;
>> }
>>
> Could you give me a revision number this patch should work with? I
> tried it with both the revision I was using before and the current git
> and in both cases, the patch failed:
>
> patching file arch/x86/boot/compressed/eboot.c
> Hunk #1 FAILED at 261.
> Hunk #2 FAILED at 297.
> Hunk #3 FAILED at 312.
> Hunk #4 FAILED at 358.
> 4 out of 4 hunks FAILED -- saving rejects to file
> arch/x86/boot/compressed/eboot.c.rej
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: EFI stub boots only with setup_efi_pc() commented
2013-10-02 23:22 ` Clea F. Rees
@ 2013-10-04 0:48 ` Clea F. Rees
0 siblings, 0 replies; 7+ messages in thread
From: Clea F. Rees @ 2013-10-04 0:48 UTC (permalink / raw)
To: Matt Fleming
Cc: linux-efi-u79uwXL29TY76Z2rM5mHXA,
matthew.garrett-05XSO3Yj/JvQT0dZR+AlfA, Matt Fleming
>>> Could you try the attached patch and let me know what you observe?
Using the patch below, I get a whole bunch of numbers. I hope you do
not need the entire sequence as they scroll off screen too quickly to
read. When the scrolling stops, the screen appears as follows:
11
8
9
10
11
8
9
10
<pattern then repeats several times>
11
25
26
_
What does that mean exactly? That is, why doesn't it ever get to 12?
>>> ---
>>>
>>> diff --git a/arch/x86/boot/compressed/eboot.c b/arch/x86/boot/compressed/eboot.c
>>> index b7388a4..f4a966d 100644
>>> --- a/arch/x86/boot/compressed/eboot.c
>>> +++ b/arch/x86/boot/compressed/eboot.c
>>> @@ -261,30 +261,37 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
>>> int i;
>>> struct setup_data *data;
>>>
>>> + efi_printk("1\n");
>>> data = (struct setup_data *)(unsigned long)params->hdr.setup_data;
>>>
>>> while (data && data->next)
>>> data = (struct setup_data *)(unsigned long)data->next;
>>>
>>> + efi_printk("2\n");
>>> status = efi_call_phys5(sys_table->boottime->locate_handle,
>>> EFI_LOCATE_BY_PROTOCOL, &pci_proto,
>>> NULL, &size, pci_handle);
>>>
>>> + efi_printk("3\n");
>>> if (status == EFI_BUFFER_TOO_SMALL) {
>>> status = efi_call_phys3(sys_table->boottime->allocate_pool,
>>> EFI_LOADER_DATA, size, &pci_handle);
>>>
>>> + efi_printk("4\n");
>>> if (status != EFI_SUCCESS)
>>> return status;
>>>
>>> + efi_printk("5\n");
>>> status = efi_call_phys5(sys_table->boottime->locate_handle,
>>> EFI_LOCATE_BY_PROTOCOL, &pci_proto,
>>> NULL, &size, pci_handle);
>>> + efi_printk("6\n");
>>> }
>>>
>>> if (status != EFI_SUCCESS)
>>> goto free_handle;
>>>
>>> + efi_printk("7\n");
>>> nr_pci = size / sizeof(void *);
>>> for (i = 0; i < nr_pci; i++) {
>>> void *h = pci_handle[i];
>>> @@ -297,13 +304,16 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
>>> if (status != EFI_SUCCESS)
>>> continue;
>>>
>>> + efi_printk("8\n");
>>> if (!pci)
>>> continue;
>>>
>>> + efi_printk("9\n");
>>> #ifdef CONFIG_X86_64
>>> status = efi_call_phys4(pci->attributes, pci,
>>> EfiPciIoAttributeOperationGet, 0,
>>> &attributes);
>>> + efi_printk("10\n");
>>> #else
>>> status = efi_call_phys5(pci->attributes, pci,
>>> EfiPciIoAttributeOperationGet, 0, 0,
>>> @@ -312,43 +322,54 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
>>> if (status != EFI_SUCCESS)
>>> continue;
>>>
>>> + efi_printk("11\n");
>>> if (!pci->romimage || !pci->romsize)
>>> continue;
>>>
>>> + efi_printk("12\n");
>>> size = pci->romsize + sizeof(*rom);
>>>
>>> status = efi_call_phys3(sys_table->boottime->allocate_pool,
>>> EFI_LOADER_DATA, size, &rom);
>>>
>>> + efi_printk("13\n");
>>> if (status != EFI_SUCCESS)
>>> continue;
>>>
>>> + efi_printk("14\n");
>>> rom->data.type = SETUP_PCI;
>>> rom->data.len = size - sizeof(struct setup_data);
>>> rom->data.next = 0;
>>> rom->pcilen = pci->romsize;
>>>
>>> + efi_printk("15\n");
>>> status = efi_call_phys5(pci->pci.read, pci,
>>> EfiPciIoWidthUint16, PCI_VENDOR_ID,
>>> 1, &(rom->vendor));
>>> + efi_printk("16\n");
>>>
>>> if (status != EFI_SUCCESS)
>>> goto free_struct;
>>>
>>> + efi_printk("17\n");
>>> status = efi_call_phys5(pci->pci.read, pci,
>>> EfiPciIoWidthUint16, PCI_DEVICE_ID,
>>> 1, &(rom->devid));
>>>
>>> + efi_printk("18\n");
>>> if (status != EFI_SUCCESS)
>>> goto free_struct;
>>>
>>> + efi_printk("19\n");
>>> status = efi_call_phys5(pci->get_location, pci,
>>> &(rom->segment), &(rom->bus),
>>> &(rom->device), &(rom->function));
>>>
>>> + efi_printk("20\n");
>>> if (status != EFI_SUCCESS)
>>> goto free_struct;
>>>
>>> + efi_printk("21\n");
>>> memcpy(rom->romdata, pci->romimage, pci->romsize);
>>>
>>> if (data)
>>> @@ -358,13 +379,18 @@ static efi_status_t setup_efi_pci(struct boot_params *params)
>>>
>>> data = (struct setup_data *)rom;
>>>
>>> + efi_printk("22\n");
>>> continue;
>>> free_struct:
>>> + efi_printk("23\n");
>>> efi_call_phys1(sys_table->boottime->free_pool, rom);
>>> + efi_printk("24\n");
>>> }
>>>
>>> free_handle:
>>> + efi_printk("25\n");
>>> efi_call_phys1(sys_table->boottime->free_pool, pci_handle);
>>> + efi_printk("26\n");
>>> return status;
>>> }
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2013-10-04 0:48 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-08-02 23:06 EFI stub boots only with setup_efi_pc() commented Clea F. Rees
[not found] ` <CA+n4xYYJhoXp6KSeROda9ErrXZ0hFEBM-x-0Obvk3fmHwABfKw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2013-09-05 8:29 ` Matt Fleming
[not found] ` <20130905082924.GI28598-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>
2013-09-09 11:45 ` Clea F. Rees
[not found] ` <CA+n4xYZq3wpY_j+w7BKJ1hV8s1=K-xqqYucZen8os7O_g71gSg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2013-09-16 11:58 ` Matt Fleming
[not found] ` <20130916115855.GC2697-HNK1S37rvNbeXh+fF434Mdi2O/JbrIOy@public.gmane.org>
2013-09-30 22:57 ` Clea F. Rees
[not found] ` <CA+n4xYa0qnsKHDsG5T8+PGKJOwXAEseEGcx4+8fw9scSMttkGg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2013-10-02 23:22 ` Clea F. Rees
2013-10-04 0:48 ` Clea F. Rees
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).